test(capture): specify coherent snapshot view ownership

.agents/skills/adversarial-validation/references/performance.md

@@ -6,17 +6,7 @@
 - Route every new sync/flush through the durability-mode and bucket override
   gates; mode `none` must not pay the new fsync.
 - Keep blocking filesystem/CPU work off async runtime threads, but do not split
-  one small operation into many `spawn_blocking` round trips; coalesce `open`,
-  `read`/`write`, `fsync`, and `close` into one blocking closure per disk.
-- Check for per-disk syscall amplification (`disks × parts/files`): pin a
-  directory once (`openat2` on Linux) and reuse the handle across sibling reads
-  instead of re-walking ancestor components per file or re-opening identical
-  `source_parent == destination_parent` directories.
-- After `fstat`/`file.metadata()`, require `Vec::with_capacity(len + 1)` before
-  `read_to_end` so small metadata reads do not trigger 6–7 exponential-growth
-  `read()` syscalls.
-- Avoid redundant `store.load()` or `root_missing()` round-trips under the same
-  exclusive lock when the state or original bytes were just loaded or persisted.
+  one small operation into many `spawn_blocking` round trips.
 - Measure lock hold time across I/O and compare acquisition order for ABBA.
 - Keep cleanup, extra stat/rename, and diagnostics out of the PUT commit critical
   section when they need not be there.

.agents/skills/rustfs-release-publish/SKILL.md

@@ -25,7 +25,7 @@ Read only the current phase's procedure; retain completed evidence across turns.
 | 4–5 | [Run the downloaded binary, Console CRUD, and latest-rc matrix](references/preview-acceptance.md). Every check must pass. |
 | Confirmation | Complete the gate below and end the turn. |
 | 6 | After fresh confirmation, [publish and verify the final tag](references/final-release.md) at `PREVIEW_HASH`. |
-| 7 | [Reconcile Security Advisories, milestones, installation references, and the website; reconcile main and remove the completed release branch](references/post-release-updates.md) only after complete publication. |
+| 7 | [Maintain milestones, installation references, and the website](references/post-release-updates.md) only after complete publication. |
 
 ## Human Confirmation
 
@@ -37,6 +37,6 @@ Only a new affirmative user reply to that exact report authorizes Phase 6. The o
 
 - Before the final tag exists, land fixes on main, backport to the candidate, verify target versions, and restart branch CI/acceptance. Use a new preview iteration once a preview was tagged; main may advance independently.
 - Once the final tag exists, retry publication jobs for that exact tag. A source fix requires a newly confirmed target, not retagging or a newer preview. Phase 7 retries recheck artifact availability and superseding releases.
-- If abandoned after the bump merged, report the unpublished version on main and record whether to revert or let the next release overwrite it. Deleting unpublished candidate branches or preview tags requires an explicit decision. Completed release branches follow Phase 7 cleanup; keep release tags.
+- If abandoned after the bump merged, report the unpublished version on main and record whether to revert or let the next release overwrite it. Deleting candidate branches or preview tags requires an explicit decision.
 
-Report completed phase results and incomplete work with evidence: Console previous/latest tags, release need, hash and run/Release URLs; target, branch CI/SHA, previews and hash, confirmation state; preview/final assets, latest-channel state, Console/rc results and cleanup; advisory reconciliation results and pending disclosure decisions; both milestone states, installation PR checks/merge state, and website text/link/PR/deployment state. Keep artifact publication separate from follow-up delivery; do not call an open PR a live update. Report deviations and their authorization.
+Report completed phase results and incomplete work with evidence: Console previous/latest tags, release need, hash and run/Release URLs; target, branch CI/SHA, previews and hash, confirmation state; preview/final assets, latest-channel state, Console/rc results and cleanup; both milestone states, installation PR checks/merge state, and website text/link/PR/deployment state. Keep artifact publication separate from follow-up delivery; do not call an open PR a live update. Report deviations and their authorization.

.agents/skills/rustfs-release-publish/references/post-release-updates.md

@@ -1,17 +1,7 @@
-# Post-release Security Advisory, Milestone, Installation, and Website Updates
+# Post-release Milestone, Installation, and Website Updates
 
 Read only after the final non-preview tag has passed Phase 6. Tag creation or a green binary build alone is insufficient: the release assets, source archive, container images, Helm package, and applicable latest channels must be available. Preview tags never enter this phase.
 
-## Security Advisories
-
-Reconcile advisories after every completed final release, including Console-only releases. Read-only preparation may start earlier; do not claim an unpublished artifact is an available fix.
-
-1. Fetch every page of the repository's live Security Advisories, including triage, draft, published, and closed states. A permission error or partial response blocks reconciliation; it does not mean there are no advisories. Keep unpublished advisory details in private task artifacts, never in a public PR, release note, or issue.
-2. For each potentially relevant advisory, map its fix to the exact published tag and verify the regression evidence. Main-branch reachability, a merged private fix, or a green build alone does not prove the shipped version is fixed. Check cherry-picked equivalents and reverted fixes explicitly. For embedded Console or dependency fixes, verify the shipped component version.
-3. Determine the first available fixed release and affected version ranges from evidence. Preserve an earlier correct patched version; do not relabel all fixes as the current release. An unchanged server baseline such as a Console-only release requires checking the earlier server release as well. Leave unresolved reports and fixes absent from the tag pending; do not close them merely because a release shipped.
-4. Prepare any necessary corrections to package identity, affected ranges, patched versions, fix/release links, and advisory text. Preserve unrelated fields, credit, CVE identifiers, and disclosure coordination. Apply updates only within existing user authorization; read back the result to verify it. Publishing a draft, requesting a CVE, or closing a report is a separate action: require explicit authorization for the concrete advisories and respect any embargo. Release-tag approval alone does not authorize those actions. Prepare the exact proposed changes before asking for missing authorization.
-5. Record each relevant advisory's URL, current state, fix evidence, first fixed version, action taken or proposed, and remaining blocker in private task artifacts. Report advisory follow-up separately from artifact publication. A successful release with unresolved advisory work is not complete follow-up; retain that work for the release monitor or a concrete user decision.
-
 ## Version milestones
 
 1. Query all pages of both open and closed milestones in the RustFS release repository, matching titles exactly to version strings without a `v` prefix. A failed or incomplete query is not evidence that a milestone is absent; resolve the lookup before making changes.
@@ -48,10 +38,3 @@ Use the existing announcement configuration from [website PR #112](https://githu
 3. Use the site's current package manager and required checks. For this config-only change, type-check and lint the changed file, build the site, and inspect the generated homepage for the version, visible link text, and destination. Confirm `/download` actually offers the announced version before using that link.
 4. Commit, push, and create or update the website PR when included in the authorized release scope. Include the published RustFS Release URL and checks run. Keep merge and production-deployment authority separate; reuse permission already given, but this reference grants none by itself.
 5. After an authorized merge/deployment, verify the live `https://rustfs.com/` banner and its destination. If only the PR is ready, report that state and URL; do not report the banner as deployed. A website failure does not invalidate or move the published release tag: retry this phase after resolving the failure.
-
-## Reconcile main and remove the completed release branch
-
-1. Compare the exact published tag with current main. Audit every release-side commit since their merge base, including inherited backports. An ahead count is a history difference, not proof of missing changes. Use ancestry and patch equivalence, then inspect current code and regression evidence for squashed, adapted, superseded, or reverted changes. Record each commit as present, superseded, release-only, or missing; a positive `git cherry` result alone does not prove a missing fix.
-2. Bring missing shared changes to main through a reviewed PR and the required checks before closing the follow-up. Preserve main's later changes; decide version bumps and pinned Console configuration separately. If a history merge is needed, use a merge commit and verify the release commit is an ancestor of main; squash does not establish that ancestry. Do not merge main into the published release branch or move its tags. Report unresolved changes instead of claiming reconciliation is complete.
-3. After publication and authorized follow-ups are complete, delete the remote `release/<target>` branch. Immediately before deletion, verify its HEAD still equals the published tag's commit and no open PR targets or originates from it. Preserve a branch with newer work or open PRs and report the blocker. Use a deletion with an expected-old-SHA check to avoid removing a concurrently advanced branch.
-4. Verify the remote branch is absent and the final and preview tags still resolve to their recorded commits. Remove a local release branch only if it is task-owned, unchecked out, and has no unpreserved work. Record cleanup separately from artifact publication; an explicitly authorized deletion before reconciliation does not make unresolved main changes complete.

.agents/skills/rustfs-release-version-bump/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "RustFS Release Bump"
+  short_description: "Prepare RustFS release branches like PR #2957."
+  default_prompt: "Use $rustfs-release-version-bump to prepare and verify an exact RustFS release-version bump."

.agents/skills/security-advisory-lessons/SKILL.md

@@ -36,21 +36,6 @@ changed lesson.
 
 ## Finding Standard
 
-Every candidate names the lower-trust principal, the boundary it crosses, the
-affected resource, and the observable result. A best-practice gap or an impact
-confined to the caller's own principal is hardening, not a candidate.
-
-Report each candidate under exactly one verdict:
-
-- `confirmed`: a complete `file:line` source trace and a bounded observed
-  result. Only confirmed findings carry severity; each also states attacker
-  prerequisites, concrete input/path, impact, smallest safe fix, and a
-  regression check.
-- `needs_validation`: the trace is blocked on one exact unresolved fact, such
-  as deployment or infrastructure behavior. State that fact; assign no severity
-  and do not substitute an assumption.
-- `rejected`: a disproved candidate, with the caller, invariant, or test that
-  disproves it.
-
-Do not exaggerate unauthenticated impact when the actual issue requires
-authenticated low privilege.
+Each finding includes severity, `file:line`, attacker prerequisites, concrete
+input/path, impact, smallest safe fix, and a regression check. Do not exaggerate
+unauthenticated impact when the actual issue requires authenticated low privilege.

.agents/skills/security-advisory-lessons/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "Security Advisory Lessons"
+  short_description: "Apply advisory lessons in reviews."
+  default_prompt: "Use $security-advisory-lessons for a dedicated RustFS security review grounded in past advisories."

.agents/skills/security-advisory-lessons/references/advisory-patterns.md

@@ -45,13 +45,11 @@ Update this file only when an advisory adds or changes a reusable lesson, affect
 - `GHSA-6r96-hmgc-726c`: request headers collided with lowercase server-derived condition keys such as `userid`, `groups`, `versionid`, and JWT/LDAP claims. Lesson: never let caller-controlled headers append to or replace server-derived policy context; reserve trusted condition keys and keep intentional request-header keys separate.
 - `GHSA-v9cp-qfw9-9pfp`: quantified negated string conditions applied negation after aggregation, transposing `ForAllValues` and `ForAnyValue` semantics. Lesson: push negation into the per-value predicate for quantified operators and test partially overlapping multi-value sets.
 - `GHSA-5w8r-p896-6vq2`: OPA policy mode skipped `ExistingObjectTag/*` loading, so tagged objects looked untagged to external policies. Lesson: external authorization plugins need the same object-tag and request context as built-in policy evaluation before they decide.
-- `GHSA-9wgh-5r68-49p7`: substituted policy-variable values could recreate `${...}` and loop inside a single resolution pass, never reaching the outer iteration limit. Lesson: bound substitutions within each pass and test cycles formed across otherwise literal resolver values; confirm whether attacker-controlled claims can reach those values before assigning remote impact.
 
 ### S3 object actions, copy, multipart, and upload policy validation
 
 - `GHSA-g8w9-qw9q-fghr` and `GHSA-xm99-m3gq-83g8`: a valid presigned `PutObject` accepted extra `x-amz-*` headers omitted from `SignedHeaders`; `x-amz-copy-source` could turn that upload capability into a cross-bucket read performed as the signer. Lesson: a presigned URL is a bounded capability; inspect all received security-sensitive headers and reject unsigned ones before selecting the S3 operation or reaching storage.
 - `GHSA-3ppv-fx5m-m749`: explicit `versionId` reads and copy sources authorized `s3:GetObject` instead of `s3:GetObjectVersion`. Lesson: version-specific object access must select version-specific actions for direct reads, `CopyObject`, and `UploadPartCopy`, with tests proving the backend is not reached on denial.
-- `GHSA-cg33-pxfp-qcj3` and `GHSA-65wq-gjw8-xr9w`: `HeadObject` and object-tagging paths authorized current-object actions while operating on an explicit historical `versionId`. Lesson: select the corresponding version action for metadata and tag reads and writes too; audit every operation that accepts a version ID.
 - `GHSA-gjph-xcgw-4p7w`: `PutObject` and browser POST accepted a caller-supplied `versionId` and replaced an unlocked historical version with only `s3:PutObject`. Lesson: only replication-authorized writes may choose an existing version ID; ordinary uploads must not rewrite version history.
 - `GHSA-x298-9x87-fvjq`: anonymous `ListObjectVersions` fell back to `ListBucket` and returned before public-access-block gates. Lesson: compatibility fallbacks must converge on the same post-authorization checks as direct grants, especially `RestrictPublicBuckets` and anonymous data-plane denies.
 - `GHSA-mx42-j6wv-px98`: `UploadPartCopy` missed source authorization and allowed cross-bucket object exfiltration. Lesson: multipart copy must enforce the same source and destination contract as `CopyObject`.
@@ -112,7 +110,6 @@ Update this file only when an advisory adds or changes a reusable lesson, affect
 
 - `GHSA-j548-9grx-fh4f`: Object Lock enforcement treated unreadable, fabricated, or unparsable bucket metadata as absent configuration and allowed retained objects to be deleted or expired. Lesson: retention must fail closed unless Object Lock absence is authoritative, and every delete, lifecycle, scanner, force-delete, and default-retention path needs the same state distinction.
 - `GHSA-28w4-rp2p-636w` and `GHSA-c7hx-vr97-45rg`: an authorized `DeleteBucket` with `x-rustfs-force-delete: true` removed active COMPLIANCE-retained versions even though ordinary version deletion was denied. Lesson: bucket force-delete must enforce retention across every version, including against root/owner callers.
-- `GHSA-q23w-fp7p-472v`: `PutObjectRetention` policy conditions read `x-amz-object-lock-*` headers while the handler applied mode and date from the XML body. Lesson: authorize the parsed retention that will be stored, or reject mismatched header/body values; compare retention dates as instants so offsets cannot change the applied bound.
 
 ### Serde deserialization and input validation
 
@@ -146,13 +143,13 @@ rg -n "deny_unknown_fields|serde.default|as u32|as usize|as i32" rustfs crates
 - IAM fixes: include import/update/list service-account cases with attacker-controlled parent, claims, access key, secret key, and policy.
 - Copy/upload fixes: include cross-bucket, cross-user, source-denied, destination-denied, copy-source-condition, and multipart completion cases.
 - Presigned upload fixes: include a valid presign with extra unsigned tagging, redirect, storage-class, and cross-bucket copy-source headers; require rejection before source or destination storage access, and verify explicitly signed equivalents still work.
-- Version-action fixes: include historical UUID, explicit current version, `null`, range, partNumber, presigned, STS/session, service-account, anonymous bucket-policy, copy source, multipart-copy source, `HeadObject`, and tag read/write cases.
+- Version-action fixes: include historical UUID, explicit current version, `null`, range, partNumber, presigned, STS/session, service-account, anonymous bucket-policy, copy source, and multipart-copy source cases.
 - Version-write fixes: a `PutObject`-only principal must not replace an existing version through `PutObject` or browser POST `?versionId=`; verify replication-authorized writes and versioned, suspended, and unversioned buckets separately.
-- Policy-condition fixes: include reserved-key header collisions, missing keys, partially overlapping multi-value sets, cyclic variable substitutions, plugin mode, and built-in policy mode.
+- Policy-condition fixes: include reserved-key header collisions, missing keys, partially overlapping multi-value sets, plugin mode, and built-in policy mode.
 - Path fixes: include encoded traversal, absolute path, nested traversal, archive entries with `..`, valid object keys that resemble traversal text but should be rejected, and canonical bucket/prefix boundary checks; test per-key `DeleteObjects` errors and prefix-scoped denials on versioned buckets.
 - Logging fixes: assert redacted output for structs and response bodies that may contain credentials.
 - IAM export fixes: assert exported archives omit plaintext user and service-account secrets unless the format deliberately encrypts or seals them.
 - RPC auth fixes: include captured metadata replay across two concrete methods, stale timestamps, wrong path, wrong method surrogate, wrong secret, and valid same-method calls.
 - Browser/CORS fixes: assert no credentials on reflected/default origins, correct behavior for explicit allowlists, and no same-origin script execution for previewed object content.
 - SSE fixes: inspect stored bytes and verify API metadata, read-back behavior, and on-disk ciphertext together; for ETags, upload equal plaintext under different keys and assert list-only callers cannot derive or correlate the plaintext digest, including multipart objects.
-- Object Lock fixes: include unreadable metadata, fabricated metadata defaults, unparsable config, confirmed absent config, COMPLIANCE/GOVERNANCE retention, lifecycle expiry, scanner sweeps, and force-delete paths; an authorized force-delete must preserve active COMPLIANCE versions. For `PutObjectRetention`, test conflicting header/body values and offset dates against policy bounds and read-back retention.
+- Object Lock fixes: include unreadable metadata, fabricated metadata defaults, unparsable config, confirmed absent config, COMPLIANCE/GOVERNANCE retention, lifecycle expiry, scanner sweeps, and force-delete paths; an authorized force-delete must preserve active COMPLIANCE versions.

.codex/config.toml

@@ -0,0 +1,5 @@
+# Bound individual tool outputs retained in context; retrieve relevant ranges
+# from task-owned log files when more evidence is needed.
+# https://learn.chatgpt.com/docs/config-file/config-reference
+
+tool_output_token_limit = 4000

.config/e2e-distributed-selection.txt

@@ -1,2 +1,2 @@
-sha256-linux=6190506b9e4bdb60fd392544a7bc47be708077bb4fe7a2ef689d047455e1af86
-sha256-darwin=6190506b9e4bdb60fd392544a7bc47be708077bb4fe7a2ef689d047455e1af86
+sha256-linux=233bd7a68777eda1f044b6cb77128056be2b6492583c0dfd114b21e98a6102ef
+sha256-darwin=233bd7a68777eda1f044b6cb77128056be2b6492583c0dfd114b21e98a6102ef

.config/e2e-full-selection.txt

@@ -1,2 +1,2 @@
-sha256-darwin=a67a347f26be186c598b5beaf68e6bcbe1855350a35e1423dd5b8a81bcba06a5
-sha256-linux=ad4fabe811c6cca1aba89fad34ccd404b5832a6a955875c2826f6e0fa1ea4f1f
+sha256-darwin=98dff0961c6eb21d6128bfd2d4a6100a624ddf1fb7db05dba360a5d3d0a310c2
+sha256-linux=890f5925b87e1bf3710c356253f3b2276d901e156df9ccc04ec9315374de6d22

.config/e2e-nightly-selection.txt

@@ -1,2 +1,2 @@
-sha256-darwin=3814cdde8cacf4187e18188d342ed70931cc32e268e42e5bb438d565bb4d0f81
-sha256-linux=663025a80273579f95ae3f49af02c0afcbfce94df09701c360a13e607b7637db
+sha256-darwin=efe68675f4676a0d2901242c5815b1f7af5cdec8231f7cccd408f6dacf5c75db
+sha256-linux=003ed6480935a0a95680cc061be143722fbd37bd477b8ff98607b1a064f3aef1

.config/e2e-smoke-selection.txt

@@ -1 +1 @@
-sha256=f065e7071aac9d48a7c1b301f328fa7cc14f11c4e7dc46da3e55057c4ea29bd5
+sha256=ba8f392b617cf9f2338142233a417e43927b28096635e628233ce8a60cf49d10

.config/ecstore-required-tests.json

@@ -29,7 +29,7 @@
     {
       "invariant": "list-uncommitted-version",
       "suite": "rustfs-filemeta",
-      "name": "metacache::tests::resolve_merge_keeps_the_newest_version_at_listing_quorum"
+      "name": "metacache::tests::resolve_with_write_quorum_slack_keeps_partial_latest_hidden_during_merge"
     },
     {
       "invariant": "minio-object-fixture",
@@ -241,6 +241,11 @@
       "suite": "rustfs-ecstore",
       "name": "store::list_objects::differential_tests::list_objects_shared_corpus_cancelled_quorum_walk_releases_senders_and_disk_lock"
     },
+    {
+      "invariant": "list-objects-shared-corpus-enabled-index-contracts",
+      "suite": "rustfs-ecstore",
+      "name": "store::list_objects::differential_tests::list_objects_shared_corpus_enabled_index_contracts"
+    },
     {
       "invariant": "borrowed-context-boundary",
       "suite": "rustfs-ecstore",

.config/make/tests.mak

@@ -52,10 +52,6 @@ script-tests: ## Run shell script tests
 	./scripts/check_embedded_secrets.sh --self-test
 	$(RUSTFS_PYTHON_BIN) ./scripts/check_test_wiring.py --self-test
 	$(RUSTFS_PYTHON_BIN) ./scripts/test_e2e_binary.py
-	$(RUSTFS_PYTHON_BIN) ./scripts/test_inspect_btf_prototypes.py
-	$(RUSTFS_PYTHON_BIN) ./scripts/test_release_build_workflow.py
-	$(RUSTFS_PYTHON_BIN) ./scripts/test_release_glibc.py
-	$(RUSTFS_PYTHON_BIN) ./scripts/test_upgrade_workflow.py
 	$(RUSTFS_PYTHON_BIN) ./scripts/test_migration_gate_evidence.py
 	$(RUSTFS_PYTHON_BIN) ./scripts/ci_gate.py --self-test
 	$(RUSTFS_PYTHON_BIN) ./scripts/check_security_coverage.py --self-test

.config/nextest.toml

@@ -68,17 +68,6 @@ command = ['sh', '-c', 'echo RUST_MIN_STACK=33554432 >> "$NEXTEST_ENV"']
 # boundaries. Reserve the run's capacity so unrelated fixtures cannot starve
 # their existing deadlines and typed-failure assertions (rustfs/backlog#2706).
 # Keep each test's internal concurrency and zero-retry policy unchanged.
-[[profile.default.overrides]]
-filter = """
-package(rustfs-ecstore) & (
-    test(/^set_disk::ops::multipart::tests::complete_multipart_(quorum_success_releases_logical_guards_before_physical_tail|waits_for_tail_before_releasing_guards_and_marking_capacity)$/)
-    | test(=store::heal::tests::full_format_heal_lost_after_last_save_does_not_publish_or_renew_and_latches)
-    | test(/^store::init::tests::(decommission_durable_ilm_receipt_pagination_fails_closed_on_second_page|dispatch_manifest_shutdown_(drains_admitted_delete_batch_without_tail_admission|stops_aborting_validation_tail_reads|stops_authorized_tail_reads_before_receipt_or_cas))$/)
-    | (binary(metadata_read_lock_reuse_test) & test(=replica_multipart_completes_while_metadata_writer_waits_for_its_snapshot))
-)
-"""
-threads-required = "num-test-threads"
-
 [[profile.default.overrides]]
 filter = 'package(rustfs) & test(/^connect::diagnostics::schedule::tests::/)'
 threads-required = "num-test-threads"
@@ -87,50 +76,18 @@ threads-required = "num-test-threads"
 filter = 'package(rustfs-ecstore) & (test(/^set_disk::tests::streaming_get_blocks_concurrent_/) | test(=store::object::tests::retired_marker_converges_after_bucket_recreation_on_sixteen_disks))'
 threads-required = "num-test-threads"
 
-# These real-disk lock probes must finish durable IO within production's 5s
-# acquisition budget. Reserve capacity from sibling test processes while
-# preserving the probes' internal concurrency, deadlines, and assertions.
-[[profile.default.overrides]]
-filter = '(package(rustfs-ecstore) & test(=core::pools_test::decommission_lock_order_tests::active_decommission_migration_does_not_invert_capacity_and_delete_object_tag_locks)) | (package(rustfs-heal) & test(=heal::mrf_queue::partial_write::tests::durable_retry_submission_error_keeps_checkpoint_and_uses_backoff))'
-threads-required = "num-test-threads"
-
-# These eight-disk decommission fixtures exercise concurrent retries,
-# free-version cleanup, and real capacity probes under production lock deadlines.
-# Reserve the run's capacity so unrelated IO cannot exhaust those budgets.
-[[profile.default.overrides]]
-filter = 'package(rustfs-ecstore) & (test(/^store::init::tests::decommission_entry_(retries_source_changed_without_canceling_other_bucket|skips_cleanup_only_marker_when_free_version_is_present)$/) | test(=core::pools::tests::decommission_capacity_mutation_limits_prewrite_sampling_to_target))'
-threads-required = "num-test-threads"
-
-# This registration/backup lock probe uses an isolated eight-disk store.
-# Reserve capacity so sibling disk-heavy tests cannot consume its 2s
-# registration budget; retain the queued backup and deadlock assertions.
-[[profile.default.overrides]]
-filter = 'package(rustfs) & test(=admin::handlers::table_catalog::tests::cold_strong_registration_completes_while_other_bucket_backup_waits)'
-threads-required = "num-test-threads"
-
 [[profile.default.overrides]]
 filter = 'package(rustfs-heal) & (binary(mrf_partial_write_test) | (binary(mrf_pipeline_test) & test(/^journal_replay_survives_/)) | test(=admin_selector_survives_sigkill_with_original_scope_and_receipts))'
 threads-required = "num-test-threads"
 
-# Native scanner backlog fixtures include 24-set and pool-expansion cases.
-# Reserve capacity across test processes while preserving each scenario's fixed
-# budget, assertions, and internal concurrency.
 [[profile.default.overrides]]
-filter = 'package(rustfs-scanner) & (test(/^scanner::backlog::tests::native_/) | test(=scanner_folder::tests::mrf_ownership::mrf_ownership_manager_completion_preserves_scanner_pending) | test(/^scanner_io::tests::service_cohort::/))'
+filter = 'package(rustfs-scanner) & (test(/^scanner::backlog::tests::native_(writer|retirement)_/) | test(=scanner_folder::tests::mrf_ownership::mrf_ownership_manager_completion_preserves_scanner_pending) | test(/^scanner_io::tests::service_cohort::/))'
 threads-required = "num-test-threads"
 
 [[profile.default.scripts]]
 filter = 'package(rustfs-ecstore) & test(/^(bucket::lifecycle::bucket_lifecycle_ops::tests::manual_transition_worker_result_recovery_marks_unknown_for_corrupt_marker|services::rebalance::entry::tests::real_rebalance_run_fence_loss_blocks_multipart_publication|store::init::tests::(batch_transitioned_delete_uses_free_version_per_item|decommission_entry_(allows_free_version_consumed_before_source_lock|rejects_subquorum_free_version_conflict_and_retains_source|skips_cleanup_only_marker_when_free_version_is_present)|dispatched_tier_delete_recovery_(checks_later_pool_then_commits_after_source_removal|finds_directory_source_on_encoded_set|retains_journal_on_source_metadata_error)|force_tier_remove_blocks_on_physical_free_version_hidden_by_other_pool|legacy_unknown_transition_delete_falls_back_for_single_batch_and_blocks_prefix|multi_pool_(recursive_prefix_rejects_legacy_or_hidden_merge_loser_before_delete|same_remote_tuple_(batch|single)_delete_waits_for_all_sources|same_tuple_recursive_prefix_uses_one_journal_owner|transitioned_delete_persists_one_free_version_per_remote_tuple)|recursive_prefix_partial_(pool|set)_failure_keeps_prepared_cleanup_owners|restored_transitioned_delete_uses_free_version_as_cleanup_owner|stable_transitioned_recursive_prefix_delete_uses_journal_owners|suspended_null_transition_delete_uses_free_version_as_sole_owner|tier_mutation_peer_handler_applies_prepare_commit_and_abort_idempotently|transition_response_loss_persists_unknown_outcome_for_provider_recovery|transition_transaction_recovery_(drops_record_after_confirmed_local_commit|keeps_cleanup_pending_local_commit)|transitioned_delete_(free_version_replays_after_store_restart|local_quorum_failure_rolls_back_without_cleanup_owner|uses_free_version_as_cleanup_owner)|versioned_delete_marker_keeps_transitioned_source_and_remote_object|versioned_explicit_transition_delete_preserves_other_version_then_allows_bucket_delete))$/)'
 setup = 'ecstore-large-stack'
 
-# These cross-crate real-storage fixtures need the same bounded stack as ECStore tests.
-# The admin target-repair fixtures compose the full bucket delete and recreate
-# path on a real store; measured with RUST_MIN_STACK on 2026-10-15 they need
-# 2016-2040 KiB of the 2 MiB libtest default and overflowed on Linux CI.
-[[profile.default.scripts]]
-filter = '(package(rustfs) & test(/^(admin::handlers::replication::target_repair_tests::[a-z_]+|admin::handlers::site_replication::tests::durable_iam_state_apply_iam_item_rejects_a_stale_recreate_after_a_replicated_delete|admin::service::federated_identity::tests::(binding_keeps_explicit_iam_mapping_strict|markerless_legacy_oidc_session_crosses_auth_and_iam_deletion_boundaries)|storage::scanner_dirty_journal::tests::(clear_callback_keeps_unknown_coverage_until_all_pending_work_is_confirmed|owned_journals_keep_peer_records_after_interleaved_writes_and_clear))$/)) | (package(rustfs-heal) & ((binary(heal_b5_versioned_regression_test) & test(/^serial_tests::(test_heal_unversioned_bucket_e2e|test_historical_null_recursive_heal_restores_physical_versions|test_recursive_deep_heal_reports_stale_delete_marker_repair)$/)) | (binary(usage_observation_read_repair_test) & test(=usage_observation_read_repair_survives_snapshot_cleanup)))) | (package(rustfs-test-utils) & test(=data_usage_snapshot_tests::data_usage_snapshot_recovery_wires_primary_backup_and_negative_cache))'
-setup = 'ecstore-base-stack'
-
 [[profile.default.scripts]]
 filter = 'package(rustfs-ecstore) | package(rustfs-s3select-api) | package(rustfs-scanner) | (package(rustfs) & test(/^(app::multipart_usecase::tests::concurrent_completions_share_durable_bucket_quota_reservations|app::object::delete::tests::compressed_delete_requests_update_observed_usage_without_releasing_quota_floor|app::object::internal_put::tests::internal_multipart_roundtrip_completes_and_abort_leaves_nothing|app::object::restore::tests::execute_restore_object_maps_failures_to_typed_s3_errors|storage::access::tests::(delete_object_access_captures_authorized_bucket_incarnation|copy_operations_reject_recreated_source_bucket_after_authorization|request_slot_keeps_bucket_policy_bound_to_its_store))$/))'
 setup = 'ecstore-base-stack'
@@ -362,17 +319,6 @@ path = "junit.xml"
 
 # Match the local durable-progress reservations. All members, deadlines,
 # assertions, and internal concurrency remain enabled without retries.
-[[profile.ci.overrides]]
-filter = """
-package(rustfs-ecstore) & (
-    test(/^set_disk::ops::multipart::tests::complete_multipart_(quorum_success_releases_logical_guards_before_physical_tail|waits_for_tail_before_releasing_guards_and_marking_capacity)$/)
-    | test(=store::heal::tests::full_format_heal_lost_after_last_save_does_not_publish_or_renew_and_latches)
-    | test(/^store::init::tests::(decommission_durable_ilm_receipt_pagination_fails_closed_on_second_page|dispatch_manifest_shutdown_(drains_admitted_delete_batch_without_tail_admission|stops_aborting_validation_tail_reads|stops_authorized_tail_reads_before_receipt_or_cas))$/)
-    | (binary(metadata_read_lock_reuse_test) & test(=replica_multipart_completes_while_metadata_writer_waits_for_its_snapshot))
-)
-"""
-threads-required = "num-test-threads"
-
 [[profile.ci.overrides]]
 filter = 'package(rustfs) & test(/^connect::diagnostics::schedule::tests::/)'
 threads-required = "num-test-threads"
@@ -381,47 +327,18 @@ threads-required = "num-test-threads"
 filter = 'package(rustfs-ecstore) & (test(/^set_disk::tests::streaming_get_blocks_concurrent_/) | test(=store::object::tests::retired_marker_converges_after_bucket_recreation_on_sixteen_disks))'
 threads-required = "num-test-threads"
 
-# These real-disk lock probes must finish durable IO within production's 5s
-# acquisition budget. Reserve capacity from sibling test processes while
-# preserving the probes' internal concurrency, deadlines, and assertions.
-[[profile.ci.overrides]]
-filter = '(package(rustfs-ecstore) & test(=core::pools_test::decommission_lock_order_tests::active_decommission_migration_does_not_invert_capacity_and_delete_object_tag_locks)) | (package(rustfs-heal) & test(=heal::mrf_queue::partial_write::tests::durable_retry_submission_error_keeps_checkpoint_and_uses_backoff))'
-threads-required = "num-test-threads"
-
-# These eight-disk decommission fixtures exercise concurrent retries,
-# free-version cleanup, and real capacity probes under production lock deadlines.
-# Reserve the run's capacity so unrelated IO cannot exhaust those budgets.
-[[profile.ci.overrides]]
-filter = 'package(rustfs-ecstore) & (test(/^store::init::tests::decommission_entry_(retries_source_changed_without_canceling_other_bucket|skips_cleanup_only_marker_when_free_version_is_present)$/) | test(=core::pools::tests::decommission_capacity_mutation_limits_prewrite_sampling_to_target))'
-threads-required = "num-test-threads"
-
-# This registration/backup lock probe uses an isolated eight-disk store.
-# Reserve capacity so sibling disk-heavy tests cannot consume its 2s
-# registration budget; retain the queued backup and deadlock assertions.
-[[profile.ci.overrides]]
-filter = 'package(rustfs) & test(=admin::handlers::table_catalog::tests::cold_strong_registration_completes_while_other_bucket_backup_waits)'
-threads-required = "num-test-threads"
-
 [[profile.ci.overrides]]
 filter = 'package(rustfs-heal) & (binary(mrf_partial_write_test) | (binary(mrf_pipeline_test) & test(/^journal_replay_survives_/)) | test(=admin_selector_survives_sigkill_with_original_scope_and_receipts))'
 threads-required = "num-test-threads"
 
-# Native scanner backlog fixtures include 24-set and pool-expansion cases.
-# Reserve capacity across test processes while preserving each scenario's fixed
-# budget, assertions, and internal concurrency.
 [[profile.ci.overrides]]
-filter = 'package(rustfs-scanner) & (test(/^scanner::backlog::tests::native_/) | test(=scanner_folder::tests::mrf_ownership::mrf_ownership_manager_completion_preserves_scanner_pending) | test(/^scanner_io::tests::service_cohort::/))'
+filter = 'package(rustfs-scanner) & (test(/^scanner::backlog::tests::native_(writer|retirement)_/) | test(=scanner_folder::tests::mrf_ownership::mrf_ownership_manager_completion_preserves_scanner_pending) | test(/^scanner_io::tests::service_cohort::/))'
 threads-required = "num-test-threads"
 
 [[profile.ci.scripts]]
 filter = 'package(rustfs-ecstore) & test(/^(bucket::lifecycle::bucket_lifecycle_ops::tests::manual_transition_worker_result_recovery_marks_unknown_for_corrupt_marker|services::rebalance::entry::tests::real_rebalance_run_fence_loss_blocks_multipart_publication|store::init::tests::(batch_transitioned_delete_uses_free_version_per_item|decommission_entry_(allows_free_version_consumed_before_source_lock|rejects_subquorum_free_version_conflict_and_retains_source|skips_cleanup_only_marker_when_free_version_is_present)|dispatched_tier_delete_recovery_(checks_later_pool_then_commits_after_source_removal|finds_directory_source_on_encoded_set|retains_journal_on_source_metadata_error)|force_tier_remove_blocks_on_physical_free_version_hidden_by_other_pool|legacy_unknown_transition_delete_falls_back_for_single_batch_and_blocks_prefix|multi_pool_(recursive_prefix_rejects_legacy_or_hidden_merge_loser_before_delete|same_remote_tuple_(batch|single)_delete_waits_for_all_sources|same_tuple_recursive_prefix_uses_one_journal_owner|transitioned_delete_persists_one_free_version_per_remote_tuple)|recursive_prefix_partial_(pool|set)_failure_keeps_prepared_cleanup_owners|restored_transitioned_delete_uses_free_version_as_cleanup_owner|stable_transitioned_recursive_prefix_delete_uses_journal_owners|suspended_null_transition_delete_uses_free_version_as_sole_owner|tier_mutation_peer_handler_applies_prepare_commit_and_abort_idempotently|transition_response_loss_persists_unknown_outcome_for_provider_recovery|transition_transaction_recovery_(drops_record_after_confirmed_local_commit|keeps_cleanup_pending_local_commit)|transitioned_delete_(free_version_replays_after_store_restart|local_quorum_failure_rolls_back_without_cleanup_owner|uses_free_version_as_cleanup_owner)|versioned_delete_marker_keeps_transitioned_source_and_remote_object|versioned_explicit_transition_delete_preserves_other_version_then_allows_bucket_delete))$/)'
 setup = 'ecstore-large-stack'
 
-# These cross-crate real-storage fixtures need the same bounded stack as ECStore tests.
-[[profile.ci.scripts]]
-filter = '(package(rustfs) & test(/^(admin::handlers::replication::target_repair_tests::[a-z_]+|admin::handlers::site_replication::tests::durable_iam_state_apply_iam_item_rejects_a_stale_recreate_after_a_replicated_delete|admin::service::federated_identity::tests::(binding_keeps_explicit_iam_mapping_strict|markerless_legacy_oidc_session_crosses_auth_and_iam_deletion_boundaries)|storage::scanner_dirty_journal::tests::(clear_callback_keeps_unknown_coverage_until_all_pending_work_is_confirmed|owned_journals_keep_peer_records_after_interleaved_writes_and_clear))$/)) | (package(rustfs-heal) & ((binary(heal_b5_versioned_regression_test) & test(/^serial_tests::(test_heal_unversioned_bucket_e2e|test_historical_null_recursive_heal_restores_physical_versions|test_recursive_deep_heal_reports_stale_delete_marker_repair)$/)) | (binary(usage_observation_read_repair_test) & test(=usage_observation_read_repair_survives_snapshot_cleanup)))) | (package(rustfs-test-utils) & test(=data_usage_snapshot_tests::data_usage_snapshot_recovery_wires_primary_backup_and_negative_cache))'
-setup = 'ecstore-base-stack'
-
 [[profile.ci.scripts]]
 filter = 'package(rustfs-ecstore) | package(rustfs-s3select-api) | package(rustfs-scanner) | (package(rustfs) & test(/^(app::multipart_usecase::tests::concurrent_completions_share_durable_bucket_quota_reservations|app::object::delete::tests::compressed_delete_requests_update_observed_usage_without_releasing_quota_floor|app::object::internal_put::tests::internal_multipart_roundtrip_completes_and_abort_leaves_nothing|app::object::restore::tests::execute_restore_object_maps_failures_to_typed_s3_errors|storage::access::tests::(delete_object_access_captures_authorized_bucket_incarnation|copy_operations_reject_recreated_source_bucket_after_authorization|request_slot_keeps_bucket_policy_bound_to_its_store))$/))'
 setup = 'ecstore-base-stack'
@@ -678,7 +595,6 @@ default-filter = """
         test(/^(delete_marker_migration_semantics|version_id_regression|list_objects_v2_pagination|list_object_versions_regression|list_objects_duplicates|list_buckets_double_slash|list_buckets_auth|list_buckets_iam_filter|leading_slash_key|special_chars|create_bucket_region|delete_objects_versioning|head_object_consistency|head_object_range|copy_object_metadata|copy_object_tagging|copy_source_invalid_date|content_encoding|compression|multipart_storage_class|storage_class_capability|ssec_copy|anonymous_access|bucket_policy_check|presigned_negative|negative_sigv4|admin_auth|notification_webhook|tls_hot_reload|console_smoke|admin_iam_crud|admin_pools|sts_query_compat|data_usage)_test::|^fake_s3_target::/)
         | test(/^replication_extension_test::(test_replication_check_succeeds_with_remote_target|test_replication_check_rejects_target_without_object_lock|test_set_remote_target_rejects_unversioned_source_bucket|test_replication_check_rejects_unversioned_source_bucket|test_replication_check_rejects_missing_replication_config|test_replication_check_rejects_invalid_bucket|test_set_remote_target_rejects_same_bucket_on_same_deployment|test_set_remote_target_rejects_unversioned_target_bucket|test_set_remote_target_update_requires_arn|test_set_remote_target_update_rejects_missing_target|test_set_remote_target_rejects_invalid_target_url|test_set_remote_target_rejects_self_signed_https_target_without_skip_tls_verify|test_set_remote_target_rejects_private_ca_https_target_without_ca_cert_pem|test_list_remote_targets_rejects_empty_bucket|test_list_remote_targets_rejects_invalid_bucket|test_remove_remote_target_rejects_missing_target|test_remove_remote_target_rejects_missing_arn|test_remove_remote_target_rejects_invalid_bucket|test_remove_remote_target_rejects_target_used_by_replication|test_delete_bucket_replication_removes_remote_target)$/)
         | test(/^reliant::lifecycle::/)
-        | test(/^conditional_delete_objects_test::/)
         | test(/^reliant::tiering::/)
         | test(/^upgrade_compatibility_test::upgrade_write_readiness_tests::/)
         | test(/^on_demand_migration::(get_basic_test::(get_miss_pulls_inline_and_serves_locally_afterwards|head_miss_answers_from_the_source_without_persisting)|interaction_test::test_odm_admin_config_is_redacted_and_status_counts_match_the_source)$/)

.docker/observability/grafana/dashboards/rustfs.json

@@ -1748,7 +1748,7 @@
             "uid": "${datasource}"
           },
           "editorMode": "code",
-          "expr": "max by (rustfs_cluster_id, job, bucket) (rustfs_cluster_usage_buckets_objects_count{rustfs_cluster_id=\"$storage_cluster\", job=~\"$job\", bucket=~\"$bucket\"})",
+          "expr": "max by (job, bucket) (rustfs_cluster_usage_buckets_objects_count{job=~\"$job\", bucket=~\"$bucket\"})",
           "legendFormat": "{{bucket}}",
           "range": true,
           "refId": "A"
@@ -1848,7 +1848,7 @@
             "uid": "${datasource}"
           },
           "editorMode": "code",
-          "expr": "max by (rustfs_cluster_id, job, bucket) (rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id=\"$storage_cluster\", job=~\"$job\", bucket=~\"$bucket\"})",
+          "expr": "max by (job, bucket) (rustfs_cluster_usage_buckets_total_bytes{job=~\"$job\", bucket=~\"$bucket\"})",
           "legendFormat": "{{bucket}}",
           "range": true,
           "refId": "A"

.docker/observability/prometheus.yml

@@ -61,7 +61,6 @@ scrape_configs:
 otlp:
   promote_resource_attributes:
     - service.instance.id
-    - process.pid
     - service.name
     - service.namespace
     - cloud.availability_zone

.docker/observability/tests/bucket-usage.test.yml

@@ -1,48 +0,0 @@
-evaluation_interval: 1m
-tests:
-- name: bucket usage keeps clusters separate and deduplicates reporting nodes
-  interval: 1m
-  input_series:
-  - series: rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id="a",job="rustfs",bucket="shared",service_instance_id="a1"}
-    values: '4096 1024'
-  - series: rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id="a",job="rustfs",bucket="shared",service_instance_id="a2"}
-    values: '4096 1024'
-  - series: rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id="b",job="rustfs",bucket="shared",service_instance_id="b1"}
-    values: '65536 65536'
-  - series: rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id="a",job="rustfs",bucket="empty",service_instance_id="a1"}
-    values: '0 0'
-  - series: rustfs_cluster_usage_buckets_objects_count{rustfs_cluster_id="a",job="rustfs",bucket="shared",service_instance_id="a1"}
-    values: '3 1'
-  - series: rustfs_cluster_usage_buckets_objects_count{rustfs_cluster_id="a",job="rustfs",bucket="shared",service_instance_id="a2"}
-    values: '3 1'
-  - series: rustfs_cluster_usage_buckets_objects_count{rustfs_cluster_id="b",job="rustfs",bucket="shared",service_instance_id="b1"}
-    values: '70 70'
-  - series: rustfs_cluster_usage_buckets_objects_count{rustfs_cluster_id="a",job="rustfs",bucket="empty",service_instance_id="a1"}
-    values: '0 0'
-  promql_expr_test:
-  - expr: max by (rustfs_cluster_id, job, bucket) (rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id="a",job=~"rustfs",bucket=~".*"})
-    eval_time: 1m
-    exp_samples:
-    - labels: '{rustfs_cluster_id="a",job="rustfs",bucket="shared"}'
-      value: 1024
-    - labels: '{rustfs_cluster_id="a",job="rustfs",bucket="empty"}'
-      value: 0
-  - expr: max by (rustfs_cluster_id, job, bucket) (rustfs_cluster_usage_buckets_objects_count{rustfs_cluster_id="a",job=~"rustfs",bucket=~".*"})
-    eval_time: 1m
-    exp_samples:
-    - labels: '{rustfs_cluster_id="a",job="rustfs",bucket="shared"}'
-      value: 1
-    - labels: '{rustfs_cluster_id="a",job="rustfs",bucket="empty"}'
-      value: 0
-  - expr: max by (rustfs_cluster_id, job, bucket) (rustfs_cluster_usage_buckets_total_bytes)
-    eval_time: 1m
-    exp_samples:
-    - labels: '{rustfs_cluster_id="a",job="rustfs",bucket="shared"}'
-      value: 1024
-    - labels: '{rustfs_cluster_id="a",job="rustfs",bucket="empty"}'
-      value: 0
-    - labels: '{rustfs_cluster_id="b",job="rustfs",bucket="shared"}'
-      value: 65536
-  - expr: rustfs_cluster_usage_buckets_total_bytes{rustfs_cluster_id="a",bucket="unknown"}
-    eval_time: 1m
-    exp_samples: []

.github/AGENTS.md

@@ -23,11 +23,6 @@ duplicate its rules here.
   pass by fixing the cause, not by weakening the gate).
 - If `shellcheck` is installed, actionlint also lints `run:` scripts; treat
   those findings as part of the gate.
-- Runner placement and billing rules live in `.github/workflows/README.md`.
-  A new job defaults to `ubuntu-latest` (free for this public repository);
-  a self-hosted `runs-on` needs one of the reasons listed there, stated in a
-  comment next to it, and a `rust-cache` key is never shared between the
-  hosted and self-hosted images.
 
 ## CI Alignment
 

.github/actions/schedule-failure-issue/README.md

@@ -74,6 +74,7 @@ Notes:
 - `.github/workflows/e2e-s3tests.yml` (weekly full s3-tests sweep)
 - `.github/workflows/mint.yml` (weekly multi-SDK mint run)
 - `.github/workflows/fuzz.yml` (nightly fuzz corpus jobs)
+- `.github/workflows/performance-ab.yml` (nightly warp A/B gate)
 - **Pending:** the ci-7 e2e nightly-full workflow does not exist yet
   (backlog#1149 ci-7). When it lands, wire it up with the snippet above.
 

.github/actions/setup-docker-compose/action.yml

@@ -1,39 +0,0 @@
-# Copyright 2024 RustFS Team
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-name: Set up Docker Compose
-description: Install the verified Compose CLI used by native Linux x86_64 Connect acceptance
-
-runs:
-  using: composite
-  steps:
-    - name: Install verified Docker Compose CLI plugin
-      shell: bash
-      run: |
-        set -euo pipefail
-        [[ $(uname -s) == Linux && $(uname -m) == x86_64 ]]
-        plugin_dir="${DOCKER_CONFIG:-$HOME/.docker}/cli-plugins"
-        mkdir -p "$plugin_dir"
-        compose_download=$(mktemp "${RUNNER_TEMP:?}/connect-compose.XXXXXX")
-        compose_staged=$(mktemp "$plugin_dir/.docker-compose.XXXXXX")
-        trap 'rm -f "$compose_download" "$compose_staged"' EXIT
-        curl --fail --location --retry 3 \
-          https://github.com/docker/compose/releases/download/v2.40.3/docker-compose-linux-x86_64 \
-          --output "$compose_download"
-        printf '%s  %s\n' \
-          dba9d98e1ba5bfe11d88c99b9bd32fc4a0624a30fafe68eea34d61a3e42fd372 \
-          "$compose_download" | sha256sum --check --strict
-        install -m 0755 "$compose_download" "$compose_staged"
-        mv -f "$compose_staged" "$plugin_dir/docker-compose"
-        [[ $(docker compose version --short) == 2.40.3 ]]

.github/actions/setup/action.yml

@@ -36,16 +36,6 @@ inputs:
     description: "Install cross-compilation tools"
     required: false
     default: "false"
-  zig-version:
-    description: >-
-      Zig version for cargo-zigbuild builds. Pinned so the `<target>.<glibc>`
-      baseline semantics cannot drift with an untested zig release.
-    required: false
-    default: "0.17.0"
-  cargo-zigbuild-version:
-    description: "cargo-zigbuild version installed for cross builds."
-    required: false
-    default: "0.23.4"
   target:
     description: "Target architecture to add"
     required: false
@@ -109,14 +99,10 @@ runs:
     - name: Install Zig
       if: inputs.install-cross-tools == 'true'
       uses: mlugg/setup-zig@d1434d08867e3ee9daa34448df10607b98908d29 # v2
-      with:
-        version: ${{ inputs.zig-version }}
 
     - name: Install cargo-zigbuild
       if: inputs.install-cross-tools == 'true'
       uses: taiki-e/install-action@a21ae4029b089b9ddc45704028756f51ab8abe48 # cargo-zigbuild
-      with:
-        tool: cargo-zigbuild@${{ inputs.cargo-zigbuild-version }}
 
     - name: Install cargo-nextest
       if: inputs.install-test-tools == 'true'
@@ -140,5 +126,5 @@ runs:
         # normalises, so cargo fingerprints stay valid.
         cache-all-crates: false
         cache-on-failure: true
-        shared-key: ${{ inputs.cache-shared-key }}${{ inputs.install-cross-tools == 'true' && format('-zig-{0}-zigbuild-{1}', inputs.zig-version, inputs.cargo-zigbuild-version) || '' }}
+        shared-key: ${{ inputs.cache-shared-key }}
         save-if: ${{ inputs.cache-save-if }}

.github/scheduled-validations.json

@@ -1,6 +1,6 @@
 [
   { "workflow": ".github/workflows/audit.yml", "max_age_hours": 36 },
-  { "workflow": ".github/workflows/build.yml", "max_age_hours": 36 },
+  { "workflow": ".github/workflows/build.yml", "max_age_hours": 192 },
   { "workflow": ".github/workflows/ci.yml", "max_age_hours": 192 },
   { "workflow": ".github/workflows/coverage.yml", "max_age_hours": 192 },
   { "workflow": ".github/workflows/e2e-replication-nightly.yml", "max_age_hours": 36 },
@@ -18,6 +18,7 @@
     "never_ran_grace_until": "2026-09-08T00:00:00Z"
   },
   { "workflow": ".github/workflows/nightly-gnu.yml", "max_age_hours": 36 },
+  { "workflow": ".github/workflows/performance-ab.yml", "max_age_hours": 36 },
   {
     "workflow": ".github/workflows/runner-hygiene.yml",
     "max_age_hours": 792,

.github/workflows/README.md

@@ -1,74 +0,0 @@
-# Workflow runners: what is billed and where a job belongs
-
-**Use this when:** adding a job, changing a `runs-on`, moving a lane between runner pools, or deciding whether a new check is worth its runner cost.
-**Source of truth:** `runs-on` in each workflow file for current placement; `.github/actionlint.yaml` for the declared self-hosted labels; the organisation's Settings → Billing → Usage page for what is actually charged; `docs/testing/ci-gates.md` for which checks block a merge (that does not change when a runner changes).
-
-## Billing facts
-
-This repository is public. For a public repository, GitHub's standard hosted runners (`ubuntu-latest`, `windows-latest`, `macos-latest`) cost nothing and have no minute cap, whatever plan the organisation is on. GitHub bills only two things: larger hosted runners, which have custom labels and are not used here, and standard runners in private repositories.
-
-Every other label in this repository is capacity RustFS pays for. ARC pods and lab machines accrue cost while the machine exists. A Ubicloud label is metered for the minutes a job holds its VM:
-
-| Label | Paid by | What it is | Use it for |
-|---|---|---|---|
-| `ubuntu-latest` | nobody | GitHub-hosted, Ubuntu 24.04, 4 vCPU, 16 GB RAM, about 14 GB of free disk. Ships a C toolchain, `docker`, `gh`, `curl`, `wget`, `jq`, Python. Can `mount` tmpfs and loop devices; io_uring works | the default for every job |
-| `windows-latest`, `macos-latest` | nobody | GitHub-hosted standard runners | Windows and macOS targets |
-| `sm-standard-2` | RustFS | ARC pod, Ubuntu 24.04, minimal image: no `gh`, no C toolchain; Docker only on the `dind-` label | only a job that a hosted runner cannot run (see the placement rule) |
-| `sm-standard-4` | RustFS | ARC pod, Ubuntu 24.04, 14 CPUs and 28 GB (the label is not the core count), large disk | workspace-scale Cargo builds |
-| `ubicloud-standard-16-arm` | RustFS | Ubicloud VM, alias of `ubicloud-standard-16-arm-ubuntu-2404`: Ubuntu 24.04 arm64, 16 vCPU, 48 GB, 300 GB disk. Reached through the `build.yml` matrix, so `actionlint.yaml` does not list it | native aarch64 release builds (`linux-aarch64-musl`, `linux-aarch64-gnu`) |
-| `dind-sm-standard-2` | RustFS | ARC pod with Docker-in-Docker | image builds that need `buildx` and a registry push |
-| `smoke-testing`, `pf-testing` | RustFS | long-lived lab machines | the packaged functional chain and performance suites |
-
-The self-hosted pods and the hosted image are both Ubuntu 24.04, so a binary built on one runs on the other; `ci.yml` builds the debug binary on `sm-standard-4` and runs it on both pools.
-
-## Placement rule
-
-A job runs on `ubuntu-latest` unless it has one of these reasons, and the reason goes in a comment next to its `runs-on`:
-
-1. **Disk.** It restores or produces the workspace `target/`: `cargo build` or `cargo test` with `--all-targets`, `cargo nextest run --all`, or any reader of the `ci-dev` / `ci-feat-*` caches. The workspace test build leaves about 64 GB in `target/`; a hosted runner has about 14 GB.
-2. **Critical-path CPU.** It is a build that several downstream lanes wait on and would blow its timeout at 4 vCPU (`build-rustfs-debug-binary` gates the e2e and s3-tests lanes).
-3. **Hardware.** Native aarch64, Docker-in-Docker, or the lab machines.
-4. **Cache ownership.** It reads a `rust-cache` key that a self-hosted job writes (see the cache rule below).
-
-Jobs that do only the following have no reason to be self-hosted: shell or Python scripts, `cargo fmt`, `cargo deny`, typos, actionlint, API calls (`gh`, dependency review, issue and stale bots, CLA checks), downloading an artifact and running a prebuilt binary, or a `cargo build -p rustfs` with its own cache key. `e2e-distributed.yml` and `e2e-upgrade.yml` build the server and the `e2e_test` harness on `ubuntu-latest` inside their budgets, so "it compiles something" alone is not a reason.
-
-A hosted runner is also the only choice for a few things the pods cannot do: `mount -o loop` and `mount -t tmpfs` fail inside the ARC pods, and a container runtime cannot guarantee that `io_uring_setup` is allowed. `uring-integration` and `e2e-distributed` are on `ubuntu-latest` for that reason, not for cost.
-
-## Cache rule
-
-`Swatinem/rust-cache` builds its key from the OS, architecture, toolchain and lockfile. It does not include the runner label or image, so a cache written on an `sm-standard-4` pod and one written on `ubuntu-latest` would overwrite each other under the same `shared-key`.
-
-- Every `shared-key` has exactly one writer. `cache-warm.yml` owns `ci-dev`, `ci-feat-rio`, `ci-feat-proto` and `ci-uring`; every other `./.github/actions/setup` call states `cache-save-if` explicitly, and `scripts/security/check_cache_save_if.sh` fails the build when one does not.
-- All readers of a key run on the same image as its writer. Keys that live on hosted runners say so in their name (`ci-uring` is warmed on `ubuntu-latest`; `ci-e2e-distributed-hosted`, `e2e-upgrade-server-hosted`, `e2e-upgrade-tests-hosted`).
-- Moving a job between pools therefore means renaming its key. The new key is empty until a run from `main` (a scheduled run, or a manual dispatch of that workflow from `main`) saves it; until then that lane does a cold build. The `build.yml` aarch64 release legs append `-ubicloud-arm` for this reason. `cache-warm.yml` does not write those keys.
-- `CARGO_BUILD_JOBS` and every other `CARGO*`, `CC*`, `CFLAGS*`, `CXX*`, `CMAKE*` and `RUST*` variable is hashed into the key. `ci.yml` sets `CARGO_BUILD_JOBS` at step level for that reason; promoting it to job level rotates the key for every lane that shares it.
-
-## Where self-hosted minutes leak
-
-These are the patterns that have cost real money here without validating anything more:
-
-- An `if: always()` summary or alert job on a self-hosted label inside a `workflow_run`-triggered workflow. The upstream workflow completes many times a day, the real jobs skip, and the summary still starts a pod each time. Gate it on the upstream job having run, and keep it on `ubuntu-latest`.
-- Per-push workflows on `main`. A merge triggers the full `ci.yml` matrix, `cache-warm.yml` and the `build.yml` development build; with tens of merges a day that is as much self-hosted time as all pull requests together. Anything that does not verify the merged tree itself belongs on a schedule or behind a `paths` filter.
-- `apt-get install` of tools the hosted image already ships. It is only needed on the minimal pod images.
-- A throttled `CARGO_BUILD_JOBS` on a 14-CPU pod. The limit exists for link-storm reasons documented in `ci.yml`; measure before raising it, and measure before lowering it further.
-- Tiny jobs on ARC pods. Each one pays roughly a minute of pod start-up before its first step; a compile-free job that takes ten seconds costs more to schedule than to run.
-
-## Measuring a run
-
-To see which pool each job of a run used and for how long:
-
-```bash
-gh api "repos/rustfs/rustfs/actions/runs/<run_id>/jobs?per_page=100" \
-  --jq '.jobs[] | "\(.conclusion)\t\(.labels | join(","))\t\(.started_at)\t\(.completed_at)\t\(.name)"'
-```
-
-Only rows whose label starts with `sm-`, `dind-`, `ubicloud-`, `smoke-` or `pf-` cost anything. At the October 2026 review, a full pull-request run of `ci.yml` spent about 150 self-hosted minutes, a third of it in jobs that never compiled; treat that as the scale of what a placement mistake costs, not as a current figure.
-
-## Checklist for a new or moved job
-
-- `runs-on` is a literal label. A new self-hosted label is declared in `.github/actionlint.yaml` in the same change (`.github/AGENTS.md`).
-- The job declares `timeout-minutes` (`scripts/security/check_job_timeouts.sh`).
-- `actionlint` reports nothing.
-- A self-hosted `runs-on` carries a comment naming its reason from the placement rule.
-- Every `./.github/actions/setup` call states `cache-save-if`, and a key read from one pool is not written from the other.
-- Moving a job does not change its check name, so the `main` ruleset and `docs/testing/ci-gates.md` stay as they are; renaming it does, and then that document's change checklist applies.

.github/workflows/audit.yml

@@ -29,6 +29,7 @@ on:
       - 'scripts/test_package_service_scripts.sh'
       - 'scripts/test_package_versions.sh'
       - 'scripts/test_docker_workflow.py'
+      - 'scripts/security/check_performance_ab_workflow.sh'
       - 'scripts/security/check_preview_release_workflow.sh'
       - 'scripts/security/check_tier_artifact_workflow.sh'
       - 'scripts/security/check_workflow_pins.sh'
@@ -47,6 +48,7 @@ on:
       - 'scripts/test_package_service_scripts.sh'
       - 'scripts/test_package_versions.sh'
       - 'scripts/test_docker_workflow.py'
+      - 'scripts/security/check_performance_ab_workflow.sh'
       - 'scripts/security/check_preview_release_workflow.sh'
       - 'scripts/security/check_tier_artifact_workflow.sh'
       - 'scripts/security/check_workflow_pins.sh'
@@ -75,7 +77,7 @@ jobs:
   cancel-closed-pr-runs:
     name: Cancel Closed PR Runs
     if: github.event_name == 'pull_request' && github.event.action == 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Explain cancellation run
@@ -87,7 +89,7 @@ jobs:
   cargo-deny:
     name: Cargo Deny
     if: github.repository == 'rustfs/rustfs' && (github.event_name != 'pull_request' || github.event.action != 'closed')
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 20
     steps:
       - name: Checkout repository
@@ -130,7 +132,7 @@ jobs:
   workflow-pin-report:
     name: Workflow Pin Report
     if: github.repository == 'rustfs/rustfs' && (github.event_name != 'pull_request' || github.event.action != 'closed')
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 5
     steps:
       - name: Checkout repository
@@ -153,6 +155,9 @@ jobs:
       - name: Check preview release workflow policy
         run: ./scripts/security/check_preview_release_workflow.sh
 
+      - name: Check performance A/B workflow trust boundary
+        run: ./scripts/security/check_performance_ab_workflow.sh
+
       - name: Check tier evidence workflow isolation
         run: ./scripts/security/check_tier_artifact_workflow.sh
 
@@ -164,7 +169,7 @@ jobs:
 
   dependency-review:
     name: Dependency Review
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     if: github.event_name == 'pull_request' && github.event.action != 'closed'
     permissions:

.github/workflows/btf-prototype-reference.yml

@@ -1,70 +0,0 @@
-name: BTF prototype reference
-
-on:
-  pull_request:
-    paths:
-      - "scripts/inspect_btf_prototypes.py"
-      - "scripts/test_inspect_btf_prototypes.py"
-      - "scripts/fixtures/btf_prototypes.c"
-      - ".github/workflows/btf-prototype-reference.yml"
-  workflow_dispatch:
-
-permissions:
-  contents: read
-
-concurrency:
-  group: btf-prototype-reference-${{ github.ref }}
-  cancel-in-progress: true
-
-jobs:
-  reference:
-    runs-on: ubuntu-latest
-    timeout-minutes: 10
-    steps:
-      - uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
-        with:
-          persist-credentials: false
-      - uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.12"
-      - name: Check prototype and CLI contracts
-        run: python3 -B scripts/test_inspect_btf_prototypes.py
-      - name: Install only missing native reference tools
-        run: |
-          select_bpftool() {
-            local candidate
-            local candidates=()
-            if command -v bpftool >/dev/null; then
-              candidates+=("$(command -v bpftool)")
-            fi
-            for candidate in /usr/lib/linux-tools-*/bpftool /usr/lib/linux-tools/*/bpftool; do
-              if [[ -x "$candidate" ]]; then
-                candidates+=("$candidate")
-              fi
-            done
-            while IFS= read -r candidate; do
-              if [[ -x "$candidate" ]] && file -Lb "$candidate" | grep -q 'ELF' \
-                && "$candidate" version >/dev/null 2>&1; then
-                printf '%s\n' "$candidate"
-                return 0
-              fi
-            done < <(printf '%s\n' "${candidates[@]}" | sort -Vr)
-            return 1
-          }
-          packages=()
-          command -v clang >/dev/null || packages+=(clang)
-          command -v llvm-objcopy >/dev/null || packages+=(llvm)
-          select_bpftool >/dev/null || packages+=(linux-tools-generic)
-          if (( ${#packages[@]} )); then
-            sudo apt-get update
-            sudo apt-get install --no-install-recommends -y "${packages[@]}"
-          fi
-          # File parsing needs a working binary, not a host-kernel wrapper.
-          selected_bpftool="$(select_bpftool)"
-          "$selected_bpftool" version
-          printf 'BPFTOOL=%s\n' "$selected_bpftool" >> "$GITHUB_ENV"
-      - name: Compare compiled declarations with bpftool
-        run: |
-          clang -target bpf -O2 -g -c scripts/fixtures/btf_prototypes.c -o "$RUNNER_TEMP/prototypes.o"
-          llvm-objcopy --dump-section ".BTF=$RUNNER_TEMP/prototypes.btf" "$RUNNER_TEMP/prototypes.o"
-          python3 -B scripts/test_inspect_btf_prototypes.py --native-reference "$RUNNER_TEMP/prototypes.btf"

.github/workflows/build.yml

@@ -22,12 +22,9 @@
 # 3. Trigger docker.yml to build and push images using the uploaded binaries
 #
 # Platform scope:
-# - Development builds come from the nightly schedule and manual dispatch and
-#   cover the full platform matrix; they feed the dev download channel
-#   (artifacts/rustfs/dev on R2, the *-main-latest packages). Tags build the
-#   release matrix. Main pushes do not build: at the usual merge rate most
-#   per-merge builds were cancelled by the next merge before publishing
-#   anything, and the dev channel does not need per-merge granularity.
+# - Pushes to main (development builds) build the Linux targets only — they
+#   feed the dev download channel and Docker dev images. Tags, the weekly
+#   schedule and manual dispatch build the full platform matrix.
 #
 # Manual Parameters:
 # - publish: Publish binaries/releases for manual runs (default: false)
@@ -39,8 +36,24 @@ name: Build and Release
 on:
   push:
     tags: [ "*.*.*" ]
+    branches: [ main ]
+    paths-ignore:
+      - "**.md"
+      - "**.txt"
+      - ".github/**"
+      - "docs/**"
+      - "deploy/**"
+      - "scripts/dev_*.sh"
+      - "LICENSE*"
+      - "README*"
+      - "**/*.png"
+      - "**/*.jpg"
+      - "**/*.svg"
+      - ".gitignore"
+      - ".dockerignore"
+      - "flake.lock"
   schedule:
-    - cron: "13 1 * * *" # Nightly 01:13 UTC
+    - cron: "13 1 * * 0" # Weekly on Sunday 01:13 UTC
   workflow_dispatch:
     inputs:
       publish:
@@ -80,7 +93,7 @@ jobs:
   # Build strategy check - determine build type based on trigger
   build-check:
     name: Build Strategy Check
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     outputs:
       should_build: ${{ steps.check.outputs.should_build }}
@@ -131,6 +144,12 @@ jobs:
               build_type="release"
               echo "📦 Release build detected: $tag_name"
             fi
+          elif [[ "${{ github.ref }}" == "refs/heads/main" ]]; then
+            # Main branch push - development build
+            should_build=true
+            build_type="development"
+            version="dev-${short_sha}"
+            echo "🛠️  Development build detected"
           elif [[ "${{ github.event_name }}" == "schedule" ]] || \
                [[ "${{ github.event_name }}" == "workflow_dispatch" ]]; then
             # Scheduled or manual build
@@ -159,7 +178,7 @@ jobs:
   prepare-platform-matrix:
     name: Prepare Platform Matrix
     needs: build-check
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     outputs:
       matrix: ${{ steps.select.outputs.matrix }}
@@ -181,20 +200,25 @@ jobs:
             selected="all"
           fi
 
-          # linux-aarch64-* need a real aarch64 host (placement rule: hardware)
-          # so the packaged binary and console smoke execute the artifact.
-          # ubicloud-standard-16-arm is that host: Ubuntu 24.04 arm64, alias of
-          # ubicloud-standard-16-arm-ubuntu-2404 (16 vCPU, 48 GB). linux-x86_64-* stay on the
-          # ARC sm-standard-4 pool. GNU rows keep cross=true so cargo-zigbuild
-          # pins the glibc 2.31 floor on the same architecture; musl stays a
-          # native cargo build.
+          # Per-merge development builds only feed the dev download channel
+          # and the Docker dev images, which consume the Linux binaries; at
+          # the usual merge cadence most per-merge builds are cancelled by the
+          # next merge anyway. macOS and Windows stay covered by tag builds,
+          # the weekly scheduled build and manual dispatch.
+          if [[ "${selected}" == "all" \
+            && "${{ github.event_name }}" == "push" \
+            && "${{ needs.build-check.outputs.build_type }}" == "development" ]]; then
+            selected="linux-x86_64-musl,linux-aarch64-musl,linux-x86_64-gnu,linux-aarch64-gnu"
+            echo "Development (main push) build: restricting to Linux targets"
+          fi
+
           all='{"include":[
-            {"target_id":"linux-x86_64-musl","os":"sm-standard-4","target":"x86_64-unknown-linux-musl","cross":false,"platform":"linux","rustflags":"","features":"","glibc":""},
-            {"target_id":"linux-aarch64-musl","os":"ubicloud-standard-16-arm","target":"aarch64-unknown-linux-musl","cross":false,"platform":"linux","rustflags":"","features":"","glibc":""},
-            {"target_id":"linux-x86_64-gnu","os":"sm-standard-4","target":"x86_64-unknown-linux-gnu","cross":true,"platform":"linux","rustflags":"","features":"pyroscope","glibc":"2.31"},
-            {"target_id":"linux-aarch64-gnu","os":"ubicloud-standard-16-arm","target":"aarch64-unknown-linux-gnu","cross":true,"platform":"linux","rustflags":"","features":"pyroscope","glibc":"2.31"},
-            {"target_id":"macos-aarch64","os":"macos-latest","target":"aarch64-apple-darwin","cross":false,"platform":"macos","rustflags":"","features":"pyroscope","glibc":""},
-            {"target_id":"windows-x86_64","os":"windows-latest","target":"x86_64-pc-windows-msvc","cross":false,"platform":"windows","rustflags":"","features":"","glibc":""}
+            {"target_id":"linux-x86_64-musl","os":"sm-standard-4","target":"x86_64-unknown-linux-musl","cross":false,"platform":"linux","rustflags":"","features":""},
+            {"target_id":"linux-aarch64-musl","os":"sm-standard-4-arm","target":"aarch64-unknown-linux-musl","cross":false,"platform":"linux","rustflags":"","features":""},
+            {"target_id":"linux-x86_64-gnu","os":"sm-standard-4","target":"x86_64-unknown-linux-gnu","cross":false,"platform":"linux","rustflags":"","features":"pyroscope"},
+            {"target_id":"linux-aarch64-gnu","os":"sm-standard-4-arm","target":"aarch64-unknown-linux-gnu","cross":false,"platform":"linux","rustflags":"","features":"pyroscope"},
+            {"target_id":"macos-aarch64","os":"macos-latest","target":"aarch64-apple-darwin","cross":false,"platform":"macos","rustflags":"","features":"pyroscope"},
+            {"target_id":"windows-x86_64","os":"windows-latest","target":"x86_64-pc-windows-msvc","cross":false,"platform":"windows","rustflags":"","features":""}
           ]}'
 
           if [[ "${selected}" == "all" ]]; then
@@ -222,8 +246,7 @@ jobs:
           echo "Selected platforms: ${selected}"
 
   build-rustfs:
-    # Acceptance workflows select artifacts by this stable platform name.
-    name: Build RustFS (${{ matrix.target_id }})
+    name: Build RustFS
     needs: [ build-check, prepare-platform-matrix ]
     if: needs.build-check.outputs.should_build == 'true' && needs.prepare-platform-matrix.result == 'success'
     runs-on: ${{ matrix.os }}
@@ -250,11 +273,7 @@ jobs:
         with:
           rust-version: stable
           target: ${{ matrix.target }}
-          # rust-cache keys include OS and architecture but not the runner image.
-          # The Ubicloud aarch64 legs therefore use their own suffix so they
-          # cannot restore a cache written on a different arm64 image.
-          # x86_64, macOS, and Windows keys stay unsuffixed.
-          cache-shared-key: build-${{ matrix.target }}${{ matrix.glibc && format('-glibc-{0}', matrix.glibc) || '' }}${{ matrix.os == 'ubicloud-standard-16-arm' && '-ubicloud-arm' || '' }}
+          cache-shared-key: build-${{ matrix.target }}
           # main only. A cache saved on refs/tags/X is scoped to that tag: no
           # other tag, no main run and no PR can restore it, so every release
           # cycle wrote up to 12 entries of 1-2GB (preview tag plus final tag,
@@ -391,53 +410,35 @@ jobs:
         shell: bash
         env:
           BUILD_TARGET: ${{ matrix.target }}
-          BUILD_OS: ${{ matrix.platform }}-${{ startsWith(matrix.target, 'aarch64-') && 'aarch64' || 'x86_64' }}
         run: |
           # Force rebuild by touching build.rs
           touch rustfs/build.rs
 
+          BINARY_NAMES=(rustfs)
+          if [[ "${{ matrix.platform }}" == "linux" ]]; then
+            BINARY_NAMES+=(rustfs-cli)
+          fi
           FEATURE_ARGS=()
           if [[ -n "${{ matrix.features }}" ]]; then
             FEATURE_ARGS=(--features "${{ matrix.features }}")
           fi
 
-          # gnu targets pin an explicit glibc baseline (`<target>.<glibc>` for
-          # cargo-zigbuild) so release binaries link no symbols newer than it
-          # and run on older distros such as RHEL/Rocky 9 (glibc 2.34); #8062.
-          TARGET_ARG="$BUILD_TARGET"
-          if [[ -n "${{ matrix.glibc }}" ]]; then
-            TARGET_ARG="${TARGET_ARG}.${{ matrix.glibc }}"
-          fi
-
-          PROFILE_ARGS=()
-          if [[ "$BUILD_TARGET" == *-linux-gnu && "${{ matrix.features }}" == "pyroscope" ]]; then
-            # The in-process profiler resolves names from the final executable.
-            # Retain function symbols without enabling DWARF source information.
-            PROFILE_ARGS=(--config 'profile.release.strip="none"')
-          fi
-          if [[ "${{ matrix.cross }}" == "true" ]]; then
-            cargo zigbuild --release --locked --target "$TARGET_ARG" "${FEATURE_ARGS[@]}" "${PROFILE_ARGS[@]}" -p rustfs --bin rustfs
-          else
-            cargo build --release --locked --target "$TARGET_ARG" "${FEATURE_ARGS[@]}" "${PROFILE_ARGS[@]}" -p rustfs --bin rustfs
-          fi
-
-          # All release targets use same-architecture runners, including GNU
-          # builds that use Zig only to select the glibc ABI floor.
-          bin="target/${{ matrix.target }}/release/rustfs"
-          if [[ -f "${bin}.exe" ]]; then
-            bin="${bin}.exe"
-          fi
-          version_output=$("$bin" --version)
-          if ! grep -Fxq "build target : ${BUILD_TARGET}" <<< "$version_output" || \
-             ! grep -Fxq "build os     : ${BUILD_OS}" <<< "$version_output" || \
-             ! grep -Fxq "rust channel : stable-${BUILD_TARGET}" <<< "$version_output"; then
-            echo "release binary reports an unexpected compile target" >&2
-            exit 1
-          fi
-          if [[ -n "${{ matrix.glibc }}" ]]; then
-            python3 scripts/release/check_glibc.py "${{ matrix.glibc }}" \
-              "target/${{ matrix.target }}/release/rustfs"
-          fi
+          # Build one binary at a time so release LTO links cannot exhaust a
+          # self-hosted runner by running concurrently.
+          for binary in "${BINARY_NAMES[@]}"; do
+            PROFILE_ARGS=()
+            if [[ "$binary" == "rustfs" && "$BUILD_TARGET" == *-linux-gnu && "${{ matrix.features }}" == "pyroscope" ]]; then
+              # The in-process profiler resolves names from the final executable.
+              # Retain function symbols without enabling DWARF source information.
+              PROFILE_ARGS=(--config 'profile.release.strip="none"')
+            fi
+            if [[ "${{ matrix.cross }}" == "true" ]]; then
+              # All cross targets in the matrix are Linux; zigbuild handles them.
+              cargo zigbuild --release --target ${{ matrix.target }} "${FEATURE_ARGS[@]}" "${PROFILE_ARGS[@]}" -p rustfs --bin "$binary"
+            else
+              cargo build --release --target ${{ matrix.target }} "${FEATURE_ARGS[@]}" "${PROFILE_ARGS[@]}" -p rustfs --bin "$binary"
+            fi
+          done
 
       - name: Verify macOS binary has no dynamic liblzma dependency
         if: matrix.platform == 'macos'
@@ -553,6 +554,9 @@ jobs:
           fi
 
           BINARY_FILES=("$BINARY_NAME")
+          if [[ "${{ matrix.platform }}" == "linux" ]]; then
+            BINARY_FILES+=("rustfs-cli")
+          fi
 
           if [[ "$TARGET" == *-linux-gnu && "${{ matrix.features }}" == "pyroscope" ]]; then
             BINARY_FILES+=(rustfs.cpu-symbol-catalog.json rustfs.cpu-symbol-catalog.sha256)
@@ -625,7 +629,7 @@ jobs:
               LATEST_FILES="$LATEST_FILE"
             fi
           elif [[ "$BUILD_TYPE" == "development" ]]; then
-            # Development builds (nightly schedule and manual dispatch)
+            # Development builds (only main branch triggers development builds)
             # Create main-latest version filename
             # Convert from rustfs-linux-x86_64-dev-abc123 to rustfs-linux-x86_64-main-latest
             MAIN_LATEST_FILE="${PACKAGE_BASENAME}-main-latest.zip"
@@ -667,7 +671,7 @@ jobs:
           echo "🔧 Build type: ${BUILD_TYPE}"
           echo "📊 Version: ${VERSION}"
 
-      - name: Verify packaged Linux binary
+      - name: Verify packaged Linux binaries
         if: matrix.platform == 'linux'
         shell: bash
         run: |
@@ -675,7 +679,7 @@ jobs:
 
           package_file="${{ steps.package.outputs.package_file }}"
           archive_members="$(unzip -Z1 "$package_file" | sort)"
-          expected_members='rustfs'
+          expected_members=$'rustfs\nrustfs-cli'
           target="${{ matrix.target }}"
           features="${{ matrix.features }}"
           if [[ "$target" == *-linux-gnu && "$features" == "pyroscope" ]]; then
@@ -687,14 +691,13 @@ jobs:
             exit 1
           fi
 
-          # musl targets build natively; the gnu legs are same-arch
-          # cargo-zigbuild builds pinned to the release glibc baseline. Either
-          # way the binaries run on their runner, so executing them here
+          # Every Linux target builds natively, so executing the binaries here
           # proves each artifact matches its runner architecture.
-          binary_dir="$(mktemp -d)"
-          trap 'rm -rf "$binary_dir"' EXIT
-          unzip -q "$package_file" rustfs -d "$binary_dir"
-          "$binary_dir/rustfs" --version
+          cli_dir="$(mktemp -d)"
+          trap 'rm -rf "$cli_dir"' EXIT
+          unzip -q "$package_file" rustfs rustfs-cli -d "$cli_dir"
+          "$cli_dir/rustfs" --version
+          "$cli_dir/rustfs-cli" --help >/dev/null
 
       - name: Verify packaged console
         if: matrix.target == 'x86_64-unknown-linux-gnu' || matrix.target == 'aarch64-unknown-linux-gnu'
@@ -774,9 +777,9 @@ jobs:
             exit 0
           fi
 
-          # ARC pods (the x86_64 legs) do not ship the aws CLI. Ubicloud arm64
-          # images follow the GitHub runner image and often do. Install the
-          # official CLI bundle when it is missing so R2 uploads work on either.
+          # The self-hosted Linux runners do not ship the aws CLI (GitHub-hosted
+          # images did). Install the official CLI bundle on demand so R2 uploads
+          # survive a fresh runner without relying on a distro package.
           if ! command -v aws >/dev/null 2>&1; then
             echo "aws CLI not found on runner; installing..."
             if command -v apt-get >/dev/null 2>&1; then
@@ -840,7 +843,7 @@ jobs:
     name: Build Summary
     needs: [ build-check, build-rustfs ]
     if: always() && needs.build-check.outputs.should_build == 'true'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Build completion summary
@@ -1029,7 +1032,7 @@ jobs:
     name: Update Latest Version
     needs: [ build-check, publish-release ]
     if: needs.build-check.outputs.should_publish == 'true' && startsWith(github.ref, 'refs/tags/') && (needs.build-check.outputs.build_type == 'release' || needs.build-check.outputs.build_type == 'prerelease')
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     steps:
       - name: Update latest.json

.github/workflows/cache-warm.yml

@@ -42,23 +42,26 @@
 name: Cache Warm
 
 on:
-  schedule:
-    # Every six hours; the 22:41 UTC run lands before ci.yml's nightly 00:11
-    # UTC matrix. A reader restores the newest entry under its key and
-    # rebuilds whatever its own diff changed, so a cache up to six hours old
-    # costs a PR the handful of crates merged since. Warming after every merge
-    # (fourteen completed runs a day, about two sm-standard-4 hours each)
-    # spent far more than that to avoid it.
-    - cron: "41 4,10,16,22 * * *"
   push:
     branches: [ main ]
-    # A toolchain or lockfile change invalidates most of every key; those
-    # re-warm at once instead of waiting for the next slot.
-    paths:
-      - "Cargo.lock"
-      - "rust-toolchain.toml"
-      - ".github/actions/setup/**"
-      - ".github/workflows/cache-warm.yml"
+    # Mirrors ci.yml's push paths-ignore: if a commit cannot change what ci.yml
+    # compiles, it cannot change what ci.yml needs restored either.
+    paths-ignore:
+      - "**.md"
+      - "docs/**"
+      - "deploy/**"
+      - "scripts/dev_*.sh"
+      - "scripts/probe.sh"
+      - "LICENSE*"
+      - ".gitignore"
+      - ".dockerignore"
+      - "README*"
+      - "**/*.png"
+      - "**/*.jpg"
+      - "**/*.svg"
+      - ".github/workflows/build.yml"
+      - ".github/workflows/docker.yml"
+      - ".github/workflows/audit.yml"
   workflow_dispatch:
     inputs:
       emit_timings:
@@ -72,7 +75,7 @@ on:
 permissions:
   contents: read
 
-# Scoped by event. Scheduled, push and dispatch runs do not compete: GitHub keeps
+# Scoped by event. A push run and a dispatch run do not compete: GitHub keeps
 # one running plus one pending per group, so with a single shared group a
 # manually dispatched run was displaced as pending by the next merge and
 # cancelled — observed three times in a row, which made the --timings gate in
@@ -96,7 +99,7 @@ env:
   RUST_BACKTRACE: 1
 
 jobs:
-  # Readers: test-and-lint (workspace and ILM), build-rustfs-debug-binary,
+  # Readers: test-and-lint, test-ilm-integration-serial, build-rustfs-debug-binary,
   # e2e-tests, e2e-full.
   warm-ci-dev:
     name: Warm ci-dev
@@ -159,17 +162,6 @@ jobs:
           cargo build --workspace --all-targets
           cargo build -p rustfs --bins --features e2e-test-hooks
 
-      # ci.yml runs Clippy on every pull request. Clippy's check-mode artifacts
-      # are separate from the build above, so warm them with the exact
-      # invocation ci.yml uses; a pull request then re-checks only what its diff
-      # touches. A lint failure on main must not stop the warm-up: the cache is
-      # still saved (cache-on-failure), and the failure belongs to ci.yml.
-      - name: Warm Clippy check artifacts
-        run: |
-          if ! cargo clippy --all-targets -- -D warnings; then
-            echo "::warning::cargo clippy reports findings on this commit; check artifacts are still cached"
-          fi
-
       # Runs before rust-cache's post step, so these are the sizes it is about
       # to archive. Reported so the cache-all-crates decision stays evidence-led:
       # registry/src is what that flag prunes, registry/cache is what the pruned
@@ -193,10 +185,9 @@ jobs:
             echo '```'
           } >> "$GITHUB_STEP_SUMMARY"
 
-  # Readers: test-and-lint-rio-v2 (nightly schedule, merge group, dispatch),
-  # build-rustfs-debug-binary-rio-v2 (nightly schedule / manual dispatch only —
-  # dormant rio-v2 variant, see rustfs/backlog#1835 and
-  # docs/architecture/minio-file-format-compat.md).
+  # Readers: test-and-lint-rio-v2 (per-PR), build-rustfs-debug-binary-rio-v2
+  # (weekly schedule / manual dispatch only — dormant rio-v2 variant, see
+  # rustfs/backlog#1835 and docs/architecture/minio-file-format-compat.md).
   # The second build below stays despite the reduced cadence: it warms the
   # rio-v2,e2e-test-hooks feature resolution the scheduled build restores,
   # which keeps that lane inside its 30-minute timeout.

.github/workflows/ci.yml

@@ -42,7 +42,7 @@ on:
   merge_group:
     types: [ checks_requested ]
   schedule:
-    - cron: "11 0 * * *" # Nightly 00:11 UTC, after the 22:41 UTC cache-warm run
+    - cron: "11 0 * * 0" # Weekly on Sunday 00:11 UTC
   workflow_dispatch:
 
 permissions:
@@ -62,15 +62,10 @@ env:
 
 jobs:
 
-  # Compile-free lanes, and lanes that only run a prebuilt binary, use
-  # GitHub-hosted runners: they are free for this public repository, and the
-  # self-hosted pool is reserved for the lanes that need its CPU count and
-  # disk (the workspace test build leaves a ~64GB target/, hosted runners
-  # have 14GB). Placement and billing rules: README.md in this directory.
   cancel-closed-pr-runs:
     name: Cancel Closed PR Runs
     if: github.event_name == 'pull_request' && github.event.action == 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Explain cancellation run
@@ -79,7 +74,7 @@ jobs:
   classify-changes:
     name: Select CI scope
     if: github.event_name != 'pull_request' || github.event.action != 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     outputs:
       mode: ${{ steps.scope.outputs.mode }}
@@ -105,12 +100,16 @@ jobs:
   typos:
     name: Typos
     if: github.event_name != 'pull_request' || github.event.action != 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
         with:
           persist-credentials: false
+      - name: Install typos download dependency
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y --no-install-recommends wget
       - name: Typos check with custom config file
         uses: crate-ci/typos@512fc24f32f44ab01972217aaaf3dc86ec234d53 # v1.50.2
 
@@ -118,7 +117,7 @@ jobs:
   quick-checks:
     name: Quick Checks
     if: github.event_name != 'pull_request' || github.event.action != 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Checkout repository
@@ -169,12 +168,12 @@ jobs:
             echo "started_at=$(date --utc --iso-8601=seconds)"
           } > artifacts/test-and-lint/run-metadata.txt
 
-      # Clippy gates every event, pull requests included. When it ran only after
-      # merge, a lint regression merged freely and then failed every main run
-      # until someone fixed it (10 of 12 main runs on 2026-10-17/18). cache-warm
-      # writes the same invocation's check artifacts into ci-dev, so a pull
-      # request re-checks only the crates its diff touches. Main keeps the step
-      # because pull requests can merge without being up to date with main.
+      # Clippy runs before the test pass: lint failures are the most common
+      # CI-only breakage and should surface in minutes, not after 20+ minutes
+      # of tests.
+      # Sampled too: clippy is the natural control arm for any CARGO_BUILD_JOBS
+      # experiment, since --all-targets is check-only for workspace members and
+      # never links the ~100 test binaries the limit exists to throttle.
       - name: Run clippy lints
         run: |
           ./scripts/ci/resource_sampler.sh start clippy
@@ -183,27 +182,31 @@ jobs:
 
       - name: Run nextest tests
         env:
-          # #5394 mitigation, settled by the backlog#1601 experiment.
+          # #5394 mitigation, now under a measured experiment (backlog#1601).
           #
           # 2 was chosen when three concurrent workspace test links were believed
           # to saturate the runner's overlay I/O and wedge Cargo until the 75m
           # timeout. cgroup v2 readings from the sampler show the pod actually
-          # has 14 CPUs and 28GB (peak use 2.1GB): the label name "sm-standard-4"
-          # had led everyone, including the original mitigation, to assume 4
-          # cores, and memory was never the constraint.
+          # has 14 CPUs and 28GB (peak use 2.1GB), so 2 throttles compilation to
+          # a seventh of what is available and memory was never the constraint —
+          # the label name "sm-standard-4" had led everyone, including the
+          # original mitigation, to assume 4 cores.
           #
-          # Measured on the same ci-dev cache, the test-profile compile took
-          # 28m21s at 2 on a pull request and 16-23 min at 3 across three
-          # successful main runs. The first pull-request run at 3 that met a
-          # busy node (run 39193827281) then hit the 75m timeout before any
-          # test started: the sampler shows cargo spinning for 70 minutes with
-          # the pod's io.pressure "full" at 86-88% and CPU pressure at zero,
-          # while node-level IO pressure was already 36% from co-tenants when
-          # the step began. The constraint is the pod's shared storage, not
-          # CPU or memory, and the limit that storage sustains on a contended
-          # node is 2. Main pushes and dispatches keep 3 (rarer, and three
-          # successes); pull requests stay at 2 until the pods get local disk
-          # for target/, which is the lever that would unlock the 14 cores.
+          # Raised to 3 on main pushes and manual dispatches; PRs keep 2 so the
+          # merge path is untouched while the experiment runs.
+          #
+          # During baseline collection, main pushes cancelled superseded runs:
+          # only 4 of 20 Test and Lint jobs reached a terminal state. Running
+          # main validations now finish; dispatch remains an independent group
+          # for collecting additional samples on demand.
+          #
+          # Baseline over 17 samples at 2:
+          # median nextest/clippy step ratio 1.95, spread 1.85-2.06. The gate-2
+          # criterion is that ratio dropping at least 10% (below ~1.76) with no
+          # 75m timeout and no run showing three consecutive samples of
+          # rustc/collect2/rust-lld in D state. If it does not, the conclusion is
+          # "this limit is not the bottleneck" — fix it back at 2 and record the
+          # experiment, which is a result, not a failure.
           #
           # Must stay step-level: rust-cache hashes CARGO/CC/CFLAGS/CXX/CMAKE/RUST
           # prefixed variables from process.env into the cache key, so promoting
@@ -280,55 +283,10 @@ jobs:
       - name: Verify rebalance/decommission migration proofs
         run: ./scripts/check_migration_gate_count.sh evidence artifacts/test-and-lint/core-test-listing.json target/nextest/ci/junit.xml
 
-      # The workspace already compiled these ignored tests. Keep the same Cargo
-      # package selection to reuse their binaries, then run them serially in
-      # separate processes because they share fixed ports and global state.
-      # test_transition_and_restore_flows was re-enabled by rustfs/backlog#1303:
-      # its "missing xl.meta on disk2" was a test-util bug (open_disk hardcoded
-      # disk_index 0), not an EC metadata-distribution issue.
-      # restore_object_usecase_reports_ongoing_conflict_and_completion was
-      # re-enabled by backlog#1304 (restore accepts serialize on a short CAS
-      # guard; the copy-back no longer holds the #4877 whole-copy-back lock,
-      # so the mid-restore ongoing read and fast 409 rejection it asserts are
-      # the implemented contract).
-      - name: Run ignored ILM integration tests serially
-        if: needs.classify-changes.outputs.mode == 'full'
-        env:
-          # Match the preceding workspace build and reuse its feature graph.
-          CARGO_BUILD_JOBS: ${{ (github.event_name == 'push' || github.event_name == 'workflow_dispatch') && '3' || '2' }}
-        run: |
-          mkdir -p artifacts/ilm-integration
-          set +e
-          NEXTEST_HIDE_PROGRESS_BAR=1 timeout --verbose --signal=TERM --kill-after=30s 80m \
-            cargo nextest run --profile default -j1 --run-ignored ignored-only \
-            --all --exclude e2e_test \
-            -E 'binary(lifecycle_integration_test) or (package(rustfs) and test(lifecycle_transition_api_test))' \
-            --status-level all --final-status-level all \
-            2>&1 | tee artifacts/ilm-integration/nextest.log
-          status=${PIPESTATUS[0]}
-          {
-            echo "exit_status=${status}"
-            echo "finished_at=$(date --utc --iso-8601=seconds)"
-            echo
-            echo "Remaining test-related processes:"
-            pgrep -af 'cargo|nextest|target/.*/deps/' || true
-          } > artifacts/ilm-integration/diagnostics.txt
-          exit "${status}"
-
-      - name: Upload ILM test diagnostics
-        if: always() && needs.classify-changes.outputs.mode == 'full'
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
-        with:
-          name: ilm-integration-${{ github.run_number }}-${{ github.run_attempt }}
-          path: artifacts/ilm-integration
-          retention-days: 3
-
-  # Both Connect boundaries rebuild only the rustfs crate on top of the restored
-  # ci-dev cache: the two e2e features are empty, so every dependency keeps its
-  # cached dev-profile artifact and each feature set recompiles rustfs once.
-  # The budget covers a ci-dev miss, where the first step builds the whole graph.
-  connect-feature-boundaries:
-    name: Connect Feature Boundaries
+  # This gate builds into a fresh target directory to isolate the E2E root.
+  # Give its cold build a separate budget from workspace tests and migration proofs.
+  offline-enrollment-e2e:
+    name: Offline Enrollment E2E Root Boundary
     if: needs.classify-changes.outputs.mode == 'full' && (github.event_name != 'pull_request' || github.event.action != 'closed')
     needs: [ quick-checks, classify-changes ]
     runs-on: sm-standard-4
@@ -353,58 +311,86 @@ jobs:
       - name: Protect Connect test home
         run: chmod go-w "$(realpath "$HOME")"
 
-      - name: Run short credential behavior tests
-        env:
-          CARGO_BUILD_JOBS: "2"
-        run: |
-          cargo test -p rustfs --test connect_registration \
-            --features connect-e2e-short-credentials \
-            registration_enforces_build_profile_credential_lifetime -- --exact
-          cargo test -p rustfs --test connect_registration \
-            --features connect-e2e-short-credentials \
-            rotation_waits_for_threshold_and_stops_on_revocation -- --exact
-
-      # The guard in rustfs/src/connect/registration.rs is cfg(not(debug_assertions)),
-      # and it must be tripped by a profile that turns debug assertions off for
-      # the whole graph. Handing `-C debug-assertions=off` to the rustfs lib alone
-      # through `cargo rustc` leaves its dependencies on the dev-profile cfg, so
-      # the lib also fails on items gated on one side only (E0599 on
-      # Filenames::Embedded) and "fails solely at the compile_error" can never
-      # hold. `cargo check --release` is metadata-only and needs no
-      # link-concurrency limit; it is cold because ci-dev holds no release
-      # artifacts.
-      - name: Reject short credentials in release builds
-        run: |
-          log="$(mktemp)"
-          set +e
-          CARGO_TERM_COLOR=never cargo check -p rustfs --release \
-            --features connect-e2e-short-credentials >"$log" 2>&1
-          status=$?
-          set -e
-          cat "$log"
-          expected='error: connect-e2e-short-credentials is restricted to debug builds'
-          summary="error: could not compile \`rustfs\` (lib) due to 1 previous error"
-          expected_count="$(grep -Fxc "$expected" "$log" || true)"
-          summary_count="$(grep -Fc "$summary" "$log" || true)"
-          error_count="$(grep -Ec '^error(:|\[)' "$log" || true)"
-          if [ "$status" -ne 101 ] || [ "$expected_count" -ne 1 ] \
-            || [ "$summary_count" -ne 1 ] || [ "$error_count" -ne 2 ]; then
-            echo "feature gate did not fail solely at the expected compile_error" >&2
-            rm -f "$log"
-            exit 1
-          fi
-          rm -f "$log"
-
       - name: Check offline enrollment E2E root boundary
         run: ./scripts/check_offline_enrollment_e2e.sh
 
+  # Dedicated serial lane for the ILM / lifecycle integration tests. These tests
+  # drive the object layer through process-global singletons (the GLOBAL_ENV
+  # ECStore, the global tier-config manager, background-expiry workers) and bind
+  # fixed ports (9002 for the scanner suite, 9003 for the app suite), so they are
+  # marked #[ignore] to keep them out of the default parallel `cargo nextest run
+  # --all` pass. Run them here explicitly with `--run-ignored ignored-only` and
+  # `-j1` so no two of them share global state or race for a port at once.
+  # serial_test's #[serial] does NOT serialize across nextest's process-per-test
+  # boundary (see .config/nextest.toml); `-j1` is what actually serializes them.
+  # See rustfs/backlog#1148 (ilm-1) and #1155.
+  test-ilm-integration-serial:
+    name: ILM Integration (serial)
+    if: needs.classify-changes.outputs.mode == 'full' && (github.event_name != 'pull_request' || github.event.action != 'closed')
+    needs: [ quick-checks, classify-changes ]
+    runs-on: sm-standard-4
+    timeout-minutes: 90
+    env:
+      FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
+        with:
+          persist-credentials: false
+
+      - name: Setup Rust environment
+        uses: ./.github/actions/setup
+        with:
+          rust-version: stable
+          cache-shared-key: ci-dev
+          cache-save-if: 'false'
+          install-build-packaging-tools: 'false'
+
+      # test_transition_and_restore_flows was re-enabled by rustfs/backlog#1303:
+      # its "missing xl.meta on disk2" was a test-util bug (open_disk hardcoded
+      # disk_index 0), not an EC metadata-distribution issue.
+      # restore_object_usecase_reports_ongoing_conflict_and_completion was
+      # re-enabled by backlog#1304 (restore accepts serialize on a short CAS
+      # guard; the copy-back no longer holds the #4877 whole-copy-back lock,
+      # so the mid-restore ongoing read and fast 409 rejection it asserts are
+      # the implemented contract).
+      - name: Run ignored ILM integration tests serially
+        env:
+          # Match the measured Test and Lint link budget. The default exposed
+          # all 14 pod CPUs and a cold cache spent the full 80m compiling
+          # without starting one ILM test (main run 32982910990).
+          CARGO_BUILD_JOBS: ${{ (github.event_name == 'push' || github.event_name == 'workflow_dispatch') && '3' || '2' }}
+        run: |
+          mkdir -p artifacts/ilm-integration
+          set +e
+          NEXTEST_HIDE_PROGRESS_BAR=1 timeout --verbose --signal=TERM --kill-after=30s 80m \
+            cargo nextest run -j1 --run-ignored ignored-only \
+            -p rustfs-scanner -p rustfs \
+            -E 'binary(lifecycle_integration_test) or (package(rustfs) and test(lifecycle_transition_api_test))' \
+            --status-level all --final-status-level all \
+            2>&1 | tee artifacts/ilm-integration/nextest.log
+          status=${PIPESTATUS[0]}
+          {
+            echo "exit_status=${status}"
+            echo "finished_at=$(date --utc --iso-8601=seconds)"
+            echo
+            echo "Remaining test-related processes:"
+            pgrep -af 'cargo|nextest|target/.*/deps/' || true
+          } > artifacts/ilm-integration/diagnostics.txt
+          exit "${status}"
+
+      - name: Upload ILM test diagnostics
+        if: always()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: ilm-integration-${{ github.run_number }}-${{ github.run_attempt }}
+          path: |
+            artifacts/ilm-integration
+            target/nextest/ci/junit.xml
+
   test-and-lint-rio-v2:
     name: Test and Lint (rio-v2)
-    # Alternate feature configurations run on the nightly schedule, merge
-    # groups and manual dispatch. The default workspace remains the PR gate and
-    # the per-push main baseline: repeating four feature builds on every merge
-    # bought nothing the nightly run does not report within a day.
-    if: needs.classify-changes.outputs.mode == 'full' && contains(fromJSON('["merge_group", "schedule", "workflow_dispatch"]'), github.event_name)
+    if: needs.classify-changes.outputs.mode == 'full' && (github.event_name != 'pull_request' || github.event.action != 'closed')
     needs: [ quick-checks, classify-changes ]
     runs-on: sm-standard-4
     timeout-minutes: 90
@@ -460,10 +446,67 @@ jobs:
           retention-days: 3
           if-no-files-found: error
 
+  connect-short-credential-boundary:
+    name: Connect Short Credential Boundary
+    if: needs.classify-changes.outputs.mode == 'full' && (github.event_name != 'pull_request' || github.event.action != 'closed')
+    needs: [ quick-checks, classify-changes ]
+    runs-on: sm-standard-4
+    timeout-minutes: 60
+    env:
+      FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
+        with:
+          persist-credentials: false
+
+      - name: Setup Rust environment
+        uses: ./.github/actions/setup
+        with:
+          rust-version: stable
+          cache-shared-key: ci-dev
+          cache-save-if: 'false'
+          install-build-packaging-tools: 'false'
+          install-test-tools: 'false'
+
+      - name: Run short credential behavior tests
+        env:
+          CARGO_BUILD_JOBS: "2"
+        run: |
+          cargo test -p rustfs --test connect_registration \
+            --features connect-e2e-short-credentials \
+            registration_enforces_build_profile_credential_lifetime -- --exact
+          cargo test -p rustfs --test connect_registration \
+            --features connect-e2e-short-credentials \
+            rotation_waits_for_threshold_and_stops_on_revocation -- --exact
+
+      - name: Reject short credentials in release builds
+        env:
+          CARGO_BUILD_JOBS: "2"
+        run: |
+          log="$(mktemp)"
+          set +e
+          CARGO_TERM_COLOR=never cargo check -p rustfs --release \
+            --features connect-e2e-short-credentials >"$log" 2>&1
+          status=$?
+          set -e
+          cat "$log"
+          expected='error: connect-e2e-short-credentials is restricted to debug builds'
+          summary="error: could not compile \`rustfs\` (lib) due to 1 previous error"
+          expected_count="$(grep -Fxc "$expected" "$log" || true)"
+          summary_count="$(grep -Fc "$summary" "$log" || true)"
+          error_count="$(grep -Ec '^error(:|\[)' "$log" || true)"
+          if [ "$status" -ne 101 ] || [ "$expected_count" -ne 1 ] \
+            || [ "$summary_count" -ne 1 ] || [ "$error_count" -ne 2 ]; then
+            echo "release feature gate did not fail solely at the expected compile_error" >&2
+            rm -f "$log"
+            exit 1
+          fi
+          rm -f "$log"
+
   test-and-lint-protocols:
     name: "Test and Lint (${{ matrix.features.name }})"
-    # Same cadence as test-and-lint-rio-v2 above.
-    if: needs.classify-changes.outputs.mode == 'full' && contains(fromJSON('["merge_group", "schedule", "workflow_dispatch"]'), github.event_name)
+    if: needs.classify-changes.outputs.mode == 'full' && (github.event_name != 'pull_request' || github.event.action != 'closed')
     needs: [ quick-checks, classify-changes ]
     runs-on: sm-standard-4
     timeout-minutes: 90
@@ -619,7 +662,7 @@ jobs:
   build-rustfs-debug-binary-rio-v2:
     name: Build RustFS Debug Binary (rio-v2)
     # Dormant rio-v2 variant (rustfs/backlog#1835): the feature ships in no
-    # default build, so this full-suite lane runs only on the nightly schedule
+    # default build, so this full-suite lane runs only on the weekly schedule
     # and manual dispatch. Per-PR cfg-seam coverage stays with
     # test-and-lint-rio-v2. Lifecycle and the promote-or-delete condition:
     # docs/architecture/minio-file-format-compat.md ("rio-v2 variant lifecycle").
@@ -1089,7 +1132,7 @@ jobs:
   s3-implemented-tests:
     name: S3 Implemented Tests
     needs: [ build-rustfs-debug-binary, e2e-tests ]
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-4
     timeout-minutes: 60
     steps:
       - name: Checkout repository
@@ -1158,12 +1201,12 @@ jobs:
     name: S3 Lifecycle Behavior Tests
     # Also gated on e2e-tests, matching s3-implemented-tests: when the e2e smoke
     # suite is already red this lane cannot tell us anything new, and it holds a
-    # runner for up to 30 minutes doing so. Both lanes only download the
+    # sm-standard-4 for up to 30 minutes doing so. Both lanes only download the
     # prebuilt debug binary (no cargo build), and s3-implemented-tests — which
     # already waits on e2e-tests — finishes later anyway, so a green PR's total
     # wall clock is unchanged.
     needs: [ build-rustfs-debug-binary, e2e-tests ]
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-4
     timeout-minutes: 30
     steps:
       - name: Checkout repository
@@ -1220,8 +1263,10 @@ jobs:
       - typos
       - quick-checks
       - test-and-lint
-      - connect-feature-boundaries
+      - offline-enrollment-e2e
+      - test-ilm-integration-serial
       - test-and-lint-rio-v2
+      - connect-short-credential-boundary
       - test-and-lint-protocols
       - build-rustfs-debug-binary
       - uring-integration
@@ -1231,7 +1276,7 @@ jobs:
       - build-rustfs-debug-binary-rio-v2
       - e2e-tests-rio-v2
       - e2e-full
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
@@ -1247,11 +1292,13 @@ jobs:
     name: Alert on scheduled failure
     needs:
       - classify-changes
-      - connect-feature-boundaries
+      - connect-short-credential-boundary
       - required-checks
       - typos
       - quick-checks
       - test-and-lint
+      - offline-enrollment-e2e
+      - test-ilm-integration-serial
       - test-and-lint-rio-v2
       - test-and-lint-protocols
       - build-rustfs-debug-binary

.github/workflows/cla.yml

@@ -46,7 +46,7 @@ jobs:
     # Echoes one line; the run exists only so the concurrency group cancels the
     # in-flight run of a closed PR.
     permissions: {}
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Explain cancellation run
@@ -61,7 +61,7 @@ jobs:
       checks: write
       issues: write
       pull-requests: write
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     steps:
       - name: Report CLA result for merge queue

.github/workflows/connect-diagnostic-schedule-acceptance.yml

@@ -57,14 +57,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -102,7 +94,7 @@ jobs:
           ' <<<"$run" >/dev/null
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e '
-            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu\\)$"))] as $matches
+            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu, [a-z0-9-]+, x86_64-unknown-linux-gnu, false, linux, pyroscope\\)$"))] as $matches
             | ($matches | length) == 1 and $matches[0].conclusion == "success"
           ' <<<"$jobs" >/dev/null
           artifact=$(gh api "repos/${GITHUB_REPOSITORY}/actions/artifacts/${ARTIFACT_ID}")
@@ -145,7 +137,7 @@ jobs:
               assert outer.testzip() is None
               with zipfile.ZipFile(io.BytesIO(outer.read(package))) as inner:
                   entries = inner.infolist()
-                  assert {item.filename for item in entries} == {'rustfs', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'} and len(entries) == 3
+                  assert {item.filename for item in entries} == {'rustfs', 'rustfs-cli'} and len(entries) == 2
                   assert all(regular(item) and 0 < item.file_size <= 1073741824 for item in entries)
                   assert inner.testzip() is None
                   pathlib.Path('binary').mkdir()
@@ -170,7 +162,7 @@ jobs:
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
           (cd connect-harness && ./tests/connectivity/diagnostic-schedule-evidence.test.sh)
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_DIAGNOSTIC_SCHEDULE_EVIDENCE="$GITHUB_WORKSPACE/diagnostic-schedule-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=diagnostic-schedule

.github/workflows/connect-health-service-acceptance.yml

@@ -57,14 +57,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -135,7 +127,7 @@ jobs:
           ' <<<"$run" >/dev/null
           jobs=$(gh api --paginate --slurp "repos/rustfs/rustfs/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e '
-            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu\\)$"))] as $matches
+            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu, [a-z0-9-]+, x86_64-unknown-linux-gnu, false, linux, pyroscope\\)$"))] as $matches
             | ($matches | length) == 1 and $matches[0].conclusion == "success"
           ' <<<"$jobs" >/dev/null
           artifact=$(gh api "repos/rustfs/rustfs/actions/artifacts/${ARTIFACT_ID}")
@@ -180,7 +172,7 @@ jobs:
               assert outer.testzip() is None
               with zipfile.ZipFile(io.BytesIO(outer.read(package))) as inner:
                   entries = inner.infolist()
-                  assert {entry.filename for entry in entries} == {'rustfs', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'} and len(entries) == 3
+                  assert {entry.filename for entry in entries} == {'rustfs', 'rustfs-cli', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'} and len(entries) == 4
                   assert all(regular(entry) and 0 < entry.file_size <= 1073741824 for entry in entries)
                   assert inner.testzip() is None
                   pathlib.Path('binary').mkdir()
@@ -207,7 +199,7 @@ jobs:
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
           (cd connect-harness && ./tests/connectivity/health-service-evidence.test.sh)
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_HEALTH_SERVICE_EVIDENCE="$GITHUB_WORKSPACE/health-service-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=health-service

.github/workflows/connect-performance-drive-acceptance.yml

@@ -73,14 +73,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -115,7 +107,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.[].jobs[] | select(.name == $name)] as $matches
@@ -162,7 +154,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_PERFORMANCE_DRIVE_EVIDENCE="$GITHUB_WORKSPACE/performance-drive-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=performance-drive

.github/workflows/connect-performance-network-acceptance.yml

@@ -73,14 +73,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -115,7 +107,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.[].jobs[] | select(.name == $name)] as $matches
@@ -162,7 +154,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_PERFORMANCE_NETWORK_EVIDENCE="$GITHUB_WORKSPACE/performance-network-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=performance-network

.github/workflows/connect-profile-cpu-acceptance.yml

@@ -77,14 +77,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -126,7 +118,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(github_api "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.jobs[] | select(.name == $name)] as $matches
@@ -181,7 +173,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_CPU_CATALOG="$GITHUB_WORKSPACE/binary/rustfs.cpu-symbol-catalog.json" \
             CONNECT_E2E_CPU_CATALOG_SHA256="$CATALOG_SHA256" \

.github/workflows/connect-profile-memory-acceptance.yml

@@ -57,14 +57,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -135,7 +127,7 @@ jobs:
           ' <<<"$run" >/dev/null
           jobs=$(gh api --paginate --slurp "repos/rustfs/rustfs/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e '
-            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu\\)$"))] as $matches
+            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu, [a-z0-9-]+, x86_64-unknown-linux-gnu, false, linux, pyroscope\\)$"))] as $matches
             | ($matches | length) == 1 and $matches[0].conclusion == "success"
           ' <<<"$jobs" >/dev/null
           artifact=$(gh api "repos/rustfs/rustfs/actions/artifacts/${ARTIFACT_ID}")
@@ -180,7 +172,7 @@ jobs:
               assert outer.testzip() is None
               with zipfile.ZipFile(io.BytesIO(outer.read(package))) as inner:
                   entries = inner.infolist()
-                  expected = {'rustfs', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'}
+                  expected = {'rustfs', 'rustfs-cli', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'}
                   assert {entry.filename for entry in entries} == expected and len(entries) == len(expected)
                   assert all(regular(entry) and 0 < entry.file_size <= 1073741824 for entry in entries)
                   assert inner.testzip() is None
@@ -206,7 +198,7 @@ jobs:
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
           (cd connect-harness && ./tests/connectivity/profile-memory-evidence.test.sh)
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_PROFILE_MEMORY_EVIDENCE="$GITHUB_WORKSPACE/profile-memory-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=profile-memory

.github/workflows/connect-profile-threads-acceptance.yml

@@ -81,14 +81,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -156,7 +148,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.[].jobs[] | select(.name == $name)] as $matches
@@ -209,7 +201,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_PROFILE_THREADS_EVIDENCE="$GITHUB_WORKSPACE/profile-threads-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=profile-threads

.github/workflows/connect-top-api-acceptance.yml

@@ -73,14 +73,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -115,7 +107,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.[].jobs[] | select(.name == $name)] as $matches
@@ -162,7 +154,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_TOP_API_EVIDENCE="$GITHUB_WORKSPACE/top-api-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=top-api

.github/workflows/connect-top-disk-acceptance.yml

@@ -68,14 +68,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -146,7 +138,7 @@ jobs:
           ' <<<"$run" >/dev/null
           jobs=$(gh api --paginate --slurp "repos/rustfs/rustfs/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e '
-            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu\\)$"))] as $matches
+            [.[].jobs[] | select(.name | test("^Build RustFS \\(linux-x86_64-gnu, [a-z0-9-]+, x86_64-unknown-linux-gnu, false, linux, pyroscope\\)$"))] as $matches
             | ($matches | length) == 1 and $matches[0].conclusion == "success"
           ' <<<"$jobs" >/dev/null
           artifact=$(gh api "repos/rustfs/rustfs/actions/artifacts/${ARTIFACT_ID}")
@@ -191,7 +183,7 @@ jobs:
               assert outer.testzip() is None
               with zipfile.ZipFile(io.BytesIO(outer.read(package))) as inner:
                   entries = inner.infolist()
-                  expected = {'rustfs', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'}
+                  expected = {'rustfs', 'rustfs-cli', 'rustfs.cpu-symbol-catalog.json', 'rustfs.cpu-symbol-catalog.sha256'}
                   assert {entry.filename for entry in entries} == expected and len(entries) == len(expected)
                   assert all(regular(entry) and 0 < entry.file_size <= 1073741824 for entry in entries)
                   assert inner.testzip() is None
@@ -221,7 +213,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_TOP_DISK_EVIDENCE="$GITHUB_WORKSPACE/top-disk-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=top-disk

.github/workflows/connect-top-locks-acceptance.yml

@@ -73,14 +73,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -115,7 +107,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.[].jobs[] | select(.name == $name)] as $matches
@@ -162,7 +154,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_TOP_LOCKS_EVIDENCE="$GITHUB_WORKSPACE/top-locks-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=top-locks

.github/workflows/connect-top-rpc-acceptance.yml

@@ -73,14 +73,6 @@ jobs:
           ref: ${{ inputs.connect_sha }}
           token: ${{ secrets.PF_TESTING_GH_TOKEN }}
 
-      - name: Set up Docker Compose
-        uses: ./.github/actions/setup-docker-compose
-
-      - name: Set up Python
-        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
-        with:
-          python-version: "3.14"
-
       - name: Set up Node.js
         uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
@@ -115,7 +107,7 @@ jobs:
           [[ $(jq -r '.head_repository.full_name' <<<"$run") == "$GITHUB_REPOSITORY" ]]
           [[ $(jq -r '.name' <<<"$run") == "Build and Release" ]]
 
-          expected_job='Build RustFS (linux-x86_64-gnu)'
+          expected_job='Build RustFS (linux-x86_64-gnu, sm-standard-4, x86_64-unknown-linux-gnu, false, linux, pyroscope)'
           jobs=$(gh api --paginate --slurp "repos/${GITHUB_REPOSITORY}/actions/runs/${BUILD_RUN_ID}/jobs?per_page=100")
           jq -e --arg name "$expected_job" '
             [.[].jobs[] | select(.name == $name)] as $matches
@@ -162,7 +154,7 @@ jobs:
           npm --prefix connect-harness/web ci
           connect-harness/web/node_modules/.bin/playwright install --with-deps chromium
           make -C connect-harness e2e-connected-dispatch-check
-          TMPDIR="$GITHUB_WORKSPACE" RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
+          RUSTFS_BINARY="$GITHUB_WORKSPACE/binary/rustfs" \
             RUSTFS_WORKTREE="$GITHUB_WORKSPACE/rustfs-source" \
             CONNECT_E2E_TOP_RPC_EVIDENCE="$GITHUB_WORKSPACE/top-rpc-service-job-evidence.json" \
             make -C connect-harness e2e-connected E2E_SCENARIO=top-rpc

.github/workflows/coverage.yml

@@ -45,7 +45,7 @@ on:
     # 07:00 UTC Sunday — staggered clear of the other Sunday crons: ci (00:00),
     # build (01:00), e2e-s3tests (02:00), audit (03:00), nix-flake-update
     # (05:00), mint (06:00), and the daily fuzz (02:00), minio-interop (03:17),
-    # and e2e-replication-nightly (04:00) lanes.
+    # e2e-replication-nightly (04:00) and performance-ab (06:00) lanes.
     - cron: "43 7 * * 0"
 
 concurrency:

.github/workflows/docker.yml

@@ -83,7 +83,7 @@ jobs:
       (github.event.workflow_run.conclusion == 'success' &&
        github.event.workflow_run.event == 'push' &&
        github.event.workflow_run.head_branch != 'main')
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     outputs:
       should_build: ${{ steps.check.outputs.should_build }}
@@ -501,7 +501,7 @@ jobs:
     name: Scan Docker Images
     needs: [ build-check, build-docker ]
     if: needs.build-check.outputs.should_build == 'true' && needs.build-check.outputs.should_push == 'true'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     permissions:
       contents: read
@@ -556,7 +556,7 @@ jobs:
     name: Docker Build Summary
     needs: [ build-check, build-docker ]
     if: always() && needs.build-check.outputs.should_build == 'true'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Docker build completion summary

.github/workflows/e2e-distributed.yml

@@ -153,7 +153,7 @@ jobs:
 
       - name: Build rustfs binary
         run: |
-          python3 scripts/e2e_binary.py build --bins --features e2e-test-hooks
+          python3 scripts/e2e_binary.py build --bins
 
       - name: Verify distributed e2e membership
         env:
@@ -170,9 +170,9 @@ jobs:
           set -euo pipefail
           rm -f target/nextest/e2e-distributed/junit.xml
           if [ -n "${FILTER}" ]; then
-            python3 scripts/e2e_binary.py run --features e2e-test-hooks -- cargo nextest run --profile e2e-distributed -p e2e_test -E "${FILTER}"
+            python3 scripts/e2e_binary.py run -- cargo nextest run --profile e2e-distributed -p e2e_test -E "${FILTER}"
           else
-            python3 scripts/e2e_binary.py run --features e2e-test-hooks -- cargo nextest run --profile e2e-distributed -p e2e_test --no-tests=fail
+            python3 scripts/e2e_binary.py run -- cargo nextest run --profile e2e-distributed -p e2e_test --no-tests=fail
           fi
           test -s target/nextest/e2e-distributed/junit.xml
 

.github/workflows/e2e-replication-nightly.yml

@@ -38,8 +38,8 @@ name: e2e-nightly
 on:
   workflow_dispatch:
   schedule:
-    # 04:00 UTC nightly — staggered clear of fuzz/e2e-s3tests (02:00)
-    # and stale (01:30).
+    # 04:00 UTC nightly — staggered clear of fuzz/e2e-s3tests (02:00),
+    # stale (01:30) and performance-ab (06:00).
     - cron: "29 4 * * *"
 
 # Only alert-on-failure needs more than read access; it declares its own

.github/workflows/e2e-s3tests.yml

@@ -371,7 +371,7 @@ jobs:
     name: Upstream HEAD classification canary
     if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
     continue-on-error: true
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 20
     steps:
       - uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7

.github/workflows/e2e-upgrade.yml

@@ -19,15 +19,12 @@ on:
     paths:
       - ".github/workflows/e2e-upgrade.yml"
       - "crates/e2e_test/src/common.rs"
-      - "crates/e2e_test/src/distributed/**"
       - "crates/e2e_test/src/fake_s3_target/**"
       - "crates/e2e_test/src/lib.rs"
       - "crates/e2e_test/src/replication_extension_test.rs"
-      - "crates/e2e_test/src/rpc_fault_proxy.rs"
       - "crates/e2e_test/src/upgrade_compatibility_test.rs"
       - "crates/ecstore/**"
       - "crates/filemeta/**"
-      - "crates/heal/**"
       - "crates/kms/**"
       - "crates/storage-api/**"
       - "rustfs/**"
@@ -49,11 +46,14 @@ concurrency:
 env:
   CARGO_TERM_COLOR: always
   RUST_BACKTRACE: 1
+  UPGRADE_SOURCE_VERSION: 1.0.0-rc.5
+  UPGRADE_SOURCE_ASSET: rustfs-linux-x86_64-gnu-v1.0.0-rc.5.zip
+  UPGRADE_SOURCE_SHA256: 3ee8df71e8edcfada533be452c4135868f697bc515460ae97b027313eade7a3d
 
 jobs:
   build:
     name: Build upgrade candidate
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 60
     env:
       FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
@@ -66,11 +66,19 @@ jobs:
       - name: Setup Rust environment
         uses: ./.github/actions/setup
         with:
-          cache-shared-key: e2e-upgrade-server-hosted
+          cache-shared-key: e2e-upgrade-server
           cache-save-if: ${{ github.ref == 'refs/heads/main' }}
           install-build-packaging-tools: "false"
           install-test-tools: "false"
 
+      - name: Download pinned previous release
+        run: |
+          mkdir -p target/debug
+          archive="target/debug/$UPGRADE_SOURCE_ASSET"
+          curl --fail --location --retry 3 --output "$archive" \
+            "https://github.com/${GITHUB_REPOSITORY}/releases/download/${UPGRADE_SOURCE_VERSION}/${UPGRADE_SOURCE_ASSET}"
+          echo "$UPGRADE_SOURCE_SHA256  $archive" | sha256sum --check --strict
+
       - name: Build current RustFS binary
         run: python3 scripts/e2e_binary.py build
 
@@ -81,31 +89,49 @@ jobs:
           path: |
             target/debug/rustfs
             target/debug/rustfs.e2e.json
+            target/debug/${{ env.UPGRADE_SOURCE_ASSET }}
           if-no-files-found: error
           retention-days: 3
           compression-level: 0
           overwrite: true
 
   upgrade:
-    name: Upgrade compatibility (${{ matrix.source.version }})
+    name: ${{ matrix.name }}
     needs: build
-    runs-on: ubuntu-latest
-    timeout-minutes: 60
     strategy:
       fail-fast: false
       matrix:
-        source:
-          - version: 1.0.0-rc.5
-            asset: rustfs-linux-x86_64-gnu-v1.0.0-rc.5.zip
-            sha256: 3ee8df71e8edcfada533be452c4135868f697bc515460ae97b027313eade7a3d
-          - version: 1.0.1
-            asset: rustfs-linux-x86_64-gnu-v1.0.1.zip
-            sha256: 9742e183cb0dfb0d48fa909797af572fe47b7f827569036baa4156e0d3958ee5
+        include:
+          # The two `_from_rc2_` tests keep their names: they assert
+          # release-independent object contracts and pass unchanged against the
+          # newer pinned source, so renaming them would only churn history and
+          # the CI required-check names. UPGRADE_SOURCE_VERSION above is the
+          # single source of truth for which release they actually run against.
+          - name: Direct upgrade from the previous release
+            test: direct_upgrade_from_rc2_preserves_object_contracts
+            artifact: direct-upgrade
+          - name: Mixed-version rolling upgrade from the previous release
+            test: rolling_upgrade_from_rc2_preserves_mixed_version_contracts
+            artifact: mixed-version-upgrade
+          - name: Bucket configuration survives the upgrade
+            test: direct_upgrade_from_previous_release_preserves_bucket_configuration
+            artifact: bucket-config-upgrade
+          - name: Rollback reads current bucket metadata
+            test: rollback_to_previous_release_reads_current_bucket_metadata
+            artifact: bucket-config-rollback
+          - name: ODM configuration recovery after rc.5 rollback
+            test: rc5_rollback_requires_restoring_odm_configuration
+            artifact: odm-config-rollback
+          - name: Multipart layouts survive the rc.5 upgrade
+            test: direct_upgrade_from_rc5_preserves_multipart_layouts
+            artifact: multipart-layout-upgrade
+          - name: rc.5 multipart replication baseline
+            test: rc5_baseline_replicates_multipart_layouts
+            artifact: multipart-layout-baseline
+    runs-on: sm-standard-2
+    timeout-minutes: 60
     env:
       FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
-      UPGRADE_SOURCE_VERSION: ${{ matrix.source.version }}
-      UPGRADE_SOURCE_ASSET: ${{ matrix.source.asset }}
-      UPGRADE_SOURCE_SHA256: ${{ matrix.source.sha256 }}
     steps:
       - name: Checkout repository
         uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
@@ -115,8 +141,9 @@ jobs:
       - name: Setup Rust environment
         uses: ./.github/actions/setup
         with:
-          cache-shared-key: e2e-upgrade-tests-hosted
-          cache-save-if: ${{ github.ref == 'refs/heads/main' }}
+          cache-shared-key: e2e-upgrade-tests
+          # One main-branch consumer warms the shared test-harness dependencies.
+          cache-save-if: ${{ github.ref == 'refs/heads/main' && matrix.artifact == 'direct-upgrade' }}
           install-build-packaging-tools: "false"
           install-test-tools: "false"
 
@@ -130,15 +157,12 @@ jobs:
         run: chmod +x target/debug/rustfs
 
       - name: Verify and unpack pinned previous release
-        id: verify_source
         env:
           SOURCE_DIR: ${{ runner.temp }}/rustfs-upgrade-source
         run: |
           set -euo pipefail
           mkdir -p "$SOURCE_DIR"
           archive="target/debug/$UPGRADE_SOURCE_ASSET"
-          curl --fail --location --retry 3 --output "$archive" \
-            "https://github.com/rustfs/rustfs/releases/download/${UPGRADE_SOURCE_VERSION}/${UPGRADE_SOURCE_ASSET}"
           echo "$UPGRADE_SOURCE_SHA256  $archive" | sha256sum --check --strict
           unzip -q "$archive" -d "$SOURCE_DIR"
           chmod +x "$SOURCE_DIR/rustfs"
@@ -146,76 +170,22 @@ jobs:
           echo "RUSTFS_UPGRADE_SOURCE_BINARY=$SOURCE_DIR/rustfs" >> "$GITHUB_ENV"
           echo "RUSTFS_E2E_LOG_DIR=$RUNNER_TEMP/rustfs-upgrade-logs" >> "$GITHUB_ENV"
 
-      # Reuse one compiled harness, while keeping each case in its own process.
-      # Continue after a failure to retain the former matrix's complete verdict.
-      - name: Run upgrade compatibility tests
-        shell: bash
+      - name: Run upgrade compatibility test
         env:
+          RUSTFS_SCANNER_HEAL_G09_EVIDENCE_DIR: ${{ runner.temp }}/rustfs-upgrade-g09-evidence/${{ matrix.artifact }}
           NO_PROXY: 127.0.0.1,localhost,::1
         run: |
           export no_proxy="$NO_PROXY"
-          cases=(
-            direct-upgrade:direct_upgrade_from_previous_release_preserves_object_contracts
-            mixed-version-upgrade:rolling_upgrade_from_previous_release_preserves_mixed_version_contracts
-            bucket-config-upgrade:direct_upgrade_from_previous_release_preserves_bucket_configuration
-            bucket-config-rollback:rollback_to_previous_release_reads_current_bucket_metadata
-            multipart-layout-upgrade:direct_upgrade_from_previous_release_preserves_multipart_layouts
-            multipart-layout-baseline:previous_release_baseline_replicates_multipart_layouts
-          )
-          if [[ "$UPGRADE_SOURCE_VERSION" == "1.0.0-rc.5" ]]; then
-            cases+=(odm-config-rollback:rc5_rollback_requires_restoring_odm_configuration)
-            cases+=(checkpoint-upgrade:rc5_checkpoint_upgrade_rewinds_signed_progress_without_losing_objects)
-          else
-            cases+=(legacy-delete-reservation:stable_upgrade_deletes_foreign_reservations_without_losing_live_objects)
-          fi
-          status=0
-          for entry in "${cases[@]}"; do
-            artifact="${entry%%:*}"
-            test="${entry#*:}"
-            export RUSTFS_SCANNER_HEAL_G09_EVIDENCE_DIR="$RUNNER_TEMP/rustfs-upgrade-g09-evidence/$artifact"
-            export RUSTFS_E2E_LOG_DIR="$RUNNER_TEMP/rustfs-upgrade-logs/$artifact"
-            mkdir -p "$RUSTFS_E2E_LOG_DIR"
-            echo "::group::$test"
-            if python3 scripts/e2e_binary.py run -- cargo test --locked -p e2e_test \
-              "upgrade_compatibility_test::$test" -- --ignored --exact --nocapture \
-              2>&1 | tee "$RUSTFS_E2E_LOG_DIR/test.log"; then
-              echo "PASS: $test"
-            else
-              echo "::error::Upgrade compatibility failed: $test"
-              status=1
-            fi
-            echo "::endgroup::"
-          done
-          exit "$status"
-
-      - name: Run mixed-version MPU recovery
-        if: ${{ !cancelled() && steps.verify_source.outcome == 'success' }}
-        timeout-minutes: 15
-        env:
-          NO_PROXY: 127.0.0.1,localhost,::1
-          HTTP_PROXY: ""
-          HTTPS_PROXY: ""
-          ALL_PROXY: ""
-          RUSTFS_E2E_LOG_FILTER: "warn,rustfs_ecstore::cluster::rpc::client=debug"
-        run: |
-          set -euo pipefail
-          export no_proxy="$NO_PROXY"
-          export RUSTFS_E2E_LOG_DIR="$RUNNER_TEMP/rustfs-upgrade-logs/mpu-mixed-recovery"
-          mkdir -p "$RUSTFS_E2E_LOG_DIR"
-          mpu_test='distributed::multipart_fence_acceptance_test::mixed_version_failed_complete_requires_current_coordinator_for_abort'
-          python3 scripts/e2e_binary.py run -- cargo test --locked -p e2e_test --lib \
-            "$mpu_test" -- --list --exact | tee "$RUSTFS_E2E_LOG_DIR/list.log"
-          test "$(grep -Fxc "${mpu_test}: test" "$RUSTFS_E2E_LOG_DIR/list.log")" -eq 1
-          python3 scripts/e2e_binary.py run -- cargo test --locked -p e2e_test --lib \
-            "$mpu_test" -- --exact --nocapture --test-threads=1 \
-            2>&1 | tee "$RUSTFS_E2E_LOG_DIR/test.log"
+          python3 scripts/e2e_binary.py run -- cargo test --locked -p e2e_test \
+            "upgrade_compatibility_test::${{ matrix.test }}" \
+            -- --ignored --exact --nocapture
 
       - name: Upload scanner/heal G09 evidence
         if: always()
         uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
-          name: upgrade-g09-evidence-${{ matrix.source.version }}-${{ github.run_number }}
-          path: ${{ runner.temp }}/rustfs-upgrade-g09-evidence
+          name: ${{ matrix.artifact }}-g09-evidence-${{ github.run_number }}
+          path: ${{ runner.temp }}/rustfs-upgrade-g09-evidence/${{ matrix.artifact }}
           if-no-files-found: ignore
           retention-days: 14
 
@@ -223,7 +193,7 @@ jobs:
         if: always()
         uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
-          name: upgrade-server-logs-${{ matrix.source.version }}-${{ github.run_number }}
+          name: ${{ matrix.artifact }}-server-logs-${{ github.run_number }}
           path: ${{ runner.temp }}/rustfs-upgrade-logs
           if-no-files-found: warn
           retention-days: 14

.github/workflows/fuzz.yml

@@ -60,7 +60,7 @@ jobs:
   cancel-closed-pr-runs:
     name: Cancel Closed PR Runs
     if: github.event_name == 'pull_request' && github.event.action == 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Explain cancellation run

.github/workflows/gateway-integration-sync.yml

@@ -6,20 +6,29 @@ name: Gateway Integration Sync
 # Merges `main` into the long-lived `gateway/integration` branch every weekday
 # so the s3s -> gateway migration (rustfs/backlog#2734, task T0.1 in #2736)
 # never accumulates merge debt. A clean merge lands as a `--no-ff` merge commit
-# pushed straight to the branch. A conflicting merge is aborted and fails the
-# run after its summary records the source and target SHAs, run URL, and exact
-# conflicting paths. It creates no sync branch or pull request. Resolve the
-# merge in an isolated worktree based on the latest target, preserve the
-# two-parent merge commit, and push only after the required checks pass and the
-# target still matches the merge's first parent. If the target moves, restart
-# from its new head.
+# pushed straight to the branch. A conflicting merge is aborted and surfaced as
+# a pull request from `sync/<source>-<date>`, whose head is the un-merged
+# source commit, so the conflict is resolved the same day, in review, with no
+# other change riding along. The run itself stays green in that case: the pull
+# request is the signal. Any other failure (push rejected, pull request not
+# created) fails the run and, on the schedule, files the shared failure issue.
 #
-# Write scope: only `gateway/integration` is ever pushed. The push uses a
-# one-hour installation token of the organisation's GitHub App (contents:
-# write) because that app is the only non-admin bypass actor of the
+# Resolving a conflict pull request: merge `origin/gateway/integration` into
+# the sync branch, resolve, push, get the review. Land it with a fast-forward
+# push of the resolved merge commit to `gateway/integration`, not with the
+# squash button: a squash drops the merge ancestry and the same hunks conflict
+# again on the next run.
+#
+# Write scope: only `gateway/integration` and `sync/*` are ever pushed, and
+# both push targets below are literal refs. The pushes use a one-hour
+# installation token of the organisation's own GitHub App (contents: write)
+# because that app is the only non-admin bypass actor of the
 # `gateway/integration` ruleset; GITHUB_TOKEN cannot be a ruleset bypass actor,
-# so it cannot push through the required reviews and checks. GITHUB_TOKEN is
-# read-only here.
+# so it cannot push through the required reviews and checks. The conflict pull
+# request is opened with FLAKE_UPDATE_TOKEN, the PAT that already opens the
+# weekly flake.lock pull requests, because this repository keeps "Allow GitHub
+# Actions to create and approve pull requests" switched off. GITHUB_TOKEN is
+# therefore read-only here.
 #
 # `schedule` fires only from the default branch, so the cron is live once this
 # file exists on `main`. `workflow_dispatch` runs from any ref that carries the
@@ -85,7 +94,6 @@ jobs:
         id: merge
         env:
           SOURCE_REF: ${{ inputs.source_ref || 'main' }}
-          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
         run: |
           set -euo pipefail
           if ! [[ "$SOURCE_REF" =~ ^[A-Za-z0-9][A-Za-z0-9._/-]*$ ]] \
@@ -111,25 +119,13 @@ jobs:
             fi
             exit 0
           fi
-          conflicts=$(git diff --name-only --diff-filter=U)
-          if [[ -z "$conflicts" ]]; then
-            git merge --abort || true
+          git diff --name-only --diff-filter=U > "$RUNNER_TEMP/conflicts.txt"
+          git merge --abort
+          if [[ ! -s "$RUNNER_TEMP/conflicts.txt" ]]; then
             echo "::error::git merge failed without reporting conflicting paths; this is not a conflict"
             exit 1
           fi
-          {
-            echo "## Gateway integration sync stopped on conflicts"
-            echo
-            echo "- Run: ${RUN_URL}"
-            echo "- Source: \`${SOURCE_REF}\` at ${source_sha}"
-            echo "- Target: \`${TARGET_BRANCH}\` at ${target_sha}"
-            echo
-            echo "Conflicting files:"
-            printf '%s\n' "$conflicts" | sed 's/^/- /'
-          } >> "$GITHUB_STEP_SUMMARY"
-          git merge --abort
-          echo "::error::${SOURCE_REF} cannot be merged cleanly into ${TARGET_BRANCH}; no branch or pull request was created"
-          exit 1
+          echo "result=conflict" >> "$GITHUB_OUTPUT"
 
       - name: Push the merge commit
         if: steps.merge.outputs.result == 'merged'
@@ -141,12 +137,50 @@ jobs:
           git push origin HEAD:refs/heads/gateway/integration
           echo "Merged \`${SOURCE_REF}\` (${SOURCE_SHA}) into \`${TARGET_BRANCH}\` as $(git rev-parse HEAD)." >> "$GITHUB_STEP_SUMMARY"
 
+      - name: Open the conflict pull request
+        if: steps.merge.outputs.result == 'conflict'
+        env:
+          GH_TOKEN: ${{ secrets.FLAKE_UPDATE_TOKEN }}
+          SOURCE_REF: ${{ steps.merge.outputs.source_ref }}
+          SOURCE_SHA: ${{ steps.merge.outputs.source_sha }}
+          TARGET_SHA: ${{ steps.merge.outputs.target_sha }}
+          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+        run: |
+          set -euo pipefail
+          date_utc=$(date -u +%Y-%m-%d)
+          sync_branch="sync/${SOURCE_REF//\//-}-${date_utc}"
+          git push origin "refs/remotes/origin/${SOURCE_REF}:refs/heads/${sync_branch}"
+          {
+            echo "The scheduled merge of \`${SOURCE_REF}\` into \`${TARGET_BRANCH}\` stopped on conflicts and was aborted; \`${TARGET_BRANCH}\` is unchanged."
+            echo
+            echo "This branch is the un-merged \`${SOURCE_REF}\` head. Merge \`origin/${TARGET_BRANCH}\` into it, resolve the files below, push, and get the review. Land the resolved merge commit with a fast-forward push to \`${TARGET_BRANCH}\`; a squash merge drops the merge ancestry and the same hunks conflict again on the next run. Do not include any change that is not conflict resolution."
+            echo
+            echo "- Run: ${RUN_URL}"
+            echo "- Source: \`${SOURCE_REF}\` at ${SOURCE_SHA}"
+            echo "- Target: \`${TARGET_BRANCH}\` at ${TARGET_SHA}"
+            echo
+            echo "## Conflicting files"
+            echo
+            echo "Output of \`git diff --name-only --diff-filter=U\` captured before \`git merge --abort\`:"
+            echo
+            sed 's/^/- /' "$RUNNER_TEMP/conflicts.txt"
+          } > "$RUNNER_TEMP/pr-body.md"
+          title="sync: merge ${SOURCE_REF} ${date_utc} (conflicts)"
+          existing=$(gh pr list --repo "$GITHUB_REPOSITORY" --base "$TARGET_BRANCH" --head "$sync_branch" --state open --json number --jq '.[0].number // empty')
+          if [[ -n "$existing" ]]; then
+            gh pr edit "$existing" --repo "$GITHUB_REPOSITORY" --title "$title" --body-file "$RUNNER_TEMP/pr-body.md"
+            pr_url=$(gh pr view "$existing" --repo "$GITHUB_REPOSITORY" --json url --jq .url)
+          else
+            pr_url=$(gh pr create --repo "$GITHUB_REPOSITORY" --base "$TARGET_BRANCH" --head "$sync_branch" --title "$title" --body-file "$RUNNER_TEMP/pr-body.md")
+          fi
+          echo "Conflicts in \`${SOURCE_REF}\` -> \`${TARGET_BRANCH}\`; pull request: ${pr_url}" >> "$GITHUB_STEP_SUMMARY"
+
   alert-on-failure:
     name: Alert on scheduled failure
     needs: [sync]
     # Same ci-8 mechanism audit.yml and coverage.yml use: a scheduled run that
-    # could not push or encountered merge conflicts must not stay visible only
-    # in the Actions tab.
+    # could not push or could not open the conflict pull request must not stay
+    # visible only in the Actions tab.
     if: github.repository == 'rustfs/rustfs' && (always() && github.event_name == 'schedule' && contains(needs.*.result, 'failure'))
     runs-on: ubuntu-latest
     timeout-minutes: 10

.github/workflows/helm-package.yml

@@ -31,7 +31,7 @@ permissions:
 
 jobs:
   build-helm-package:
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     if: |
       (github.event_name == 'workflow_dispatch' && !contains(github.event.inputs.version, '-preview')) ||
@@ -166,7 +166,7 @@ jobs:
           retention-days: 1
 
   publish-helm-package:
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     needs: [ build-helm-package ]
     if: needs.build-helm-package.result == 'success'

.github/workflows/issue-translator.yml

@@ -32,7 +32,7 @@ permissions:
 
 jobs:
   build:
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - uses: usthe/issues-translate-action@b41f55ddc81d7d54bd542a4f289fe28ec081898e # v2.7

.github/workflows/nightly-gnu.yml

@@ -45,7 +45,7 @@ jobs:
   resolve-source:
     if: github.repository == 'rustfs/rustfs'
     name: Resolve nightly source
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     outputs:
       source_sha: ${{ steps.source.outputs.sha }}
@@ -80,24 +80,13 @@ jobs:
       - name: Setup Rust environment
         uses: ./.github/actions/setup
         with:
-          cache-shared-key: build-x86_64-unknown-linux-gnu-glibc-2.31
+          cache-shared-key: build-x86_64-unknown-linux-gnu
           cache-save-if: 'false'
-          install-cross-tools: 'true'
-          target: x86_64-unknown-linux-gnu
           install-build-packaging-tools: 'false'
           install-test-tools: 'false'
 
       - name: Build RustFS
-        # Pin the glibc baseline (`<target>.<glibc>` for cargo-zigbuild) so
-        # nightlies run on older distros such as RHEL/Rocky 9 (glibc 2.34); #8062.
-        run: cargo zigbuild --release --locked --target x86_64-unknown-linux-gnu.2.31 -p rustfs --bins
-
-      - name: Verify glibc symbol baseline
-        run: |
-          python3 scripts/release/check_glibc.py 2.31 \
-            target/x86_64-unknown-linux-gnu/release/rustfs \
-            target/x86_64-unknown-linux-gnu/release/rustfs-cli
-          target/x86_64-unknown-linux-gnu/release/rustfs --version | grep -Fx 'build target : x86_64-unknown-linux-gnu'
+        run: cargo build --release --locked --target x86_64-unknown-linux-gnu -p rustfs --bins
 
       - name: Build DEB package
         id: deb

.github/workflows/nix-flake-update.yml

@@ -34,7 +34,7 @@ jobs:
   update-flake:
     if: github.repository == 'rustfs/rustfs'
     name: Update flake.lock
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 90
     env:
       FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"

.github/workflows/nix.yml

@@ -54,7 +54,7 @@ jobs:
   cancel-closed-pr-runs:
     name: Cancel Closed PR Runs
     if: github.event_name == 'pull_request' && github.event.action == 'closed'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - name: Explain cancellation run

.github/workflows/oidc-keycloak.yml

@@ -20,9 +20,8 @@ on:
       - ".github/workflows/oidc-keycloak.yml"
       - "crates/config/src/constants/oidc.rs"
       - "crates/iam/src/federation/**"
-      - "crates/iam/src/oidc/**"
+      - "crates/iam/src/oidc.rs"
       - "rustfs/src/admin/handlers/oidc.rs"
-      - "rustfs/src/admin/service/oidc_config.rs"
       - "rustfs/src/admin/handlers/sts.rs"
       - "scripts/test/oidc_keycloak_live.sh"
       - "scripts/test/fixtures/keycloak-rustfs-ci-realm.json"
@@ -32,9 +31,8 @@ on:
       - ".github/workflows/oidc-keycloak.yml"
       - "crates/config/src/constants/oidc.rs"
       - "crates/iam/src/federation/**"
-      - "crates/iam/src/oidc/**"
+      - "crates/iam/src/oidc.rs"
       - "rustfs/src/admin/handlers/oidc.rs"
-      - "rustfs/src/admin/service/oidc_config.rs"
       - "rustfs/src/admin/handlers/sts.rs"
       - "scripts/test/oidc_keycloak_live.sh"
       - "scripts/test/fixtures/keycloak-rustfs-ci-realm.json"

.github/workflows/on-demand-migration-interop.yml

@@ -186,7 +186,7 @@ jobs:
   cloud-source:
     name: ${{ matrix.provider }} source (three-case minimum)
     if: github.repository == 'rustfs/rustfs'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 45
     strategy:
       fail-fast: false

.github/workflows/package.yml

@@ -618,10 +618,8 @@ jobs:
   summary:
     name: Summary
     needs: [ resolve, package ]
-    # Every completed Build and Release run triggers this workflow; when
-    # resolve declined it (not a release tag), there is nothing to summarise.
-    if: always() && needs.resolve.result != 'skipped'
-    runs-on: ubuntu-latest
+    if: always()
+    runs-on: sm-standard-2
     timeout-minutes: 5
     steps:
       - name: Print summary

.github/workflows/performance-ab.yml

@@ -0,0 +1,384 @@
+# Copyright 2024 RustFS Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Warp A/B budget gate for the hotpath series (rustfs/backlog#935 HP-14, item 4).
+#
+# Two entry points, honestly scoped:
+#   * schedule (nightly, on main): post-merge detection — catches a regression
+#     within 24h of landing, not before merge.
+#   * workflow_dispatch: an explicitly selected trusted ref.
+# The dispatch input can run the gate with --allow-regression so a deliberate
+# correctness cost (e.g. the #4221 fsync durability fix) is recorded, not
+# blocked (rustfs/backlog#935 correction 1).
+
+name: Performance A/B
+
+on:
+  schedule:
+    - cron: "31 6 * * *" # 06:31 UTC nightly, against main
+  workflow_dispatch:
+    inputs:
+      duration:
+        description: "warp duration per round"
+        required: false
+        default: "12s"
+        type: string
+      allow_regression:
+        description: "Pass the gate despite a FAIL (deliberate tradeoff)"
+        required: false
+        default: false
+        type: boolean
+permissions:
+  actions: read
+  contents: read
+
+env:
+  CARGO_TERM_COLOR: always
+  RUST_BACKTRACE: 1
+
+jobs:
+  warp-ab:
+    if: github.repository == 'rustfs/rustfs'
+    name: Warp A/B budget gate
+    runs-on: sm-standard-2
+    # A normal nightly restores the last successful binary and builds only the
+    # candidate; daily access keeps that cache warm. A cache miss may build both
+    # and needs room for the A/B run plus artifact and cache publication.
+    timeout-minutes: 180
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
+        with:
+          persist-credentials: false
+          fetch-depth: 0 # baseline may be an earlier successful scheduled head
+
+      - name: Setup Rust environment
+        uses: ./.github/actions/setup
+        with:
+          rust-version: stable
+          cache-shared-key: warp-ab-${{ hashFiles('**/Cargo.lock') }}
+          cache-save-if: ${{ github.ref == 'refs/heads/main' }}
+
+      - name: Install warp
+        run: |
+          set -euo pipefail
+          WARP_VERSION="1.0.0"
+          curl -fsSL "https://github.com/minio/warp/releases/download/v${WARP_VERSION}/warp_Linux_x86_64.tar.gz" \
+            | sudo tar -xz -C /usr/local/bin warp
+          warp --version
+
+      - name: Decide exemption
+        id: exempt
+        env:
+          INPUT_ALLOW_REGRESSION: ${{ github.event.inputs.allow_regression }}
+        run: |
+          allow="false"
+          if [[ "$INPUT_ALLOW_REGRESSION" == "true" ]]; then
+            allow="true"
+          fi
+          echo "allow_regression=$allow" >> "$GITHUB_OUTPUT"
+
+      # A failed regression run must keep comparing against the last known-good
+      # scheduled head. Otherwise the next nightly would absorb the regression
+      # into its baseline and turn green without a fix.
+      - name: Find last successful scheduled baseline
+        id: scheduled_baseline
+        if: github.event_name == 'schedule'
+        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
+        with:
+          result-encoding: string
+          script: |
+            const { data } = await github.rest.actions.listWorkflowRuns({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              workflow_id: "performance-ab.yml",
+              event: "schedule",
+              status: "success",
+              per_page: 1,
+            });
+            return data.workflow_runs[0]?.head_sha ?? "";
+
+      # Manual runs compare a selected ref with current main. Scheduled runs
+      # compare current main with the last successful scheduled head. With no
+      # history, the first run measures the candidate against itself and seeds
+      # that head only if the complete rig succeeds.
+      - name: Resolve baseline / candidate commits
+        id: commits
+        env:
+          SCHEDULED_BASELINE_SHA: ${{ steps.scheduled_baseline.outputs.result }}
+        run: |
+          set -euo pipefail
+          candidate_sha="$(git rev-parse HEAD)"
+          if [[ "${{ github.event_name }}" == "schedule" ]]; then
+            baseline_sha="${SCHEDULED_BASELINE_SHA:-$candidate_sha}"
+          else
+            baseline_sha="$(git rev-parse origin/main)"
+          fi
+          git cat-file -e "${baseline_sha}^{commit}"
+          if ! git merge-base --is-ancestor "$baseline_sha" "$candidate_sha"; then
+            echo "::error::baseline $baseline_sha is not an ancestor of candidate $candidate_sha; update the selected ref before comparing" >&2
+            exit 1
+          fi
+          echo "baseline_sha=$baseline_sha" >> "$GITHUB_OUTPUT"
+          echo "candidate_sha=$candidate_sha" >> "$GITHUB_OUTPUT"
+          echo "baseline commit: $baseline_sha"
+          echo "candidate commit: $candidate_sha"
+
+      # Exact-key restore of the candidate binary saved by its successful
+      # scheduled run. A miss leaves cache-hit unset and falls back to a source
+      # build of that known-good head.
+      - name: Restore cached baseline binary
+        id: baseline_cache
+        uses: actions/cache/restore@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6
+        with:
+          path: baseline-bin/rustfs
+          key: rustfs-baseline-${{ steps.commits.outputs.baseline_sha }}
+
+      # Self-heal: on a nightly/dispatch run where the candidate commit IS the
+      # baseline commit, a cache miss would make the rig build the same commit
+      # twice (~65min per side with the post-#4806 LTO profile — no job budget
+      # fits that). Build it once here, reuse it for both phases, and save it
+      # back to the cache so the next run hits.
+      - name: Build baseline on cache miss (same-commit self-heal)
+        id: selfheal
+        if: >-
+          steps.baseline_cache.outputs.cache-hit != 'true' &&
+          steps.commits.outputs.baseline_sha == steps.commits.outputs.candidate_sha
+        run: |
+          set -euo pipefail
+          cargo build --release --bin rustfs
+          mkdir -p baseline-bin
+          cp target/release/rustfs baseline-bin/rustfs
+          echo "built=true" >> "$GITHUB_OUTPUT"
+
+      - name: Build baseline on cache miss (different candidate)
+        id: baseline_build
+        if: >-
+          steps.baseline_cache.outputs.cache-hit != 'true' &&
+          steps.commits.outputs.baseline_sha != steps.commits.outputs.candidate_sha
+        run: |
+          set -euo pipefail
+          baseline_root="$RUNNER_TEMP/rustfs-baseline-${{ github.run_id }}"
+          baseline_target="$RUNNER_TEMP/rustfs-baseline-target-${{ github.run_id }}"
+          git worktree add --detach "$baseline_root" "${{ steps.commits.outputs.baseline_sha }}"
+          cargo build --release --manifest-path "$baseline_root/Cargo.toml" --bin rustfs --target-dir "$baseline_target"
+          mkdir -p baseline-bin
+          cp "$baseline_target/release/rustfs" baseline-bin/rustfs
+          git worktree remove --force "$baseline_root"
+          echo "built=true" >> "$GITHUB_OUTPUT"
+
+      - name: Build candidate binary
+        id: candidate_build
+        if: steps.commits.outputs.baseline_sha != steps.commits.outputs.candidate_sha
+        run: |
+          set -euo pipefail
+          cargo build --release --bin rustfs
+          mkdir -p candidate-bin
+          cp target/release/rustfs candidate-bin/rustfs
+          echo "built=true" >> "$GITHUB_OUTPUT"
+
+      - name: Save self-healed baseline to cache
+        if: steps.selfheal.outputs.built == 'true' || steps.baseline_build.outputs.built == 'true'
+        uses: actions/cache/save@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6
+        with:
+          path: baseline-bin/rustfs
+          key: rustfs-baseline-${{ steps.commits.outputs.baseline_sha }}
+
+      - name: Run warp A/B and gate
+        id: ab
+        env:
+          INPUT_DURATION: ${{ github.event.inputs.duration }}
+        run: |
+          set -euo pipefail
+          # The formal runner executes A1 baseline -> B1 candidate -> B2 candidate
+          # -> A2 baseline for each workload and drive-sync cell. It requires three
+          # rounds per leg to emit tail latency and error-rate evidence.
+          # --health-timeout 180 outlasts the server's own 120s startup-readiness
+          # budget, which the rig's previous 60s health poll undershot (the first
+          # two nightly failures). perf-6 recalibrates these once the noise study
+          # lands.
+          duration="${INPUT_DURATION:-12s}"
+          baseline_sha="${{ steps.commits.outputs.baseline_sha }}"
+          candidate_sha="${{ steps.commits.outputs.candidate_sha }}"
+          baseline_hit="${{ steps.baseline_cache.outputs.cache-hit }}"
+          selfheal_built="${{ steps.selfheal.outputs.built }}"
+          baseline_built="${{ steps.baseline_build.outputs.built }}"
+          candidate_built="${{ steps.candidate_build.outputs.built }}"
+
+          args=(--duration "$duration" --rounds 3 --cooldown 5 --health-timeout 180 --baseline-revision "$baseline_sha" --candidate-revision "$candidate_sha")
+
+          if [[ "$baseline_hit" == "true" || "$selfheal_built" == "true" || "$baseline_built" == "true" ]]; then
+            chmod +x baseline-bin/rustfs
+            base_bin="$PWD/baseline-bin/rustfs"
+            args+=(--baseline-bin "$base_bin")
+            if [[ "$baseline_hit" == "true" ]]; then
+              base_src="actions-cache (rustfs-baseline-$baseline_sha)"
+            elif [[ "$selfheal_built" == "true" ]]; then
+              base_src="source build (cache self-heal, saved as rustfs-baseline-$baseline_sha)"
+            else
+              base_src="isolated baseline source build (saved as rustfs-baseline-$baseline_sha)"
+            fi
+            if [[ "$candidate_sha" == "$baseline_sha" ]]; then
+              # No commits landed since the last successful baseline, so reuse
+              # the one binary for both phases and measure only rig drift.
+              args+=(--candidate-bin "$base_bin")
+              cand_src="same binary as baseline (same commit)"
+            elif [[ "$candidate_built" == "true" ]]; then
+              chmod +x candidate-bin/rustfs
+              args+=(--candidate-bin "$PWD/candidate-bin/rustfs")
+              cand_src="source build of the checked-out ref"
+            else
+              echo "::error::candidate binary was not built" >&2
+              exit 2
+            fi
+          else
+            echo "::error::baseline binary was not restored or built" >&2
+            exit 2
+          fi
+
+          echo "baseline binary: $base_src"
+          echo "candidate binary: $cand_src"
+
+          if [[ "${{ steps.exempt.outputs.allow_regression }}" == "true" ]]; then
+            args+=(--allow-regression --exemption-reason "workflow dispatch override")
+          fi
+          # Do not let a gate FAIL abort the job here; capture status and surface
+          # it after the step summary is written.
+          set +e
+          bash scripts/run_hotpath_warp_abba.sh "${args[@]}"
+          echo "status=$?" >> "$GITHUB_OUTPUT"
+          set -e
+          # Locate the newest run dir + gate.md for the summary/comment/artifact
+          # steps. On a startup failure there is no gate.md, but the run dir still
+          # holds server-logs/ for diagnosis.
+          # Run dirs are UTC-timestamp names (no special chars); ls is safe here.
+          # shellcheck disable=SC2012
+          run_dir="$(ls -td target/hotpath-abba/*/ 2>/dev/null | head -n1 || true)"
+          echo "run_dir=${run_dir%/}" >> "$GITHUB_OUTPUT"
+          # shellcheck disable=SC2012
+          gate_md="$(ls -t target/hotpath-abba/*/candidate_gate.md 2>/dev/null | head -n1 || true)"
+          echo "gate_md=$gate_md" >> "$GITHUB_OUTPUT"
+
+      - name: Upload A/B results
+        if: always()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: hotpath-warp-ab-${{ github.run_number }}
+          # Includes per-cell median_summary.csv / baseline_compare.csv, both gates,
+          # and server-logs/ (rustfs.log + startup env per phase) so a failed run
+          # is diagnosable. Short retention: this is churny nightly debug data.
+          path: target/hotpath-abba/
+          if-no-files-found: warn
+          retention-days: 14
+
+      - name: Write gate summary
+        if: always()
+        run: |
+          set -euo pipefail
+          status="${{ steps.ab.outputs.status }}"
+          gate_md="${{ steps.ab.outputs.gate_md }}"
+          run_dir="${{ steps.ab.outputs.run_dir }}"
+          {
+            echo "## Hotpath warp A/B — run ${{ github.run_number }}"
+            echo
+            if [[ "$status" == "0" ]]; then
+              echo "Rig/gate exit: \`0\` (pass or warn)."
+            else
+              echo "Rig/gate exit: \`${status:-unknown}\` — **FAILED**."
+            fi
+            echo
+            if [[ -n "$gate_md" && -f "$gate_md" ]]; then
+              cat "$gate_md"
+            else
+              echo "No \`gate.md\` produced — the rig failed **before** the gate"
+              echo "(most likely server startup / health). Failing phase(s) below;"
+              echo "full logs in the \`hotpath-warp-ab-${{ github.run_number }}\` artifact."
+              if [[ -n "$run_dir" && -d "$run_dir/server-logs" ]]; then
+                for f in "$run_dir"/server-logs/*.log; do
+                  [[ -f "$f" ]] || continue
+                  echo
+                  echo "<details><summary>$(basename "$f")</summary>"
+                  echo
+                  echo '```'
+                  tail -n 30 "$f"
+                  echo '```'
+                  echo "</details>"
+                done
+              fi
+            fi
+          } >> "$GITHUB_STEP_SUMMARY"
+
+      - name: Stage successful candidate baseline
+        if: >-
+          steps.ab.outputs.status == '0' &&
+          steps.commits.outputs.baseline_sha != steps.commits.outputs.candidate_sha
+        run: |
+          set -euo pipefail
+          cp candidate-bin/rustfs baseline-bin/rustfs
+
+      - name: Cache successful candidate baseline
+        if: >-
+          steps.ab.outputs.status == '0' &&
+          steps.commits.outputs.baseline_sha != steps.commits.outputs.candidate_sha
+        uses: actions/cache/save@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6
+        with:
+          path: baseline-bin/rustfs
+          key: rustfs-baseline-${{ steps.commits.outputs.candidate_sha }}
+
+      # Scheduled failure alerting is handled by the alert-on-failure job below
+      # (perf-2 consuming ci-8's schedule-failure-issue composite action).
+
+      - name: Enforce gate
+        if: always()
+        run: |
+          status="${{ steps.ab.outputs.status }}"
+          if [[ -z "$status" ]]; then
+            echo "::error::warp A/B setup failed before the rig ran. Check the first failed workflow step." >&2
+            exit 1
+          fi
+          if [[ "$status" != "0" ]]; then
+            echo "::error::warp A/B budget gate failed (exit $status). See the step summary / gate.md artifact." >&2
+            exit "$status"
+          fi
+          echo "warp A/B budget gate passed."
+
+  alert-on-failure:
+    name: Alert on scheduled failure
+    needs: [warp-ab]
+    # `always()` is required: without it this job is skipped when a needed
+    # job fails. Alerts only for scheduled (nightly) runs (backlog#1149
+    # ci-8); manual dispatch failures are already watched by a human.
+    # `cancelled` is included alongside `failure` on purpose: a job that hits
+    # timeout-minutes ends as `cancelled`, and the 2026-07-11..07-14 nightly
+    # timeouts went silent precisely because the guard was failure-only. The
+    # composite action already reports cancelled/timed-out jobs in the issue
+    # body.
+    if: >-
+      github.repository == 'rustfs/rustfs' && (
+        always() && github.event_name == 'schedule' &&
+        (contains(needs.*.result, 'failure') || contains(needs.*.result, 'cancelled'))
+      )
+    runs-on: ubuntu-latest
+    timeout-minutes: 10
+    permissions:
+      contents: read
+      issues: write
+    steps:
+      - uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
+        with:
+          persist-credentials: false
+      - name: Open or update failure-tracking issue
+        uses: ./.github/actions/schedule-failure-issue
+        with:
+          github-token: ${{ secrets.BACKLOG_ISSUE_TOKEN }}

.github/workflows/rustfs-pool-expand-test.yml

@@ -279,9 +279,7 @@ jobs:
           if [ "${{ inputs.pools || '3' }}" = "3" ]; then
             STEPS="$STEPS,7,8"
             if [ "${{ inputs.run_decommission != 'false' }}" = "true" ]; then
-              # Step 10 asserts the surviving pools after the removal
-              # (rustfs#8726), so it only runs together with step 9.
-              STEPS="$STEPS,9,10"
+              STEPS="$STEPS,9"
             fi
           fi
           ARGS=(--steps "$STEPS" --with-warp -y \
@@ -295,10 +293,6 @@ jobs:
           if [ -n "${RUSTFS_POOL_PROXY_ENDPOINT}" ]; then
             ARGS+=(--proxy-endpoint "${RUSTFS_POOL_PROXY_ENDPOINT}")
           fi
-          # Finalize the removal so step 10 can exercise the surviving
-          # pools; passing it without step 9 selected would run step 10
-          # against a topology where nothing was removed.
-          case ",${STEPS}," in *,9,*) ARGS+=(--finalize-decommission) ;; esac
           if [ -n "${{ inputs.package_url }}" ]; then
             ARGS+=(--package-url "${{ inputs.package_url }}")
           elif [ -n "${{ inputs.rustfs_version }}" ]; then

.github/workflows/schedule-failure-alert-drill.yml

@@ -16,8 +16,8 @@
 #
 # Forces a job failure and then runs .github/actions/schedule-failure-issue
 # through the exact `needs` + `always() && contains(needs.*.result,
-# 'failure')` wiring used by the real consumers (e2e-s3tests, mint, fuzz).
-# Dispatch it twice to verify both the issue-creation and
+# 'failure')` wiring used by the real consumers (e2e-s3tests, mint, fuzz,
+# performance-ab). Dispatch it twice to verify both the issue-creation and
 # the dedupe-comment paths, then close the resulting
 # "[scheduled-failure] Schedule Failure Alert Drill" issue in rustfs/backlog
 # (the default target repository). Requires the BACKLOG_ISSUE_TOKEN secret
@@ -44,7 +44,7 @@ permissions:
 jobs:
   forced-failure:
     name: Forced failure
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 5
     steps:
       - name: Fail on purpose

.github/workflows/scheduled-validation-watchdog.yml

@@ -28,6 +28,7 @@ on:
       - "mint"
       - "minio-interop"
       - "Nightly GNU Build"
+      - "Performance A/B"
       - "Runner Hygiene"
     types: [completed]
 

.github/workflows/stale.yml

@@ -27,7 +27,7 @@ on:
 jobs:
   stale:
     if: github.repository == 'rustfs/rustfs'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 30
     steps:
       - uses: actions/stale@4391f3da665fdf50b6810c1a66712fb9ba21aa93 # v11.0.0

.github/workflows/star-history.yml

@@ -15,7 +15,7 @@ concurrency:
 jobs:
   update:
     if: github.repository == 'rustfs/rustfs'
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 10
     steps:
       - uses: overtrue/repo-visuals-action@ee2c632f6ce617e851fb46ea935ee8af762ebb93 # v1.4.0

.github/workflows/targets-integration.yml

@@ -42,7 +42,7 @@ jobs:
   targets-live:
     if: github.repository == 'rustfs/rustfs'
     name: PostgreSQL, MySQL, AMQP, and NATS
-    runs-on: ubuntu-latest
+    runs-on: sm-standard-2
     timeout-minutes: 90
     env:
       FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
@@ -60,7 +60,7 @@ jobs:
       - name: Setup Rust environment
         uses: ./.github/actions/setup
         with:
-          cache-shared-key: targets-live-lane-hosted
+          cache-shared-key: targets-live-lane
           cache-save-if: ${{ github.ref == 'refs/heads/main' || github.event_name == 'schedule' }}
           install-build-packaging-tools: 'false'
           install-test-tools: 'false'

.github/workflows/windows-filesystem.yml

@@ -19,21 +19,8 @@ on:
     branches: [ main ]
     paths:
       - "crates/ecstore/src/disk/**"
-      - "crates/ecstore/src/layout/**"
       - "crates/ecstore/src/store/init_format.rs"
-      - "crates/ecstore/src/set_disk/ops/locking.rs"
-      - "crates/ecstore/src/object_api/write_commit_context.rs"
-      - "crates/ecstore/src/bucket/utils.rs"
       - "crates/ecstore/Cargo.toml"
-      - "crates/utils/src/dunce.rs"
-      - "crates/utils/src/os/**"
-      - "crates/utils/Cargo.toml"
-      - "crates/utils/src/path.rs"
-      - "rustfs/build.rs"
-      - "rustfs/Cargo.toml"
-      - "rustfs/src/main.rs"
-      - "rustfs/src/server/runtime.rs"
-      - "rustfs/src/startup_*.rs"
       - "Cargo.toml"
       - "Cargo.lock"
       - ".github/actions/setup/**"
@@ -42,21 +29,8 @@ on:
     branches: [ main ]
     paths:
       - "crates/ecstore/src/disk/**"
-      - "crates/ecstore/src/layout/**"
       - "crates/ecstore/src/store/init_format.rs"
-      - "crates/ecstore/src/set_disk/ops/locking.rs"
-      - "crates/ecstore/src/object_api/write_commit_context.rs"
-      - "crates/ecstore/src/bucket/utils.rs"
       - "crates/ecstore/Cargo.toml"
-      - "crates/utils/src/dunce.rs"
-      - "crates/utils/src/os/**"
-      - "crates/utils/Cargo.toml"
-      - "crates/utils/src/path.rs"
-      - "rustfs/build.rs"
-      - "rustfs/Cargo.toml"
-      - "rustfs/src/main.rs"
-      - "rustfs/src/server/runtime.rs"
-      - "rustfs/src/startup_*.rs"
       - "Cargo.toml"
       - "Cargo.lock"
       - ".github/actions/setup/**"
@@ -103,122 +77,14 @@ jobs:
         shell: pwsh
         run: cargo test -p rustfs-ecstore --lib rename_all_ -- --nocapture
 
-      - name: Test multipart parent durability barriers
-        shell: pwsh
-        run: cargo test -p rustfs-ecstore --lib multipart_owned_record_parent_barriers_ -- --nocapture
-
-      - name: Test Windows path conversion
-        shell: pwsh
-        run: "cargo test -p rustfs-utils --no-default-features --features path --lib dunce:: -- --nocapture"
-
-      - name: Test Windows volume and disk-space probes
-        shell: pwsh
-        run: |
-          # Map a drive letter to the temp drive's administrative share so the
-          # remote-volume probe also covers mapped network drives.
-          $drive = (Split-Path -Qualifier ([System.IO.Path]::GetTempPath())).TrimEnd(':')
-          net use R: "\\localhost\$drive`$" /persistent:no
-          if ($LASTEXITCODE -eq 0) {
-            $env:RUSTFS_TEST_WINDOWS_MAPPED_DRIVE = 'R:'
-          } else {
-            Write-Output "::warning::could not map R: to \\localhost\$drive`$; mapped-drive test will skip"
-          }
-          cargo test -p rustfs-utils --features os --lib os:: -- --nocapture
-          $status = $LASTEXITCODE
-          if ($env:RUSTFS_TEST_WINDOWS_MAPPED_DRIVE) { net use R: /delete /y | Out-Null }
-          exit $status
-
       - name: Test Windows handle guards
         shell: pwsh
         run: cargo test -p rustfs-ecstore --lib windows_ -- --nocapture
 
-      - name: Test endpoint path parsing
-        shell: pwsh
-        run: "cargo test -p rustfs-ecstore --lib layout::endpoint:: -- --nocapture"
-
       - name: Test startup temporary-directory cleanup
         shell: pwsh
         run: cargo test -p rustfs-ecstore --lib cleanup_tmp_on_startup_ -- --nocapture
 
-      - name: Test object key Windows admission
-        shell: pwsh
-        run: "cargo test -p rustfs-ecstore --lib bucket::utils:: -- --nocapture"
-
-      - name: Test Windows file-name rules
-        shell: pwsh
-        run: "cargo test -p rustfs-utils --no-default-features --features path --lib dunce:: -- --nocapture"
-
       - name: Test fresh format publication
         shell: pwsh
         run: cargo test -p rustfs-ecstore --lib fresh_format_load_initializes_all_disks -- --nocapture
-
-  # The native Windows main thread only gets the 1 MiB PE stack reserve, while
-  # Linux and macOS hand out 8 MiB. Startup runs the whole root future on the
-  # thread that calls `block_on`, so a debug server once overflowed before it
-  # listened on Windows only. Start a real server to keep that covered.
-  startup-smoke:
-    name: Startup Smoke
-    runs-on: windows-latest
-    timeout-minutes: 90
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@f548e57e544e1ff5a4c46bf1e1b8685f8e4a348a # v7
-        with:
-          persist-credentials: false
-
-      - name: Setup Rust environment
-        uses: ./.github/actions/setup
-        with:
-          rust-version: stable
-          cache-shared-key: build-x86_64-pc-windows-msvc
-          cache-save-if: 'false'
-          install-build-packaging-tools: 'false'
-          install-test-tools: 'false'
-
-      - name: Build debug server binary
-        shell: pwsh
-        run: cargo build --locked -p rustfs --bin rustfs
-
-      - name: Print version
-        shell: pwsh
-        run: |
-          & .\target\debug\rustfs.exe --version
-          if ($LASTEXITCODE -ne 0) { throw "rustfs --version exited with $LASTEXITCODE" }
-
-      - name: Start server and wait for readiness
-        shell: pwsh
-        env:
-          RUSTFS_ACCESS_KEY: rustfsadmin
-          RUSTFS_SECRET_KEY: rustfsadmin
-        run: |
-          $volume = Join-Path $env:RUNNER_TEMP 'rustfs-startup-smoke'
-          $stdout = Join-Path $env:RUNNER_TEMP 'rustfs-startup-smoke.out.log'
-          $stderr = Join-Path $env:RUNNER_TEMP 'rustfs-startup-smoke.err.log'
-          New-Item -ItemType Directory -Force -Path $volume | Out-Null
-          $server = Start-Process -FilePath .\target\debug\rustfs.exe `
-            -ArgumentList @('--address', '127.0.0.1:9000', "`"$volume`"") `
-            -RedirectStandardOutput $stdout -RedirectStandardError $stderr `
-            -NoNewWindow -PassThru
-          # Cache the process handle now; otherwise ExitCode reads back empty.
-          $null = $server.Handle
-          $ready = $false
-          try {
-            for ($i = 0; $i -lt 180; $i++) {
-              if ($server.HasExited) { break }
-              try {
-                $response = Invoke-WebRequest -Uri 'http://127.0.0.1:9000/health/ready' -UseBasicParsing -NoProxy -TimeoutSec 2
-                if ($response.StatusCode -eq 200) { $ready = $true; break }
-              } catch {}
-              Start-Sleep -Seconds 1
-            }
-          } finally {
-            if (-not $server.HasExited) { Stop-Process -Id $server.Id -Force }
-            $server.WaitForExit()
-            Write-Host '--- server stdout (tail) ---'
-            Get-Content $stdout -Tail 100 -ErrorAction SilentlyContinue
-            Write-Host '--- server stderr (tail) ---'
-            Get-Content $stderr -Tail 100 -ErrorAction SilentlyContinue
-          }
-          if (-not $ready) {
-            throw ("rustfs did not report ready on /health/ready (exit code: {0})" -f $server.ExitCode)
-          }

.gitignore

@@ -90,4 +90,3 @@ worktrees/*
 # insta scratch files; the accepted .snap files ARE the assertions and are committed
 *.snap.new
 *.pending-snap
-.mimosa

.vscode/launch.json

@@ -7,9 +7,7 @@
             "name": "Debug(only) executable 'rustfs'",
             "env": {
               "RUST_LOG": "rustfs=info,ecstore=info,s3s=info,iam=info",
-              "RUSTFS_SKIP_BACKGROUND_TASK": "on",
-              // Local test volumes share one physical disk.
-              "RUSTFS_UNSAFE_BYPASS_DISK_CHECK": "true"
+              "RUSTFS_SKIP_BACKGROUND_TASK": "on"
               //"RUSTFS_OBS_LOG_DIRECTORY": "./deploy/logs",
               // "RUSTFS_POLICY_PLUGIN_URL":"http://localhost:8181/v1/data/rustfs/authz/allow",
               // "RUSTFS_POLICY_PLUGIN_AUTH_TOKEN":"your-opa-token"
@@ -46,7 +44,6 @@
             "env": {
                 "RUST_LOG": "rustfs=debug,ecstore=info,s3s=debug,iam=debug",
                 "RUSTFS_SKIP_BACKGROUND_TASK": "on",
-                "RUSTFS_UNSAFE_BYPASS_DISK_CHECK": "true",
                 //"RUSTFS_OBS_LOG_DIRECTORY": "./deploy/logs",
                 // "RUSTFS_POLICY_PLUGIN_URL":"http://localhost:8181/v1/data/rustfs/authz/allow",
                 // "RUSTFS_POLICY_PLUGIN_AUTH_TOKEN":"your-opa-token" 
@@ -120,8 +117,7 @@
             "env": {
                 "RUSTFS_ACCESS_KEY": "rustfsadmin",
                 "RUSTFS_SECRET_KEY": "rustfsadmin",
-                "RUSTFS_VOLUMES": "./target/volume/test{1...4}",
-                "RUSTFS_UNSAFE_BYPASS_DISK_CHECK": "true",
+                //"RUSTFS_VOLUMES": "./target/volume/test{1...4}",
                 "RUSTFS_ADDRESS": ":9000",
                 "RUSTFS_CONSOLE_ENABLE": "true",
                 // "RUSTFS_OBS_TRACE_ENDPOINT": "http://127.0.0.1:4318/v1/traces", // jeager otlp http endpoint
@@ -139,31 +135,42 @@
         {
             "type": "lldb",
             "request": "launch",
-            "name": "Debug test_transition_and_restore_flows",
+            "name": "Debug test_lifecycle_transition_basic",
             "cargo": {
                 "args": [
                     "test",
-                    "--no-run",
                     "-p",
                     "rustfs-scanner",
                     "--test",
                     "lifecycle_integration_test",
+                    "serial_tests::test_lifecycle_transition_basic",
                     "-j",
                     "1"
-                ],
-                "filter": {
-                    "name": "lifecycle_integration_test",
-                    "kind": "test"
-                }
+                ]
             },
-            "args": [
-                "serial_tests::test_transition_and_restore_flows",
-                "--exact",
-                "--ignored",
-                "--test-threads=1"
-            ],
+            "args": [],
             "cwd": "${workspaceFolder}"
         },
+        {
+            "name": "Debug executable target/debug/test",
+            "type":  "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/target/debug/deps/lifecycle_integration_test-5915cbfcab491b3b",
+            "args": [
+              "--skip",
+              "test_lifecycle_expiry_basic",
+              "--skip",
+              "test_lifecycle_expiry_deletemarker",
+              //"--skip",
+              //"test_lifecycle_transition_basic",
+            ],
+            "cwd": "${workspaceFolder}",
+            //"stopAtEntry": false,
+            //"preLaunchTask": "cargo build",
+            "sourceLanguages": [
+                "rust"
+            ],
+        },
         {
             "name": "Debug executable target/debug/rustfs with sse kms",
             "type":  "lldb",
@@ -278,7 +285,7 @@
                 "RUSTFS_COMPRESSION_ENABLED": "true",
                 "RUSTFS_KMS_ENABLE": "true",
                 "RUSTFS_KMS_BACKEND": "local",
-                "RUSTFS_KMS_KEY_DIR": "${workspaceFolder}/target/rio-v2/kms-key-dir",
+                "RUSTFS_KMS_KEY_DIR": "./target/rio-v2/kms-key-dir",
                 "RUSTFS_KMS_LOCAL_MASTER_KEY": "my-secret-key", // Some Password
                 "RUSTFS_KMS_DEFAULT_KEY_ID": "rustfs-master-key",
             },
@@ -300,7 +307,6 @@
                 "RUSTFS_ACCESS_KEY": "rustfsadmin",
                 "RUSTFS_SECRET_KEY": "rustfsadmin",
                 "RUSTFS_VOLUMES": "./target/e2e-test/test{1...4}",
-                "RUSTFS_UNSAFE_BYPASS_DISK_CHECK": "true",
                 "RUSTFS_REGION": "us-east-1",
                 "RUSTFS_ADDRESS": ":9000",
                 "RUSTFS_CONSOLE_ENABLE": "true",
@@ -309,7 +315,7 @@
 
                 "RUSTFS_KMS_ENABLE": "true",
                 "RUSTFS_KMS_BACKEND": "local",
-                "RUSTFS_KMS_KEY_DIR": "${workspaceFolder}/target/e2e-key-dir",
+                "RUSTFS_KMS_KEY_DIR": "./target/e2e-key-dir",
                 "RUSTFS_KMS_LOCAL_MASTER_KEY": "my-secret-key", // Some Password
                 "RUSTFS_KMS_DEFAULT_KEY_ID": "rustfs-master-key",
             },

AGENTS.md

@@ -3,9 +3,6 @@
 This file contains repository-wide rules. Use the nearest subdirectory
 `AGENTS.md` for path-specific invariants.
 
-S3-compatible object store in Rust, derived from MinIO. Erasure-coded,
-multi-pool, supports ILM tiering/lifecycle.
-
 ## Precedence
 
 1. System/developer instructions.
@@ -41,27 +38,6 @@ rules. A skill cannot expand the user's requested scope or grant authorization.
   pending. Before requesting missing approval, prepare the concrete result that
   is already authorized; retain explicit merge and release gates.
 
-- Keep individual tool outputs concise. Save large outputs to task-owned logs
-  and retrieve only the relevant ranges.
-
-## Common Commands
-
-```bash
-cargo build --release --bin rustfs   # production binary
-cargo check -p <crate>               # fast type-check one crate
-cargo test -p <crate>                # test one crate
-cargo fmt --all --check              # for Rust changes; see AGENTS.md verification tiers
-make pre-commit                      # fast gate: fmt + arch checks + quick-check (NO clippy/tests)
-make pre-pr                          # optional full gate for broad cross-module changes
-make build-docker BUILD_OS=ubuntu22.04
-```
-
-> **Docker build note**: `buildx build` without `--load` keeps the image in
-> the buildx cache only — `docker run` will use a stale local image. The
-> Makefile already includes `--load`; if you suspect a stale binary, add
-> `--no-cache` to the `buildx build` invocation inside
-> `.config/make/build-docker.mak`.
-
 ## Task-Specific Guidance
 
 Read only the reference needed for the current task, once per unchanged context:
@@ -89,19 +65,6 @@ or agent scratch notes. Durable architecture belongs under `docs/architecture/`,
 operations under `docs/operations/`, and testing references under
 `docs/testing/`. `scripts/check_no_planning_docs.sh` enforces this boundary.
 
-## Further References
-
-- Agent knowledge base index and doc-writing rules: [docs/architecture/README.md](docs/architecture/README.md)
-- Crate membership: `Cargo.toml` `[workspace].members`
-- Architecture, layering, crate map: [ARCHITECTURE.md](ARCHITECTURE.md)
-- Migration guardrails & readiness contracts: [docs/architecture/](docs/architecture/README.md)
-- CI workflow steps: `.github/workflows/`; event, timeout, and required-status
-  matrix: [docs/testing/ci-gates.md](docs/testing/ci-gates.md)
-- Test-layer taxonomy, per-layer entry commands, serial/nextest rules, flake
-  policy: [docs/testing/README.md](docs/testing/README.md)
-- Tier/ILM transition debugging (xl.meta inspection, versionId tracing):
-  [docs/operations/tier-ilm-debugging.md](docs/operations/tier-ilm-debugging.md)
-
 ## Verification
 
 Select checks from the final task-owned diff. Scoped `AGENTS.md` files may add a

CHANGELOG.md

@@ -19,17 +19,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Single-part source checksums are forwarded as `x-amz-checksum-*` headers instead of user metadata, so the replica preserves checksum responses ([#7313](https://github.com/rustfs/rustfs/pull/7313)). This changes the default outbound headers for checksummed objects.
 - Site-replication outage recovery now uses a bounded 30-second retry drain plus the 600-second full reconciliation pass, persists destructive liabilities before local deletion, and fences replay settlement and peer edits ([#7148](https://github.com/rustfs/rustfs/pull/7148)). Persisted additions are optional and ignored by older readers.
 - IAM snapshot/deletion replay, target-assigned delete-marker purges, timestamp ordering, and best-effort peer broadcast now close the control-plane gaps found by the R6 review ([#7195](https://github.com/rustfs/rustfs/pull/7195)).
-- A `set-remote-target?update=true&creds=true` request on a site-replication peer's bucket target is now rejected with `400 InvalidRequest`, naming the target and why it counts as peer-owned. It used to return success while keeping the stored endpoint, credentials, and TLS settings, so an endpoint edit from the Console's bucket replication form appeared to apply but did not. Change a peer's address with a site-replication edit instead; see "Changing a site's address" in `docs/operations/site-replication-operations.md`. The `sync`, `bandwidth`, `proxy`, and `path` update groups are unchanged.
-- Site-replication edits that omit the site name, such as `mc admin replicate update --deployment-id <ID> --endpoint <URL>`, now keep the stored name on the editing site and on every receiver instead of renaming the site after the new endpoint host. No persisted format changes.
 - Upgrade and rollback: upgrade every node in one site consecutively and verify reconciliation before moving to the next site; do not intentionally run a site mixed-version. Target-version ledger keys are harmless on rollback, although old code cannot use their routing. Before rolling back past [#7307](https://github.com/rustfs/rustfs/pull/7307), drain or repair every pending version purge: older code can free a retained version's data directory before its remote purge is acknowledged. See `docs/operations/site-replication-operations.md`.
 
 ### Security
 - **Presigned URLs honour only signed headers** (GHSA-g8w9-qw9q-fghr): a SigV4 presigned request that carries an `x-amz-*` request header not listed in `X-Amz-SignedHeaders` is now rejected with `403 AccessDenied` ("There were headers present in the request which were not signed"), matching AWS S3. Previously the holder of a presigned `PutObject` URL could add unsigned `x-amz-tagging`, `x-amz-storage-class`, `x-amz-website-redirect-location`, ACL, metadata, Object Lock or SSE headers and have them applied. Presigners that intend a property must set it before signing so the SDK lists the header in `SignedHeaders`; `x-amz-cf-id` (CloudFront) remains tolerated unsigned. Header-signed SigV4 and SigV2 requests are unchanged.
 
 ### Fixed
-- **`mc admin info` and `mc admin replicate resync` work again**: since the cluster snapshot view (#3803), `GET /v3/info` and `GET /v3/storageinfo` returned the madmin `InfoMessage` / `StorageInfo` nested under an `info` key next to `admin_discovery` (and, for `/info`, `bitrotSelftest`). madmin decodes those bodies directly, so `mc admin info` printed empty values and `mc admin replicate resync start/status` rejected every alias with "alias provided is not part of cluster replication" because the local `deploymentID` read as empty. Requests on the MinIO-compatible `/minio/admin/v3` prefix now receive the bare madmin body again; `/rustfs/admin/v3`, which rc and the console use, keeps the envelope unchanged.
-- **Replication status writebacks no longer invalidate scanner cycles**: every replication status writeback (`persist_replication_state_if_current` in `Update` mode, including the heals the scanner queues during its own walk) marked the bucket dirty on the namespace path, advanced the dirty-usage generation and woke the scanner, so a site-replication resync or any steadily replicated bucket ended every cycle `superseded`, the authoritative usage baseline was never certified, and the console reported usage freshness as stale for as long as replication ran (#8375). Those writebacks are now status class: they do not touch the dirty-usage generation, the activity snapshot, the dirty-usage wake or the durable dirty journal, and only the process-local producer coverage is recorded. They need no rescan queue, because a deployment with an active replication rule already walks every bucket on every cycle and that walk refreshes `replication_stats`. Tier-expiration cleanup, which removes data, stays on the namespace path. No persisted format, protocol, admin API or configuration changes. Design: `docs/architecture/scanner-per-bucket-convergence.md`.
-- **A bucket written during a scanner walk no longer supersedes the whole cycle**: the post-walk cluster activity proof required the entire activity snapshot and the dirty-usage generation to be unchanged, so a single-node deployment with one bucket written every few seconds never reached a `Complete` cycle, `usage_snapshot_converged` never became true and the authoritative usage baseline was never certified (#8494). Dirty usage is now certified per bucket: a structurally complete walk that the guardrail did not veto is `Complete`, buckets whose dirty generation still matches the start snapshot are cleared, and buckets written during the walk stay dirty and are published as the pending set, so `usage_snapshot_converged` means "no bucket pending" and the next cycle is scoped to the pending buckets. Such a cycle writes the authoritative usage object; the observed object is now selected by an explicit observational mark that vetoed cycles carry. The guardrail itself vetoes only on the fields that change which namespace the walk described (peer set, process instance, protocol version, topology, maintenance generation, data movement, movement generation, publication blocked) and reports the field and host in the scanner log and in the new `usage_freshness` fields `guardrail_veto_total`, `last_guardrail_veto_unix_secs` and `last_guardrail_veto_reason`. In a distributed deployment a remote peer whose dirty usage moved during the walk still supersedes the cycle until scoped acknowledgements can name its buckets. `.usage.v2.json` gains the additive, default-when-absent fields `usage_snapshot_pending_buckets`, `usage_snapshot_pending_bucket_count` and `usage_snapshot_observational`; older readers ignore them and treat the snapshot as an authoritative but unconverged publication. Quota is unchanged: an uncertified bucket keeps its last usage value as the floor and live writes do not advance it. No protocol or configuration changes. Design: `docs/architecture/scanner-per-bucket-convergence.md`.
 - **SSE-C TLS requirement now covers copy-source keys**: with `RUSTFS_SSE_C_REQUIRE_TLS=true`, a `CopyObject` or `UploadPartCopy` that carried only the `x-amz-copy-source-server-side-encryption-customer-*` headers (an SSE-C source copied to a destination that is not SSE-C) was accepted on a plaintext transport and was not counted in `rustfs_ssec_plaintext_requests_total`, although those headers carry the source object's customer key. They now count as SSE-C headers: such requests are counted, and refused with `400 InvalidRequest` when the switch is on. Requests over TLS are unaffected.
 - **SSE-KMS encryption context could replace the location entry**: data keys are wrapped under an encryption-context entry, keyed by the bucket name, that ties them to the object's location. A client `x-amz-server-side-encryption-context` entry under the same key replaced it, and because the client context is stored with the object, the replacement also applied on every read. SSE-KMS writes now refuse such a context with `400 InvalidArgument`; other keys, including the names of other buckets, are unaffected. Objects stored with such an entry by an earlier release remain readable. See `docs/operations/kms-backend-security.md`.
 - **Two admin handlers reported a missing policy or user as `500 InternalError`**: `GET /rustfs/admin/v3/info-canned-policy` and `DELETE /rustfs/admin/v3/remove-user` wrapped every IAM error as an internal fault instead of classifying it, so a lookup for something that does not exist answered `500 InternalError` where the rest of the admin API answers `404 NoSuchResource` — `GET /rustfs/admin/v3/user-info` already did, and `iam_error_to_s3_error` already mapped `NoSuchPolicy`, `NoSuchUser`, `NoSuchServiceAccount` and `NoSuchTempAccount` to `NoSuchResource`. A client could not tell "does not exist" from a real server failure, and a controller that checks existence before creating (a Kubernetes operator reconciling policies or users) could never create the object: the missing-policy lookup failed with a `500` and it retried forever. Both handlers now map their IAM errors through the shared `iam_error_to_s3_error` helper, so a missing policy or user is `404 NoSuchResource` and the `error` is preserved as the S3 error's source; the three IAM calls in the removal path (temporary-user lookup, service-account lookup, delete) are all reclassified, and every other IAM error stays `500 InternalError`. Fixes #8117.
@@ -90,7 +85,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Four-layer regression-prevention tests guard against silent feature deletion: compile-time module assertion, module-presence unit test, cross-module `Protocol` enum assertion, end-to-end SSH banner test against the running binary
 
 ### Changed
-- **Scanner bucket walkers are no longer tied to the disk count**: a new scanner key `walkers_per_disk` (`RUSTFS_SCANNER_WALKERS_PER_DISK`, `1` through `16`, default `1`) spawns that many bucket walkers per local disk of an erasure set, so a single-disk deployment can walk several buckets at once instead of one bucket, one object at a time. The walkers of one disk share its scanner admission, so a peer still sees one scan session per physical disk, and remote disks keep one walker. The walker count is what `max_concurrent_disk_scans` caps and the disk count stays the floor; the `disk_concurrency_budget` log line now reports `walkers_per_disk`, `walkers`, and the `reason` for the effective limit. The default keeps today's I/O profile because a local disk cannot yet report whether it is rotational; the design intends a higher default on flash once it can (`docs/architecture/scanner-per-bucket-convergence.md`). As groundwork for pacing, the raw-enumeration page writer inserts into its building page in place and commits pages without copying them, and a regression test pins the cost of recording 20,000 entries in one prefix so the quadratic bookkeeping removed by [#8114](https://github.com/rustfs/rustfs/pull/8114) (rustfs#8113) cannot return. No persisted format, protocol or admin API changes beyond the additional config key, which older releases preserve and ignore.
 - **Encryption and KMS work merged since `1.0.0-rc.5`** (entries were missing from this section):
   - **Persisted KMS configuration secrets** are sealed field-by-field with `RUSTFS_KMS_CONFIG_SECRET`. **When the variable is unset the secrets are persisted in cleartext and the server only warns** (`persisted KMS configuration carries cleartext secrets`); it never refuses the write. Set it, identically, on every node, and re-save the configuration to seal an existing one.
   - **New v2 ciphertext frame format** with per-frame index binding and final-frame authentication. Its write switch `RUSTFS_ENCRYPTION_FRAME_V2` is **off by default**: v2 frames are unreadable by nodes without v2 read support, and encrypted ciphertext travels verbatim through transition, decommission and SSE-C replication passthrough, so turn it on only after every node — and every RustFS warm/replication target that receives raw ciphertext — runs a release with v2 read support. Reading v2 objects needs no switch.

CLAUDE.md

@@ -0,0 +1,46 @@
+# RustFS — CLAUDE.md
+
+S3-compatible object store in Rust, derived from MinIO. Erasure-coded,
+multi-pool, supports ILM tiering/lifecycle.
+
+Agent rules (change style, verification, PR conventions, scoped per-crate
+guidance) live in [AGENTS.md](AGENTS.md) — follow it. This file only adds
+what Claude Code needs on top: commands and pointers.
+
+## Commands
+
+```bash
+cargo build --release --bin rustfs   # production binary
+cargo check -p <crate>               # fast type-check one crate
+cargo test -p <crate>                # test one crate
+cargo fmt --all --check              # for Rust changes; see AGENTS.md verification tiers
+make pre-commit                      # fast gate: fmt + arch checks + quick-check (NO clippy/tests)
+make pre-pr                          # optional full gate for broad cross-module changes
+make build-docker BUILD_OS=ubuntu22.04
+```
+
+> **Docker build note**: `buildx build` without `--load` keeps the image in
+> the buildx cache only — `docker run` will use a stale local image. The
+> Makefile already includes `--load`; if you suspect a stale binary, add
+> `--no-cache` to the `buildx build` invocation inside
+> `.config/make/build-docker.mak`.
+
+## Where to look (do not duplicate here)
+
+- Agent knowledge base index and doc-writing rules: [docs/architecture/README.md](docs/architecture/README.md)
+- Crate membership: `Cargo.toml` `[workspace].members`
+- Architecture, layering, crate map: [ARCHITECTURE.md](ARCHITECTURE.md)
+- Migration guardrails & readiness contracts: [docs/architecture/](docs/architecture/README.md)
+- CI workflow steps: `.github/workflows/`; event, timeout, and required-status
+  matrix: [docs/testing/ci-gates.md](docs/testing/ci-gates.md)
+- Test-layer taxonomy, per-layer entry commands, serial/nextest rules, flake
+  policy: [docs/testing/README.md](docs/testing/README.md)
+- Tier/ILM transition debugging (xl.meta inspection, versionId tracing):
+  [docs/operations/tier-ilm-debugging.md](docs/operations/tier-ilm-debugging.md)
+
+## Domain conventions worth knowing up front
+
+Repo-wide domain invariants (dual internal metadata keys, defensive UUID
+reads, unversioned tier buckets) live in [AGENTS.md](AGENTS.md) under
+"Cross-Cutting Storage Invariants" — read them before touching metadata or
+tiering code.

Cargo.toml

@@ -18,6 +18,7 @@ members = [
     "crates/audit", # Audit target management system with multi-target fan-out
     "crates/checksums", # client checksums
     "crates/common", # Shared utilities and data structures
+    "crates/capture-authority", # Durable capture authority storage
     "crates/data-usage", # Shared data usage models and algorithms
     "crates/config", # Configuration management
     "crates/connect-symbol-catalog", # Release-only CPU profile symbol catalog generator
@@ -66,7 +67,6 @@ members = [
     "crates/io-metrics", # I/O metrics collection for performance analysis
     "crates/io-core", # Buffered Bytes, mmap-then-copy, and aligned pread I/O helpers
     "crates/zip", # ZIP file handling and compression
-    "crates/zstd-compat", # Fixed codec for dependencies requiring zstd 0.13
 ]
 resolver = "3"
 
@@ -75,7 +75,7 @@ edition = "2024"
 license = "Apache-2.0"
 repository = "https://github.com/rustfs/rustfs"
 rust-version = "1.99.0"
-version = "1.0.2"
+version = "1.0.1"
 homepage = "https://rustfs.com"
 description = "RustFS is a high-performance distributed object storage software built using Rust, one of the most popular languages worldwide. "
 keywords = ["RustFS", "Minio", "object-storage", "filesystem", "s3"]
@@ -92,62 +92,62 @@ redundant_clone = "warn"
 
 [workspace.dependencies]
 # RustFS Internal Crates
-rustfs = { path = "./rustfs", version = "1.0.2" }
-rustfs-heal = { path = "crates/heal", version = "1.0.2" }
-rustfs-heal-contracts = { path = "crates/heal-contracts", version = "1.0.2" }
-rustfs-scanner-metrics = { path = "crates/scanner-metrics", version = "1.0.2" }
-rustfs-audit = { path = "crates/audit", version = "1.0.2" }
-rustfs-checksums = { path = "crates/checksums", version = "1.0.2" }
-rustfs-common = { path = "crates/common", version = "1.0.2" }
-rustfs-data-usage = { path = "crates/data-usage", version = "1.0.2" }
-rustfs-config = { path = "./crates/config", version = "1.0.2" }
-rustfs-concurrency = { path = "./crates/concurrency", version = "1.0.2" }
-rustfs-credentials = { path = "crates/credentials", version = "1.0.2" }
-rustfs-crypto = { path = "crates/crypto", version = "1.0.2" }
-rustfs-ecstore = { path = "crates/ecstore", version = "1.0.2" }
-rustfs-filemeta = { path = "crates/filemeta", version = "1.0.2" }
-rustfs-iam = { path = "crates/iam", version = "1.0.2" }
-rustfs-keystone = { path = "crates/keystone", version = "1.0.2" }
-rustfs-license = { path = "crates/license", version = "1.0.2" }
-rustfs-lifecycle = { path = "crates/lifecycle", version = "1.0.2" }
-rustfs-kms = { path = "crates/kms", version = "1.0.2" }
-rustfs-lock = { path = "crates/lock", version = "1.0.2" }
-rustfs-madmin = { path = "crates/madmin", version = "1.0.2" }
-rustfs-notify = { path = "crates/notify", version = "1.0.2" }
-rustfs-io-metrics = { path = "crates/io-metrics", version = "1.0.2" }
-rustfs-io-core = { path = "crates/io-core", version = "1.0.2" }
-rustfs-object-capacity = { path = "crates/object-capacity", version = "1.0.2" }
-rustfs-object-data-cache = { path = "crates/object-data-cache", version = "1.0.2", default-features = false }
-rustfs-log-analyzer = { path = "crates/log-analyzer", version = "1.0.2" }
-rustfs-obs = { path = "crates/obs", version = "1.0.2" }
-rustfs-policy = { path = "crates/policy", version = "1.0.2" }
-rustfs-protos = { path = "crates/protos", version = "1.0.2" }
-rustfs-protocols = { path = "crates/protocols", version = "1.0.2" }
-rustfs-replication = { path = "crates/replication", version = "1.0.2" }
-rustfs-rio = { path = "crates/rio", version = "1.0.2" }
-rustfs-rio-v2 = { path = "crates/rio-v2", version = "1.0.2" }
-rustfs-s3-client = { path = "crates/s3-client", version = "1.0.2" }
-rustfs-s3-types = { path = "crates/s3-types", version = "1.0.2" }
-rustfs-s3-ops = { path = "crates/s3-ops", version = "1.0.2" }
-rustfs-s3select-api = { path = "crates/s3select-api", version = "1.0.2" }
-rustfs-s3select-query = { path = "crates/s3select-query", version = "1.0.2" }
-rustfs-scanner = { path = "crates/scanner", version = "1.0.2" }
-rustfs-security-governance = { path = "crates/security-governance", version = "1.0.2" }
-rustfs-extension-schema = { path = "crates/extension-schema", version = "1.0.2" }
-rustfs-signer = { path = "crates/signer", version = "1.0.2" }
-rustfs-storage-api = { path = "crates/storage-api", version = "1.0.2" }
-rustfs-trusted-proxies = { path = "crates/trusted-proxies", version = "1.0.2" }
-rustfs-targets = { path = "crates/targets", version = "1.0.2" }
-rustfs-test-utils = { path = "crates/test-utils", version = "1.0.2" }
-rustfs-tls-runtime = { path = "crates/tls-runtime", version = "1.0.2" }
-rustfs-utils = { path = "crates/utils", version = "1.0.2" }
-rustfs-zip = { path = "./crates/zip", version = "1.0.2" }
+rustfs = { path = "./rustfs", version = "1.0.1" }
+rustfs-heal = { path = "crates/heal", version = "1.0.1" }
+rustfs-heal-contracts = { path = "crates/heal-contracts", version = "1.0.1" }
+rustfs-scanner-metrics = { path = "crates/scanner-metrics", version = "1.0.1" }
+rustfs-audit = { path = "crates/audit", version = "1.0.1" }
+rustfs-checksums = { path = "crates/checksums", version = "1.0.1" }
+rustfs-common = { path = "crates/common", version = "1.0.1" }
+rustfs-data-usage = { path = "crates/data-usage", version = "1.0.1" }
+rustfs-config = { path = "./crates/config", version = "1.0.1" }
+rustfs-concurrency = { path = "./crates/concurrency", version = "1.0.1" }
+rustfs-credentials = { path = "crates/credentials", version = "1.0.1" }
+rustfs-crypto = { path = "crates/crypto", version = "1.0.1" }
+rustfs-ecstore = { path = "crates/ecstore", version = "1.0.1" }
+rustfs-filemeta = { path = "crates/filemeta", version = "1.0.1" }
+rustfs-iam = { path = "crates/iam", version = "1.0.1" }
+rustfs-keystone = { path = "crates/keystone", version = "1.0.1" }
+rustfs-license = { path = "crates/license", version = "1.0.1" }
+rustfs-lifecycle = { path = "crates/lifecycle", version = "1.0.1" }
+rustfs-kms = { path = "crates/kms", version = "1.0.1" }
+rustfs-lock = { path = "crates/lock", version = "1.0.1" }
+rustfs-madmin = { path = "crates/madmin", version = "1.0.1" }
+rustfs-notify = { path = "crates/notify", version = "1.0.1" }
+rustfs-io-metrics = { path = "crates/io-metrics", version = "1.0.1" }
+rustfs-io-core = { path = "crates/io-core", version = "1.0.1" }
+rustfs-object-capacity = { path = "crates/object-capacity", version = "1.0.1" }
+rustfs-object-data-cache = { path = "crates/object-data-cache", version = "1.0.1", default-features = false }
+rustfs-log-analyzer = { path = "crates/log-analyzer", version = "1.0.1" }
+rustfs-obs = { path = "crates/obs", version = "1.0.1" }
+rustfs-policy = { path = "crates/policy", version = "1.0.1" }
+rustfs-protos = { path = "crates/protos", version = "1.0.1" }
+rustfs-protocols = { path = "crates/protocols", version = "1.0.1" }
+rustfs-replication = { path = "crates/replication", version = "1.0.1" }
+rustfs-rio = { path = "crates/rio", version = "1.0.1" }
+rustfs-rio-v2 = { path = "crates/rio-v2", version = "1.0.1" }
+rustfs-s3-client = { path = "crates/s3-client", version = "1.0.1" }
+rustfs-s3-types = { path = "crates/s3-types", version = "1.0.1" }
+rustfs-s3-ops = { path = "crates/s3-ops", version = "1.0.1" }
+rustfs-s3select-api = { path = "crates/s3select-api", version = "1.0.1" }
+rustfs-s3select-query = { path = "crates/s3select-query", version = "1.0.1" }
+rustfs-scanner = { path = "crates/scanner", version = "1.0.1" }
+rustfs-security-governance = { path = "crates/security-governance", version = "1.0.1" }
+rustfs-extension-schema = { path = "crates/extension-schema", version = "1.0.1" }
+rustfs-signer = { path = "crates/signer", version = "1.0.1" }
+rustfs-storage-api = { path = "crates/storage-api", version = "1.0.1" }
+rustfs-trusted-proxies = { path = "crates/trusted-proxies", version = "1.0.1" }
+rustfs-targets = { path = "crates/targets", version = "1.0.1" }
+rustfs-test-utils = { path = "crates/test-utils", version = "1.0.1" }
+rustfs-tls-runtime = { path = "crates/tls-runtime", version = "1.0.1" }
+rustfs-utils = { path = "crates/utils", version = "1.0.1" }
+rustfs-zip = { path = "./crates/zip", version = "1.0.1" }
 
 # Async Runtime and Networking
 async-channel = "2.5.0"
 async_zip = { default-features = false, version = "0.0.19" }
 mysql_async = { default-features = false, version = "0.37.1" }
-async-compression = { version = "0.4.51" }
+async-compression = { version = "0.4.50" }
 async-trait = "0.1.92"
 async-nats = { version = "0.50.0", default-features = false }
 axum = "0.8.9"
@@ -156,10 +156,9 @@ futures-core = "0.3.34"
 futures-lite = "2.6.1"
 futures-util = "0.3.34"
 pollster = "1.0.1"
-pulsar = { default-features = false, version = "=6.9.0" }
-lapin = { default-features = false, version = "4.12.3" }
+pulsar = { default-features = false, version = "6.9.0" }
+lapin = { default-features = false, version = "4.12.0" }
 hyper = { version = "1.12.0" }
-h2 = { version = "0.4.20" }
 hyper-rustls = { default-features = false, version = "0.27.10" }
 hyper-util = { version = "0.1.21" }
 http = "1.5.0"
@@ -170,11 +169,11 @@ minlz = "1.2.3"
 reqwest = "0.13.5"
 rustfs-kafka-async = { version = "1.4.0" }
 socket2 = { version = "0.6.5" }
-tokio = { version = "1.54.1" }
+tokio = { version = "1.53.2" }
 tokio-rustls = { default-features = false, version = "0.26.6" }
 tokio-stream = { version = "0.1.19" }
 tokio-test = "0.4.6"
-tokio-util = { version = "0.7.20" }
+tokio-util = { version = "0.7.19" }
 tonic = { version = "0.14.6" }
 tonic-prost = { version = "0.14.6" }
 tonic-prost-build = { version = "0.14.6" }
@@ -182,7 +181,7 @@ tower = { version = "0.5.3" }
 tower-http = { version = "0.7.1" }
 h3 = "0.0.8"
 h3-quinn = "0.0.10"
-quinn = { version = "0.11.12", default-features = false, features = [
+quinn = { version = "0.11.11", default-features = false, features = [
       "futures-io",
       "runtime-tokio",
       "rustls-aws-lc-rs",
@@ -202,7 +201,7 @@ rmp-serde = { version = "1.3.1" }
 serde = { version = "1.0.229" }
 serde_ignored = { version = "0.1" }
 serde_with = { version = "3", default-features = false, features = ["macros", "std"] }
-serde_json = { version = "1.0.154" }
+serde_json = { version = "1.0.151" }
 serde_urlencoded = "0.7.1"
 
 # Cryptography and Security
@@ -221,12 +220,12 @@ openidconnect = { default-features = false, version = "4.0" }
 pbkdf2 = "0.13.0"
 p256 = { version = "0.14.0", features = ["ecdsa", "pkcs8"] }
 rsa = { version = "=0.10.0-rc.18" }
-rustls = { default-features = false, version = "0.23.46" }
+rustls = { default-features = false, version = "0.23.45" }
 rustls-native-certs = "0.8"
 rustls-pki-types = "1.15.1"
 x509-parser = "0.18.1"
 sha1 = "0.11.0"
-sha2 = "0.11.1"
+sha2 = "0.11.0"
 subtle = "2.6"
 zeroize = { version = "1.9.1" }
 proptest = "1"
@@ -244,24 +243,24 @@ tokio-postgres-rustls = "0.14.0"
 
 # Utilities and Tools
 anyhow = "1.0.104"
-arc-swap = "1.9.3"
+arc-swap = "1.9.2"
 rustfs-tokio-tar = { version = "0.7.1" }
 # Candidate Snowball parser versions exercised by rustfs-zip compatibility fixtures.
-tar-codec = "0.0.15"
-tar-framing = "0.0.15"
-atoi = "3.1.1"
+tar-codec = "0.0.14"
+tar-framing = "0.0.14"
+atoi = "3.1.0"
 atomic_enum = "0.3.0"
-aws-config = { version = "1.13.0" }
-aws-credential-types = { version = "1.3.1" }
-aws-sdk-kms = { default-features = false, version = "1.126.0" }
-aws-sdk-s3 = { default-features = false, version = "1.155.0" }
-aws-sdk-sts = { default-features = false, version = "1.122.0" }
-aws-smithy-async = { version = "1.3.1" }
-aws-smithy-http-client = { default-features = false, version = "1.5.1" }
-aws-smithy-runtime-api = { version = "1.19.1" }
-aws-smithy-types = { version = "1.9.0" }
+aws-config = { version = "1.12.0" }
+aws-credential-types = { version = "1.3.0" }
+aws-sdk-kms = { default-features = false, version = "1.123.0" }
+aws-sdk-s3 = { default-features = false, version = "1.152.0" }
+aws-sdk-sts = { default-features = false, version = "1.119.0" }
+aws-smithy-async = { version = "1.3.0" }
+aws-smithy-http-client = { default-features = false, version = "1.5.0" }
+aws-smithy-runtime-api = { version = "1.19.0" }
+aws-smithy-types = { version = "1.8.1" }
 base64-simd = "0.8.0"
-brotli = "9.1.0"
+brotli = "9.0.0"
 clap = { version = "4.6.7" }
 const-str = { version = "1.1.0" }
 convert_case = "0.12.0"
@@ -270,15 +269,15 @@ crossbeam-queue = "0.3.14"
 crossbeam-channel = "0.5.17"
 crossbeam-deque = "0.8.8"
 crossbeam-utils = "0.8.23"
-datafusion = { default-features = false, version = "55.2.0" }
+datafusion = { default-features = false, version = "55.1.0" }
 derive_builder = "0.20.2"
 ed25519-dalek = "3.0.0"
-enumset = "1.1.15"
+enumset = "1.1.14"
 faster-hex = "1.0.0"
 flate2 = "1.1.10"
 glob = "0.3.4"
-google-cloud-storage = "1.21.0"
-google-cloud-auth = "1.18.0"
+google-cloud-storage = "1.20.0"
+google-cloud-auth = "1.17.0"
 hashbrown = { version = "0.17.1" }
 # Base32 for RFC 6238 TOTP shared secrets (RFC 4648 unpadded, the alphabet
 # every authenticator app expects). Already in the graph transitively.
@@ -288,11 +287,11 @@ highway = { version = "1.3.0" }
 hostname = "0.4.2"
 ipnetwork = { version = "0.21.1" }
 lazy_static = "1.5.1"
-libc = "0.2.191"
+libc = "0.2.190"
 libsystemd = "0.7.2"
 local-ip-address = "0.6.13"
 log = "0.4"
-memmap2 = "0.9.12"
+memmap2 = "0.9.11"
 lz4 = "1.28.1"
 matchit = "0.9.2"
 md-5 = "0.11.0"
@@ -302,8 +301,7 @@ netif = "0.1.6"
 num_cpus = { version = "1.17.0" }
 nvml-wrapper = "0.13.0"
 object = { version = "=0.40.0", default-features = false }
-parking_lot = "0.13.0"
-same-file = "1.0.6"
+parking_lot = "0.12.5"
 path-absolutize = "4.0.1"
 percent-encoding = "2.3.2"
 # Server-side QR rendering for TOTP enrollment, so neither the console nor the
@@ -317,19 +315,21 @@ ratelimit = "2.0.0"
 rayon = "1.12.0"
 rustfs-erasure-codec = { version = "9.0.0" }
 reed-solomon-simd = "3.1.0"
+openraft = { version = "=0.9.25", default-features = false, features = ["serde"] }
+redb = "=4.3.0"
 regex = { version = "1.13.1" }
 rumqttc = { package = "rumqttc-next", version = "0.34.0" }
-redis = { version = "1.8.0" }
+redis = { version = "1.7.1" }
 rustify = { version = "0.7", default-features = false }
 rustix = { version = "1.1.5" }
-rust-embed = { version = "8.13.0" }
+rust-embed = { version = "8.12.0" }
 rustc-hash = { version = "2.1.3" }
 s3s = { git = "https://github.com/s3s-project/s3s", rev = "5761ddfe4c509fcac87e3b99fd5bd9e176b442c0", features = ["minio"] }
 s3s-sigv4 = { git = "https://github.com/s3s-project/s3s", rev = "5761ddfe4c509fcac87e3b99fd5bd9e176b442c0" }
 serial_test = "4.0.1"
 shadow-rs = { default-features = false, version = "2.0.0" }
 siphasher = "1.0.4"
-smallvec = { version = "1.16.4" }
+smallvec = { version = "1.16.2" }
 compact_str = "0.10.0"
 snap = "1.1.2"
 starshard = { version = "2.4.0" }
@@ -343,12 +343,12 @@ tracing = { version = "0.1.44" }
 tracing-appender = "0.2.5"
 tracing-core = "0.1.36"
 tracing-error = "0.2.1"
-tracing-opentelemetry = { version = "0.34.0" }
+tracing-opentelemetry = { version = "0.34" }
 tracing-subscriber = { version = "0.3.23" }
 transform-stream = "0.3.1"
 url = "2.5.8"
 urlencoding = "2.1.3"
-uuid = { version = "1.29.0" }
+uuid = { version = "1.27.0" }
 vaultrs = { version = "0.8.0" }
 tar = "0.4.46"
 walkdir = "2.5.0"
@@ -356,37 +356,37 @@ winapi-util = "0.1.11"
 windows = { version = "0.62.2" }
 windows-sys = "0.61.2"
 xxhash-rust = { version = "0.8.19" }
-zip = "9.1.2"
-zstd = "0.14.3"
+zip = "8.6.0"
+zstd = "0.14.0"
 
 # Observability and Metrics
 metrics = "0.24.6"
 metrics-util = "0.20"
-dial9 = { version = "0.5.5", default-features = false }
-opentelemetry = { version = "0.33.1" }
-opentelemetry-appender-tracing = { version = "0.33.1" }
-opentelemetry-otlp = { version = "0.33.1" }
-opentelemetry-proto = { version = "0.33.1", default-features = false, features = ["metrics", "gen-tonic-messages"] }
-opentelemetry_sdk = { version = "0.33.1" }
-opentelemetry-semantic-conventions = { version = "0.33.1" }
-opentelemetry-stdout = { version = "0.33.1" }
-pyroscope = { version = "2.1.5" }
+dial9 = { version = "0.5.3", default-features = false }
+opentelemetry = { version = "0.33.0" }
+opentelemetry-appender-tracing = { version = "0.33.0" }
+opentelemetry-otlp = { version = "0.33.0" }
+opentelemetry-proto = { version = "0.33.0", default-features = false, features = ["metrics", "gen-tonic-messages"] }
+opentelemetry_sdk = { version = "0.33.0" }
+opentelemetry-semantic-conventions = { version = "0.33.0" }
+opentelemetry-stdout = { version = "0.33.0" }
+pyroscope = { version = "2.1.1" }
 
 # FTP and SFTP
 libunftp = { version = "0.23.1" }
 unftp-core = "0.1.0"
-suppaftp = { version = "12.2.1" }
+suppaftp = { version = "12.1.1" }
 rcgen = { version = "0.14.10", default-features = false, features = ["aws_lc_rs", "crypto", "pem"] }
-russh = { version = "0.65.0" }
+russh = { version = "0.64.1" }
 russh-sftp = "3.0.1"
 
 # WebDAV
 dav-server = "0.11.0"
 
 # Performance Analysis and Memory Profiling
-rustfs-mimalloc = { version = "0.7.2" }
+rustfs-mimalloc = { version = "0.5.6" }
 # Preserve Unicode focus filters until rustfs/backlog#2302 is resolved.
-hotpath = { version = "0.29.2", default-features = false }
+hotpath = { version = "0.28.5", default-features = false }
 # Snapshot testing for output format regression detection
 insta = { version = "1.49" }
 
@@ -423,20 +423,3 @@ inherits = "release"
 inherits = "release"
 debug = true
 strip = "none"
-
-# Pin upstream fixes until compatible crates.io releases are available.
-[patch.crates-io]
-zstd = { path = "crates/zstd-compat" }
-# Preserve RustFS OIDC flows while migrating the old elliptic-curve dependency chain.
-# Upstream: openidconnect-rs#244 and oauth2-rs#352.
-openidconnect = { git = "https://github.com/ramosbugs/openidconnect-rs", rev = "2f77ee73bc4554e1cb6279adb386005429fb4e41" }
-oauth2 = { git = "https://github.com/ramosbugs/oauth2-rs", rev = "89585f4a280afd7b285eaf1c9be0068874e03d22" }
-aws-sigv4 = { git = "https://github.com/smithy-lang/smithy-rs", rev = "5c1c4bca2ea7af97496e077c9f1a718eaac64bf8" }
-
-# Keep the signer's path dependencies on the SDK's existing registry types.
-[patch."https://github.com/smithy-lang/smithy-rs"]
-aws-credential-types = { version = "=1.3.1" }
-aws-smithy-eventstream = { version = "=0.61.6" }
-aws-smithy-http = { version = "=0.64.2" }
-aws-smithy-runtime-api = { version = "=1.19.1" }
-aws-smithy-types = { version = "=1.9.0" }

Dockerfile

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-FROM alpine:3.24.2 AS build
+FROM alpine:3.24.1 AS build
 
 ARG TARGETARCH
 ARG RELEASE=latest
@@ -70,7 +70,7 @@ RUN set -eux; \
     rm -rf rustfs.zip /build/.tmp || true
 
 
-FROM alpine:3.24.2
+FROM alpine:3.24.1
 
 ARG RELEASE=latest
 ARG BUILD_DATE

README.md

@@ -141,7 +141,7 @@ chown -R 10001:10001 data logs
 docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:latest
 
 # Using specific version
-docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.4
+docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.1
 ```
 
 If you use [podman](https://github.com/containers/podman) instead of docker, you can install the RustFS with the below command
@@ -220,7 +220,7 @@ For developers who want to build RustFS Docker images from source with multi-arc
 ./docker-buildx.sh --push
 
 # Build specific version
-./docker-buildx.sh --release 1.0.4 --push
+./docker-buildx.sh --release 1.0.1 --push
 
 # Build for custom registry
 ./docker-buildx.sh --registry your-registry.com --namespace yourname --push
@@ -238,7 +238,7 @@ You can also use Make targets for convenience:
 ```bash
 make docker-buildx                    # Build locally
 make docker-buildx-push               # Build and push
-make docker-buildx-version VERSION=1.0.4  # Build specific version
+make docker-buildx-version VERSION=1.0.1  # Build specific version
 make help-docker                      # Show all Docker-related commands
 ```
 

README_ZH.md

@@ -138,7 +138,7 @@ chown -R 10001:10001 data logs
 docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:latest
 
 # 使用指定版本运行
-docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.4
+docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.1
 ```
 
 如果您通过绑定挂载启用 TLS 证书目录，也请用同样方式准备该目录：
@@ -176,7 +176,7 @@ docker compose -f docker-compose-simple.yml up -d
 ./docker-buildx.sh --push
 
 # 构建指定版本
-./docker-buildx.sh --release 1.0.4 --push
+./docker-buildx.sh --release 1.0.1 --push
 
 # 构建并推送到自定义仓库
 ./docker-buildx.sh --registry your-registry.com --namespace yourname --push
@@ -193,7 +193,7 @@ docker compose -f docker-compose-simple.yml up -d
 ```bash
 make docker-buildx                    # 本地构建
 make docker-buildx-push               # 构建并推送
-make docker-buildx-version VERSION=1.0.4  # 构建指定版本
+make docker-buildx-version VERSION=1.0.1  # 构建指定版本
 make help-docker                      # 显示所有 Docker 相关命令
 ```
 

_typos.toml

@@ -51,8 +51,6 @@ unparseable = "unparseable"
 # Disaster-recovery objectives: recovery time and recovery point.
 RTO = "RTO"
 rto = "rto"
-# rustix OFlags variant (POSIX O_WRONLY equivalent).
-WRONLY = "WRONLY"
 
 [files]
 extend-exclude = []

crates/capture-authority/Cargo.toml

@@ -0,0 +1,24 @@
+[package]
+name = "rustfs-capture-authority"
+version.workspace = true
+edition.workspace = true
+license.workspace = true
+rust-version.workspace = true
+repository.workspace = true
+description = "Durable local storage for capture authority"
+
+[dependencies]
+openraft = { workspace = true }
+redb = { workspace = true }
+sha2 = { workspace = true }
+serde = { workspace = true, features = ["derive"] }
+thiserror = { workspace = true }
+tokio = { workspace = true }
+url = { workspace = true }
+
+[dev-dependencies]
+serde_json = { workspace = true }
+tempfile = { workspace = true }
+
+[lints]
+workspace = true

crates/capture-authority/src/application.rs

@@ -0,0 +1,862 @@
+fn read_native_range_in<R: std::ops::RangeBounds<u64>>(
+    tx: &redb::ReadTransaction,
+    range: R,
+) -> Result<Vec<crate::CaptureRaftEntry>, StoreError> {
+    use std::ops::Bound::{Excluded, Included, Unbounded};
+    let logs = tx.open_table(LOGS)?;
+    let start = match range.start_bound() {
+        Included(value) => *value,
+        Excluded(value) => match value.checked_add(1) {
+            Some(value) => value,
+            None => return Ok(Vec::new()),
+        },
+        Unbounded => 0,
+    };
+    let end = match range.end_bound() {
+        Included(value) => *value,
+        Excluded(value) => match value.checked_sub(1) {
+            Some(value) => value,
+            None => return Ok(Vec::new()),
+        },
+        Unbounded => u64::MAX,
+    };
+    if start > end {
+        return Ok(Vec::new());
+    }
+    let mut entries = Vec::new();
+    for row in logs.range(start..=end)? {
+        let (key, value) = row?;
+        let entry = decode_log(value.value())?;
+        if entry.id.index != key.value() {
+            return Err(StoreError::CorruptRecord);
+        }
+        entries.push(decode_native_entry(entry)?);
+    }
+    Ok(entries)
+}
+
+use crate::{
+    LOGS, LogId, META, Store, StoreError, committed::validate_committed, created::*, decode_log, encode_id, encode_log,
+    next_index, number,
+};
+use redb::{ReadTransaction, ReadableDatabase, ReadableTable, ReadableTableMetadata, TableDefinition};
+use std::collections::BTreeMap;
+
+pub(crate) const HEADS: TableDefinition<&[u8], &[u8]> = TableDefinition::new("heads");
+pub(crate) const VERSIONS: TableDefinition<&[u8], &[u8]> = TableDefinition::new("versions");
+pub(crate) const EVENTS: TableDefinition<&[u8], &[u8]> = TableDefinition::new("events");
+pub(crate) const OPERATIONS: TableDefinition<&[u8], &[u8]> = TableDefinition::new("operations");
+pub(crate) const APPLY_RESULTS: TableDefinition<u64, &[u8]> = TableDefinition::new("apply_results");
+
+fn version_key(object: &ObjectIdentity, version: [u8; 16]) -> Result<Vec<u8>, StoreError> {
+    let mut b = object_bytes(object)?;
+    b.extend(version);
+    Ok(b)
+}
+fn decode_version(b: &[u8]) -> Result<CreatedVersion, StoreError> {
+    let mut d = Decoder::new(b);
+    let v = CreatedVersion {
+        prepared: d.prepared()?,
+        decision_log_id: d.id()?,
+        binding_revision: d.u64()?,
+    };
+    d.finish()?;
+    Ok(v)
+}
+fn decode_event(b: &[u8]) -> Result<CreatedEvent, StoreError> {
+    let mut d = Decoder::new(b);
+    let e = CreatedEvent {
+        object: d.object()?,
+        prepared: d.prepared()?,
+        decision_log_id: d.id()?,
+        binding: d.binding()?,
+    };
+    d.finish()?;
+    Ok(e)
+}
+fn decode_operation(b: &[u8]) -> Result<CreatedOperation, StoreError> {
+    let mut d = Decoder::new(b);
+    let len = usize::try_from(d.u64()?).map_err(|_| StoreError::CorruptRecord)?;
+    let command = b
+        .get(8..8usize.checked_add(len).ok_or(StoreError::CorruptRecord)?)
+        .ok_or(StoreError::CorruptRecord)?;
+    DecideCreated::decode(command)?;
+    let mut d = Decoder::new(&b[8 + len..]);
+    let result = d.result()?;
+    d.finish()?;
+    Ok(CreatedOperation {
+        canonical_command: command.to_vec(),
+        result,
+    })
+}
+fn decode_apply(b: &[u8]) -> Result<CreatedApplyResult, StoreError> {
+    let mut d = Decoder::new(b);
+    let r = CreatedApplyResult {
+        log_id: d.id()?,
+        result: d.result()?,
+    };
+    d.finish()?;
+    Ok(r)
+}
+fn read_binding(meta: &impl ReadableTable<&'static str, &'static [u8]>) -> Result<CaptureBinding, StoreError> {
+    let b = meta.get("capture_binding")?.ok_or(StoreError::BindingNotPinned)?;
+    let mut d = Decoder::new(b.value());
+    let binding = d.binding()?;
+    d.finish()?;
+    Ok(binding)
+}
+fn read_applied(meta: &impl ReadableTable<&'static str, &'static [u8]>) -> Result<Option<LogId>, StoreError> {
+    let b = meta.get("last_applied")?.ok_or(StoreError::CorruptRecord)?;
+    let mut d = Decoder::new(b.value());
+    let id = match d.take::<1>()?[0] {
+        0 => None,
+        1 => Some(d.id()?),
+        _ => return Err(StoreError::CorruptRecord),
+    };
+    d.finish()?;
+    Ok(id)
+}
+fn required<T>(value: Option<T>) -> Result<T, StoreError> {
+    value.ok_or(StoreError::CorruptRecord)
+}
+fn encode_native_entry(entry: &crate::CaptureRaftEntry) -> Result<Vec<u8>, StoreError> {
+    let id: LogId = entry.log_id.into();
+    let (kind, body) = match &entry.payload {
+        openraft::EntryPayload::Blank => (0, vec![]),
+        openraft::EntryPayload::Normal(command) => (1, command.created().encode()?),
+        openraft::EntryPayload::Membership(membership) => {
+            let stored = openraft::StoredMembership::new(Some(entry.log_id), membership.clone());
+            let checked = crate::CaptureMembership::try_from(&stored)?;
+            (2, checked.encode()?)
+        }
+    };
+    let mut payload = b"CENR\x01".to_vec();
+    payload.push(kind);
+    payload.extend(
+        u64::try_from(body.len())
+            .map_err(|_| StoreError::CorruptRecord)?
+            .to_be_bytes(),
+    );
+    payload.extend(body);
+    encode_log(&crate::LogEntry { id, payload })
+}
+
+pub(crate) fn decode_native_entry(entry: crate::LogEntry) -> Result<crate::CaptureRaftEntry, StoreError> {
+    let bytes = &entry.payload;
+    if bytes.len() < 14 || &bytes[..5] != b"CENR\x01" {
+        return Err(StoreError::CorruptRecord);
+    }
+    let length = usize::try_from(number(&bytes[6..14])?).map_err(|_| StoreError::CorruptRecord)?;
+    let body = &bytes[14..];
+    if length != body.len() {
+        return Err(StoreError::CorruptRecord);
+    }
+    let payload = match bytes[5] {
+        0 if body.is_empty() => openraft::EntryPayload::Blank,
+        1 => openraft::EntryPayload::Normal(crate::CaptureCommand::try_created(DecideCreated::decode(body)?)?),
+        2 => {
+            let membership = crate::CaptureMembership::decode(body).map_err(|_| StoreError::CorruptRecord)?;
+            if *membership.log_id() != Some(entry.id) {
+                return Err(StoreError::CorruptRecord);
+            }
+            let stored = openraft::StoredMembership::try_from(&membership).map_err(|_| StoreError::CorruptRecord)?;
+            openraft::EntryPayload::Membership(stored.membership().clone())
+        }
+        _ => return Err(StoreError::CorruptRecord),
+    };
+    Ok(crate::CaptureRaftEntry {
+        log_id: entry.id.into(),
+        payload,
+    })
+}
+
+fn apply_created(
+    tx: &redb::WriteTransaction,
+    id: LogId,
+    command: &DecideCreated,
+    canonical: &[u8],
+    binding: &CaptureBinding,
+) -> Result<CreatedResult, StoreError> {
+    let mut operations = tx.open_table(OPERATIONS)?;
+    let existing = operations
+        .get(command.operation_id.as_slice())?
+        .map(|v| decode_operation(v.value()))
+        .transpose()?;
+    let result = if let Some(existing) = existing {
+        if existing.canonical_command == canonical {
+            existing.result
+        } else {
+            CreatedResult::OperationConflict
+        }
+    } else {
+        let mut heads = tx.open_table(HEADS)?;
+        let mut versions = tx.open_table(VERSIONS)?;
+        let object = object_bytes(&command.object)?;
+        let key = version_key(&command.object, command.prepared.version_id)?;
+        let actual = heads
+            .get(object.as_slice())?
+            .map(|v| {
+                let mut d = Decoder::new(v.value());
+                let id = d.uuid()?;
+                d.finish()?;
+                Ok::<_, StoreError>(id)
+            })
+            .transpose()?;
+        let outcome = if command.binding_revision != binding.revision {
+            CreatedResult::BindingMismatch
+        } else if actual != command.expected_head {
+            CreatedResult::HeadMismatch { actual_head: actual }
+        } else if versions.get(key.as_slice())?.is_some() {
+            CreatedResult::VersionConflict
+        } else {
+            let mut version = Vec::new();
+            put_prepared(&mut version, &command.prepared);
+            version.extend(encode_id(id));
+            version.extend(binding.revision.to_be_bytes());
+            versions.insert(key.as_slice(), version.as_slice())?;
+            let mut event = object;
+            put_prepared(&mut event, &command.prepared);
+            event.extend(encode_id(id));
+            event.extend(binding_bytes(&binding)?);
+            tx.open_table(EVENTS)?
+                .insert(command.operation_id.as_slice(), event.as_slice())?;
+            heads.insert(object_bytes(&command.object)?.as_slice(), command.prepared.version_id.as_slice())?;
+            CreatedResult::Created {
+                version_id: command.prepared.version_id,
+                event_id: command.operation_id,
+                decision_log_id: id,
+            }
+        };
+        let mut op = u64::try_from(canonical.len())
+            .map_err(|_| StoreError::CorruptRecord)?
+            .to_be_bytes()
+            .to_vec();
+        op.extend(canonical);
+        op.extend(result_bytes(&outcome));
+        operations.insert(command.operation_id.as_slice(), op.as_slice())?;
+        outcome
+    };
+    Ok(result)
+}
+
+fn validate_response_kind(entry: &crate::CaptureRaftEntry, response: &crate::CaptureResponse) -> Result<(), StoreError> {
+    let kind = match entry.payload {
+        openraft::EntryPayload::Blank => crate::CaptureResponseKind::Blank,
+        openraft::EntryPayload::Normal(_) => crate::CaptureResponseKind::Created,
+        openraft::EntryPayload::Membership(_) => crate::CaptureResponseKind::Membership,
+    };
+    if response.kind() != kind || *response.log_id() != LogId::from(entry.log_id) {
+        return Err(StoreError::CorruptRecord);
+    }
+    Ok(())
+}
+fn encode_response(response: &crate::CaptureResponse) -> Result<Vec<u8>, StoreError> {
+    let (kind, body) = match response.kind() {
+        crate::CaptureResponseKind::Blank => (0, vec![]),
+        crate::CaptureResponseKind::Created => (1, result_bytes(response.created_result().ok_or(StoreError::CorruptRecord)?)),
+        crate::CaptureResponseKind::Membership => (2, vec![]),
+    };
+    let mut bytes = b"CRSP\x01".to_vec();
+    bytes.push(kind);
+    bytes.extend(encode_id(*response.log_id()));
+    bytes.extend(
+        u64::try_from(body.len())
+            .map_err(|_| StoreError::CorruptRecord)?
+            .to_be_bytes(),
+    );
+    bytes.extend(body);
+    Ok(bytes)
+}
+fn decode_response(bytes: &[u8]) -> Result<crate::CaptureResponse, StoreError> {
+    if bytes.len() < 38
+        || &bytes[..5] != b"CRSP\x01"
+        || number(&bytes[30..38])? != u64::try_from(bytes.len() - 38).map_err(|_| StoreError::CorruptRecord)?
+    {
+        return Err(StoreError::CorruptRecord);
+    }
+    let id = crate::decode_id(&bytes[6..30])?;
+    match bytes[5] {
+        0 if bytes.len() == 38 => crate::CaptureResponse::blank(id),
+        1 => {
+            let mut body = encode_id(id);
+            body.extend(&bytes[38..]);
+            crate::CaptureResponse::created(decode_apply(&body)?)
+        }
+        2 if bytes.len() == 38 => crate::CaptureResponse::membership(id),
+        _ => Err(StoreError::CorruptRecord),
+    }
+}
+fn entry_command(entry: &crate::LogEntry, native: bool) -> Result<DecideCreated, StoreError> {
+    if native {
+        match decode_native_entry(entry.clone())?.payload {
+            openraft::EntryPayload::Normal(command) => Ok(command.created().clone()),
+            _ => Err(StoreError::CorruptRecord),
+        }
+    } else {
+        DecideCreated::decode(&entry.payload)
+    }
+}
+
+pub(crate) fn native_application_state(tx: &ReadTransaction) -> Result<(Option<LogId>, crate::CaptureMembership), StoreError> {
+    let meta = tx.open_table(META)?;
+    if number(required(meta.get("schema")?)?.value())? != 4 {
+        return Err(StoreError::UnsupportedSchema);
+    }
+    let last = read_applied(&meta)?;
+    let membership = crate::CaptureMembership::decode(required(meta.get("native_membership")?)?.value())
+        .map_err(|_| StoreError::CorruptRecord)?;
+    if membership
+        .log_id()
+        .is_some_and(|id| last.is_none_or(|last| id.index > last.index))
+    {
+        return Err(StoreError::CorruptRecord);
+    }
+    Ok((last, membership))
+}
+
+impl Store {
+    /// Explicitly opt an empty, bound store into native entry persistence.
+    pub fn enable_native_empty(&mut self) -> Result<(), StoreError> {
+        let tx = self.write_transaction()?;
+        {
+            let mut meta = tx.open_table(META)?;
+            let schema = number(required(meta.get("schema")?)?.value())?;
+            if !matches!(schema, 3 | 4) {
+                return Err(StoreError::UnsupportedSchema);
+            }
+            read_binding(&meta)?;
+            if schema == 4 {
+                let membership = crate::CaptureMembership::decode(required(meta.get("native_membership")?)?.value())
+                    .map_err(|_| StoreError::CorruptRecord)?;
+                let last = read_applied(&meta)?;
+                if membership
+                    .log_id()
+                    .is_some_and(|id| last.is_none_or(|last| id.index > last.index))
+                    || (last.is_none() && (!membership.configs().is_empty() || !membership.nodes().is_empty()))
+                {
+                    return Err(StoreError::CorruptRecord);
+                }
+                return Ok(());
+            }
+            if read_applied(&meta)?.is_some() || !tx.open_table(LOGS)?.is_empty()? {
+                return Err(StoreError::UnsupportedSchema);
+            }
+            for table in [HEADS, VERSIONS, EVENTS, OPERATIONS] {
+                if !tx.open_table(table)?.is_empty()? {
+                    return Err(StoreError::UnsupportedSchema);
+                }
+            }
+            if !tx.open_table(APPLY_RESULTS)?.is_empty()? || meta.get("committed")?.is_some() || meta.get("tail")?.is_some() {
+                return Err(StoreError::UnsupportedSchema);
+            }
+            if meta.get("native_membership")?.is_some() {
+                return Err(StoreError::CorruptRecord);
+            }
+            let membership = crate::CaptureMembership::new(None, vec![], BTreeMap::new())?;
+            meta.insert("native_membership", membership.encode()?.as_slice())?;
+            meta.insert("schema", 4u64.to_be_bytes().as_slice())?;
+        }
+        self.commit(tx)
+    }
+    pub fn append_native(&mut self, entries: &[crate::CaptureRaftEntry]) -> Result<(), StoreError> {
+        let tx = self.write_transaction()?;
+        {
+            let mut meta = tx.open_table(META)?;
+            if number(required(meta.get("schema")?)?.value())? != 4 {
+                return Err(StoreError::UnsupportedSchema);
+            }
+            if entries.is_empty() {
+                return Ok(());
+            }
+            let tail = meta.get("tail")?.map(|value| crate::decode_id(value.value())).transpose()?;
+            let mut next = next_index(tail.map(|id| id.index))?;
+            let mut encoded = Vec::with_capacity(entries.len());
+            for (position, entry) in entries.iter().enumerate() {
+                if entry.log_id.index != next {
+                    return Err(StoreError::InvalidLogSequence);
+                }
+                encoded.push(encode_native_entry(entry)?);
+                if position + 1 < entries.len() {
+                    next = next_index(Some(next))?;
+                }
+            }
+            let mut logs = tx.open_table(LOGS)?;
+            for (entry, bytes) in entries.iter().zip(&encoded) {
+                logs.insert(entry.log_id.index, bytes.as_slice())?;
+            }
+            let last = entries.last().ok_or(StoreError::InvalidLogSequence)?;
+            meta.insert("tail", encode_id(last.log_id.into()).as_slice())?;
+        }
+        self.commit(tx)
+    }
+    pub fn read_native_entry(&self, index: u64) -> Result<Option<crate::CaptureRaftEntry>, StoreError> {
+        let tx = self.database.begin_read()?;
+        if number(required(tx.open_table(META)?.get("schema")?)?.value())? != 4 {
+            return Err(StoreError::UnsupportedSchema);
+        }
+        tx.open_table(LOGS)?
+            .get(index)?
+            .map(|value| {
+                let entry = decode_log(value.value())?;
+                if entry.id.index != index {
+                    return Err(StoreError::CorruptRecord);
+                }
+                decode_native_entry(entry)
+            })
+            .transpose()
+    }
+    /// Read typed native entries within the requested bounds.
+    pub fn read_native_range<R: std::ops::RangeBounds<u64>>(&self, range: R) -> Result<Vec<crate::CaptureRaftEntry>, StoreError> {
+        let tx = self.database.begin_read()?;
+        if number(required(tx.open_table(META)?.get("schema")?)?.value())? != 4 {
+            return Err(StoreError::UnsupportedSchema);
+        }
+        read_native_range_in(&tx, range)
+    }
+    pub fn read_native_entries(&self, start: u64, end: u64) -> Result<Vec<crate::CaptureRaftEntry>, StoreError> {
+        let tx = self.database.begin_read()?;
+        if number(required(tx.open_table(META)?.get("schema")?)?.value())? != 4 {
+            return Err(StoreError::UnsupportedSchema);
+        }
+        if start > end {
+            return Err(StoreError::InvalidLogSequence);
+        }
+        read_native_range_in(&tx, start..end)
+    }
+
+    pub fn apply_next_native(&mut self, id: LogId) -> Result<crate::CaptureResponse, StoreError> {
+        let tx = self.write_transaction()?;
+        let response;
+        {
+            let mut meta = tx.open_table(META)?;
+            if number(required(meta.get("schema")?)?.value())? != 4 {
+                return Err(StoreError::UnsupportedSchema);
+            }
+            let binding = read_binding(&meta)?;
+            let logs = tx.open_table(LOGS)?;
+            let committed = validate_committed(&meta, &logs)?.ok_or(StoreError::InvalidCommittedLog)?;
+            let entry = decode_log(required(logs.get(id.index)?)?.value())?;
+            if entry.id != id || id.index > committed.index {
+                return Err(StoreError::InvalidCommittedLog);
+            }
+            let native = decode_native_entry(entry)?;
+            let last = read_applied(&meta)?;
+            let mut results = tx.open_table(APPLY_RESULTS)?;
+            if last.is_some_and(|last| id.index <= last.index) {
+                let saved = decode_response(required(results.get(id.index)?)?.value())?;
+                if *saved.log_id() != id {
+                    return Err(StoreError::InvalidApplySequence);
+                }
+                validate_response_kind(&native, &saved)?;
+                return Ok(saved);
+            }
+            if id.index != next_index(last.map(|last| last.index))? {
+                return Err(StoreError::InvalidApplySequence);
+            }
+            response = match native.payload {
+                openraft::EntryPayload::Blank => crate::CaptureResponse::blank(id)?,
+                openraft::EntryPayload::Normal(command) => {
+                    let canonical = command.created().encode()?;
+                    let result = apply_created(&tx, id, command.created(), &canonical, &binding)?;
+                    crate::CaptureResponse::created(CreatedApplyResult { log_id: id, result })?
+                }
+                openraft::EntryPayload::Membership(membership) => {
+                    let stored = openraft::StoredMembership::new(Some(id.into()), membership);
+                    let checked = crate::CaptureMembership::try_from(&stored)?;
+                    meta.insert("native_membership", checked.encode()?.as_slice())?;
+                    crate::CaptureResponse::membership(id)?
+                }
+            };
+            results.insert(id.index, encode_response(&response)?.as_slice())?;
+            let mut last = vec![1];
+            last.extend(encode_id(id));
+            meta.insert("last_applied", last.as_slice())?;
+        }
+        self.commit(tx)?;
+        Ok(response)
+    }
+    pub fn native_apply_response(&self, index: u64) -> Result<Option<crate::CaptureResponse>, StoreError> {
+        let tx = self.database.begin_read()?;
+        if number(required(tx.open_table(META)?.get("schema")?)?.value())? != 4 {
+            return Err(StoreError::UnsupportedSchema);
+        }
+        let results = tx.open_table(APPLY_RESULTS)?;
+        results.get(index)?.map(|v| decode_response(v.value())).transpose()
+    }
+    pub fn native_application_state(&self) -> Result<(Option<LogId>, crate::CaptureMembership), StoreError> {
+        native_application_state(&self.database.begin_read()?)
+    }
+    /// Bootstrap the immutable capture policy. The caller must be the trusted
+    /// deployment provisioning owner. This crate does not authenticate callers;
+    /// do not expose this public method through an ordinary object request API.
+    /// Prepared identities assert identity only, not physical data durability.
+    pub fn pin_capture_binding(&mut self, binding: CaptureBinding) -> Result<(), StoreError> {
+        let encoded = binding_bytes(&binding)?;
+        let tx = self.write_transaction()?;
+        {
+            let mut meta = tx.open_table(META)?;
+            if matches!(number(required(meta.get("schema")?)?.value())?, 3 | 4) {
+                if read_binding(&meta)? == binding {
+                    return Ok(());
+                }
+                return Err(StoreError::BindingConflict);
+            }
+            if meta.get("capture_binding")?.is_some() || meta.get("last_applied")?.is_some() {
+                return Err(StoreError::CorruptRecord);
+            }
+            for table in [HEADS, VERSIONS, EVENTS, OPERATIONS] {
+                if !tx.open_table(table)?.is_empty()? {
+                    return Err(StoreError::CorruptRecord);
+                }
+            }
+            if !tx.open_table(APPLY_RESULTS)?.is_empty()? {
+                return Err(StoreError::CorruptRecord);
+            }
+            meta.insert("capture_binding", encoded.as_slice())?;
+            meta.insert("last_applied", &[0][..])?;
+            meta.insert("schema", 3u64.to_be_bytes().as_slice())?;
+        }
+        self.commit(tx)
+    }
+    /// Apply the next exact persisted committed Created command atomically.
+    pub fn apply_next_created(&mut self, id: LogId) -> Result<CreatedResult, StoreError> {
+        let tx = self.write_transaction()?;
+        let result;
+        {
+            let mut meta = tx.open_table(META)?;
+            if number(required(meta.get("schema")?)?.value())? != 3 {
+                return Err(StoreError::UnsupportedSchema);
+            }
+            let binding = read_binding(&meta)?;
+            let logs = tx.open_table(LOGS)?;
+            let committed = validate_committed(&meta, &logs)?.ok_or(StoreError::InvalidCommittedLog)?;
+            let entry = decode_log(required(logs.get(id.index)?)?.value())?;
+            if entry.id != id || id.index > committed.index {
+                return Err(StoreError::InvalidCommittedLog);
+            }
+            let last = read_applied(&meta)?;
+            let mut results = tx.open_table(APPLY_RESULTS)?;
+            if last.is_some_and(|last| id.index <= last.index) {
+                let saved = decode_apply(required(results.get(id.index)?)?.value())?;
+                if saved.log_id != id {
+                    return Err(StoreError::InvalidApplySequence);
+                }
+                return Ok(saved.result);
+            }
+            if id.index != next_index(last.map(|last| last.index))? {
+                return Err(StoreError::InvalidApplySequence);
+            }
+            let command = DecideCreated::decode(&entry.payload)?;
+            result = apply_created(&tx, id, &command, &entry.payload, &binding)?;
+            let mut applied = encode_id(id);
+            applied.extend(result_bytes(&result));
+            results.insert(id.index, applied.as_slice())?;
+            let mut last = vec![1];
+            last.extend(encode_id(id));
+            meta.insert("last_applied", last.as_slice())?;
+        }
+        self.commit(tx)?;
+        Ok(result)
+    }
+    pub fn capture_binding(&self) -> Result<Option<CaptureBinding>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let meta = tx.open_table(META)?;
+        if !matches!(number(required(meta.get("schema")?)?.value())?, 3 | 4) {
+            return Ok(None);
+        }
+        Ok(Some(read_binding(&meta)?))
+    }
+    pub fn last_applied(&self) -> Result<Option<LogId>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let meta = tx.open_table(META)?;
+        if !matches!(number(required(meta.get("schema")?)?.value())?, 3 | 4) {
+            return Ok(None);
+        }
+        read_applied(&meta)
+    }
+    pub fn created_head(&self, object: &ObjectIdentity) -> Result<Option<[u8; 16]>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let t = tx.open_table(HEADS)?;
+        t.get(object_bytes(object)?.as_slice())?
+            .map(|v| {
+                let mut d = Decoder::new(v.value());
+                let id = d.uuid()?;
+                d.finish()?;
+                Ok(id)
+            })
+            .transpose()
+    }
+    pub fn created_version(&self, object: &ObjectIdentity, version: [u8; 16]) -> Result<Option<CreatedVersion>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let t = tx.open_table(VERSIONS)?;
+        t.get(version_key(object, version)?.as_slice())?
+            .map(|v| decode_version(v.value()))
+            .transpose()
+    }
+    pub fn created_event(&self, operation: [u8; 16]) -> Result<Option<CreatedEvent>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let t = tx.open_table(EVENTS)?;
+        t.get(operation.as_slice())?.map(|v| decode_event(v.value())).transpose()
+    }
+    pub fn created_operation(&self, operation: [u8; 16]) -> Result<Option<CreatedOperation>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let t = tx.open_table(OPERATIONS)?;
+        t.get(operation.as_slice())?.map(|v| decode_operation(v.value())).transpose()
+    }
+    pub fn created_apply_result(&self, index: u64) -> Result<Option<CreatedApplyResult>, StoreError> {
+        let tx = self.database.begin_read()?;
+        let t = tx.open_table(APPLY_RESULTS)?;
+        if number(required(tx.open_table(META)?.get("schema")?)?.value())? == 4 {
+            t.get(index)?
+                .map(|v| {
+                    decode_response(v.value()).map(|r| {
+                        r.created_result().cloned().map(|result| CreatedApplyResult {
+                            log_id: *r.log_id(),
+                            result,
+                        })
+                    })
+                })
+                .transpose()
+                .map(Option::flatten)
+        } else {
+            t.get(index)?.map(|v| decode_apply(v.value())).transpose()
+        }
+    }
+}
+
+pub(crate) fn validate_application(tx: &ReadTransaction) -> Result<(), StoreError> {
+    let meta = tx.open_table(META)?;
+    let schema = number(required(meta.get("schema")?)?.value())?;
+    let native = schema == 4;
+    if !matches!(schema, 3 | 4) {
+        return Ok(());
+    }
+    let binding = read_binding(&meta).map_err(|error| match error {
+        StoreError::BindingNotPinned => StoreError::CorruptRecord,
+        other => other,
+    })?;
+    let last = read_applied(&meta)?;
+    let logs = tx.open_table(LOGS)?;
+    let committed = validate_committed(&meta, &logs)?;
+    let heads = tx.open_table(HEADS)?;
+    let versions = tx.open_table(VERSIONS)?;
+    let events = tx.open_table(EVENTS)?;
+    let operations = tx.open_table(OPERATIONS)?;
+    let results = tx.open_table(APPLY_RESULTS)?;
+    if last.is_some_and(|last| committed.is_none_or(|c| last.index > c.index)) {
+        return Err(StoreError::CorruptRecord);
+    }
+    let mut next = Some(0);
+    let mut observed = None;
+    let mut first_operations = BTreeMap::new();
+    let mut expected_membership = crate::CaptureMembership::new(None, vec![], BTreeMap::new())?;
+    let mut replay_operations = BTreeMap::<[u8; 16], (Vec<u8>, CreatedResult)>::new();
+    let mut replay_heads = BTreeMap::<Vec<u8>, [u8; 16]>::new();
+    let mut replay_versions = std::collections::BTreeSet::new();
+    for row in results.iter()? {
+        let (k, v) = row?;
+        let entry = decode_log(required(logs.get(k.value())?)?.value())?;
+        let r = if native {
+            let response = decode_response(v.value())?;
+            let native_entry = decode_native_entry(entry.clone())?;
+            validate_response_kind(&native_entry, &response)?;
+            if next != Some(k.value()) || response.log_id().index != k.value() {
+                return Err(StoreError::CorruptRecord);
+            }
+            observed = Some(*response.log_id());
+            next = k.value().checked_add(1);
+            match native_entry.payload {
+                openraft::EntryPayload::Membership(membership) => {
+                    expected_membership =
+                        crate::CaptureMembership::try_from(&openraft::StoredMembership::new(Some(entry.id.into()), membership))?;
+                    continue;
+                }
+                openraft::EntryPayload::Blank => continue,
+                openraft::EntryPayload::Normal(_) => CreatedApplyResult {
+                    log_id: *response.log_id(),
+                    result: response.created_result().cloned().ok_or(StoreError::CorruptRecord)?,
+                },
+            }
+        } else {
+            decode_apply(v.value())?
+        };
+        if !native && (next != Some(k.value()) || r.log_id.index != k.value()) {
+            return Err(StoreError::CorruptRecord);
+        }
+        if entry.id != r.log_id {
+            return Err(StoreError::CorruptRecord);
+        }
+        let c = entry_command(&entry, native)?;
+        let canonical = c.encode()?;
+        let op = decode_operation(required(operations.get(c.operation_id.as_slice())?)?.value())?;
+        let canonical_matches = op.canonical_command == canonical;
+        let expected = if canonical_matches {
+            op.result
+        } else {
+            CreatedResult::OperationConflict
+        };
+        if native {
+            let replay = if let Some((previous, result)) = replay_operations.get(&c.operation_id) {
+                if *previous == canonical {
+                    result.clone()
+                } else {
+                    CreatedResult::OperationConflict
+                }
+            } else {
+                let object = object_bytes(&c.object)?;
+                let key = version_key(&c.object, c.prepared.version_id)?;
+                let actual = replay_heads.get(&object).copied();
+                let result = if c.binding_revision != binding.revision {
+                    CreatedResult::BindingMismatch
+                } else if actual != c.expected_head {
+                    CreatedResult::HeadMismatch { actual_head: actual }
+                } else if replay_versions.contains(&key) {
+                    CreatedResult::VersionConflict
+                } else {
+                    replay_heads.insert(object, c.prepared.version_id);
+                    replay_versions.insert(key);
+                    CreatedResult::Created {
+                        version_id: c.prepared.version_id,
+                        event_id: c.operation_id,
+                        decision_log_id: r.log_id,
+                    }
+                };
+                replay_operations.insert(c.operation_id, (canonical.clone(), result.clone()));
+                result
+            };
+            if replay != r.result {
+                return Err(StoreError::CorruptRecord);
+            }
+        }
+        if expected != r.result {
+            return Err(StoreError::CorruptRecord);
+        }
+        if canonical_matches {
+            first_operations.entry(c.operation_id).or_insert(r.log_id);
+        }
+        observed = Some(r.log_id);
+        next = k.value().checked_add(1)
+    }
+    if observed != last {
+        return Err(StoreError::CorruptRecord);
+    }
+    if native {
+        if native_application_state(tx)?.1 != expected_membership
+            || heads.len()? != u64::try_from(replay_heads.len()).map_err(|_| StoreError::CorruptRecord)?
+            || versions.len()? != u64::try_from(replay_versions.len()).map_err(|_| StoreError::CorruptRecord)?
+        {
+            return Err(StoreError::CorruptRecord);
+        }
+        for (object, id) in replay_heads {
+            if required(heads.get(object.as_slice())?)?.value() != id {
+                return Err(StoreError::CorruptRecord);
+            }
+        }
+    }
+    for row in operations.iter()? {
+        let (k, v) = row?;
+        let op = decode_operation(v.value())?;
+        let c = DecideCreated::decode(&op.canonical_command)?;
+        if k.value() != c.operation_id {
+            return Err(StoreError::CorruptRecord);
+        }
+        let first = required(first_operations.get(&c.operation_id))?;
+        if let CreatedResult::Created {
+            version_id,
+            event_id,
+            decision_log_id,
+        } = op.result
+        {
+            if version_id != c.prepared.version_id
+                || event_id != c.operation_id
+                || c.binding_revision != binding.revision
+                || decision_log_id != *first
+            {
+                return Err(StoreError::CorruptRecord);
+            }
+            let event = decode_event(required(events.get(event_id.as_slice())?)?.value())?;
+            if event
+                != (CreatedEvent {
+                    object: c.object.clone(),
+                    prepared: c.prepared.clone(),
+                    decision_log_id,
+                    binding: binding.clone(),
+                })
+            {
+                return Err(StoreError::CorruptRecord);
+            }
+        } else if events.get(c.operation_id.as_slice())?.is_some() {
+            return Err(StoreError::CorruptRecord);
+        }
+    }
+    for row in events.iter()? {
+        let (k, v) = row?;
+        let e = decode_event(v.value())?;
+        if e.binding != binding {
+            return Err(StoreError::CorruptRecord);
+        }
+        let op = decode_operation(required(operations.get(k.value())?)?.value())?;
+        let c = DecideCreated::decode(&op.canonical_command)?;
+        if c.object != e.object
+            || c.prepared != e.prepared
+            || op.result
+                != (CreatedResult::Created {
+                    version_id: e.prepared.version_id,
+                    event_id: c.operation_id,
+                    decision_log_id: e.decision_log_id,
+                })
+        {
+            return Err(StoreError::CorruptRecord);
+        }
+        let version =
+            decode_version(required(versions.get(version_key(&e.object, e.prepared.version_id)?.as_slice())?)?.value())?;
+        if version
+            != (CreatedVersion {
+                prepared: e.prepared.clone(),
+                decision_log_id: e.decision_log_id,
+                binding_revision: binding.revision,
+            })
+        {
+            return Err(StoreError::CorruptRecord);
+        }
+    }
+    for row in versions.iter()? {
+        let (k, v) = row?;
+        let version = decode_version(v.value())?;
+        let mut d = Decoder::new(k.value());
+        let object = d.object()?;
+        let id = d.uuid()?;
+        d.finish()?;
+        if id != version.prepared.version_id || version.binding_revision != binding.revision {
+            return Err(StoreError::CorruptRecord);
+        }
+        let entry = decode_log(required(logs.get(version.decision_log_id.index)?)?.value())?;
+        let c = entry_command(&entry, native)?;
+        if entry.id != version.decision_log_id
+            || c.object != object
+            || c.prepared != version.prepared
+            || events.get(c.operation_id.as_slice())?.is_none()
+        {
+            return Err(StoreError::CorruptRecord);
+        }
+        let head = required(heads.get(object_bytes(&object)?.as_slice())?)?;
+        let mut d = Decoder::new(head.value());
+        let head_id = d.uuid()?;
+        d.finish()?;
+        let current = decode_version(required(versions.get(version_key(&object, head_id)?.as_slice())?)?.value())?;
+        if current.decision_log_id.index < version.decision_log_id.index {
+            return Err(StoreError::CorruptRecord);
+        }
+    }
+    for row in heads.iter()? {
+        let (k, v) = row?;
+        let mut d = Decoder::new(k.value());
+        let object = d.object()?;
+        d.finish()?;
+        let mut d = Decoder::new(v.value());
+        let id = d.uuid()?;
+        d.finish()?;
+        decode_version(required(versions.get(version_key(&object, id)?.as_slice())?)?.value())?;
+    }
+    Ok(())
+}

crates/capture-authority/src/application_tests.rs

@@ -0,0 +1,531 @@
+use super::*;
+use crate::application::{APPLY_RESULTS, EVENTS, OPERATIONS, VERSIONS};
+use crate::tests::CrashBackend;
+use redb::ReadableTableMetadata;
+use std::{
+    io::{BufRead, BufReader, Write},
+    process::{Command, Stdio},
+    sync::mpsc,
+    time::Duration,
+};
+fn identity() -> NodeIdentity {
+    NodeIdentity {
+        cluster_id: [7; 16],
+        node_id: 0,
+    }
+}
+fn binding() -> CaptureBinding {
+    CaptureBinding {
+        revision: 9,
+        binding_id: [8; 16],
+        target_id: [9; 16],
+    }
+}
+fn command() -> DecideCreated {
+    DecideCreated {
+        operation_id: [1; 16],
+        object: ObjectIdentity {
+            bucket_incarnation: [2; 16],
+            key: vec![0, 255, 3],
+        },
+        expected_head: None,
+        prepared: PreparedIdentity {
+            version_id: [3; 16],
+            preparation_id: [4; 16],
+            content_digest: [5; 32],
+            content_length: 123,
+        },
+        binding_revision: 9,
+    }
+}
+fn id(index: u64) -> LogId {
+    LogId {
+        term: 8,
+        leader_node: 0,
+        index,
+    }
+}
+fn append(store: &mut Store, index: u64, c: &DecideCreated) {
+    store
+        .append(&[LogEntry {
+            id: id(index),
+            payload: c.encode().expect("encode command"),
+        }])
+        .expect("append command");
+}
+fn apply(store: &mut Store, index: u64, c: &DecideCreated) -> CreatedResult {
+    append(store, index, c);
+    store.save_committed(id(index)).expect("commit log");
+    store.apply_next_created(id(index)).expect("apply command")
+}
+fn success(c: &DecideCreated, index: u64) -> CreatedResult {
+    CreatedResult::Created {
+        version_id: c.prepared.version_id,
+        event_id: c.operation_id,
+        decision_log_id: id(index),
+    }
+}
+fn complete(store: &Store) {
+    let c = command();
+    assert_eq!(store.capture_binding().expect("binding"), Some(binding()));
+    assert_eq!(
+        store.created_head(&c.object).expect("head"),
+        Some(c.prepared.version_id),
+        "U11 acknowledged head lost"
+    );
+    assert_eq!(
+        store.created_version(&c.object, c.prepared.version_id).expect("version"),
+        Some(CreatedVersion {
+            prepared: c.prepared.clone(),
+            decision_log_id: id(0),
+            binding_revision: 9
+        })
+    );
+    assert_eq!(
+        store.created_event(c.operation_id).expect("event"),
+        Some(CreatedEvent {
+            object: c.object.clone(),
+            prepared: c.prepared.clone(),
+            decision_log_id: id(0),
+            binding: binding()
+        }),
+        "U11 complete event missing"
+    );
+    assert_eq!(
+        store.created_operation(c.operation_id).expect("operation"),
+        Some(CreatedOperation {
+            canonical_command: c.encode().expect("encode"),
+            result: success(&c, 0)
+        })
+    );
+    assert_eq!(
+        store.created_apply_result(0).expect("result"),
+        Some(CreatedApplyResult {
+            log_id: id(0),
+            result: success(&c, 0)
+        })
+    );
+    assert_eq!(store.last_applied().expect("last applied"), Some(id(0)));
+}
+fn fixture() -> (tempfile::TempDir, Store) {
+    let dir = tempfile::tempdir().expect("tempdir");
+    let mut store = Store::initialize(dir.path().join("db"), identity()).expect("initialize");
+    store.pin_capture_binding(binding()).expect("pin");
+    (dir, store)
+}
+fn reopen(path: &std::path::Path) -> Store {
+    let result = Store::open(path, identity());
+    assert!(result.is_ok(), "U11 complete reopen: {:?}", result.as_ref().err());
+    result.expect("asserted successful reopen")
+}
+#[test]
+fn created_codec_is_canonical_and_strict() {
+    let c = command();
+    let b = c.encode().expect("encode");
+    let mut golden = b"CRTD\x01".to_vec();
+    golden.extend([1; 16]);
+    golden.extend([2; 16]);
+    golden.extend(3u32.to_be_bytes());
+    golden.extend([0, 255, 3]);
+    golden.push(0);
+    golden.extend([3; 16]);
+    golden.extend([4; 16]);
+    golden.extend([5; 32]);
+    golden.extend(123u64.to_be_bytes());
+    golden.extend(9u64.to_be_bytes());
+    assert_eq!(b, golden);
+    assert_eq!(DecideCreated::decode(&b).expect("decode"), c);
+    let mut variants = Vec::new();
+    let mut v = c.clone();
+    v.operation_id = [6; 16];
+    variants.push(v);
+    let mut v = c.clone();
+    v.object.bucket_incarnation = [6; 16];
+    variants.push(v);
+    let mut v = c.clone();
+    v.object.key.push(4);
+    variants.push(v);
+    let mut v = c.clone();
+    v.expected_head = Some([6; 16]);
+    variants.push(v);
+    let mut v = c.clone();
+    v.prepared.version_id = [6; 16];
+    variants.push(v);
+    let mut v = c.clone();
+    v.prepared.preparation_id = [6; 16];
+    variants.push(v);
+    let mut v = c.clone();
+    v.prepared.content_digest = [6; 32];
+    variants.push(v);
+    let mut v = c.clone();
+    v.prepared.content_length += 1;
+    variants.push(v);
+    let mut v = c.clone();
+    v.binding_revision += 1;
+    variants.push(v);
+    for v in variants {
+        assert_ne!(v.encode().expect("variant"), b)
+    }
+    for n in 0..b.len() {
+        assert!(DecideCreated::decode(&b[..n]).is_err())
+    }
+    let mut bad = b.clone();
+    bad[4] = 2;
+    assert!(DecideCreated::decode(&bad).is_err());
+    let mut bad = b.clone();
+    bad[44] = 2;
+    assert!(DecideCreated::decode(&bad).is_err());
+    let mut bad = b.clone();
+    bad.push(0);
+    assert!(DecideCreated::decode(&bad).is_err());
+    let mut bad = b;
+    bad[37..41].copy_from_slice(&4097u32.to_be_bytes());
+    assert!(DecideCreated::decode(&bad).is_err());
+    let mut bad = c.clone();
+    bad.operation_id = [0; 16];
+    assert!(bad.encode().is_err());
+    bad = c;
+    bad.object.key = vec![0; MAX_OBJECT_KEY_LENGTH + 1];
+    assert!(bad.encode().is_err());
+}
+#[test]
+fn created_committed_order_and_identity_are_enforced() {
+    let (_dir, mut s) = fixture();
+    let c = command();
+    append(&mut s, 0, &c);
+    append(&mut s, 1, &c);
+    assert!(s.apply_next_created(id(0)).is_err());
+    assert_eq!(s.last_applied().expect("no apply"), None);
+    s.save_committed(id(1)).expect("committed");
+    assert!(matches!(s.apply_next_created(id(1)), Err(StoreError::InvalidApplySequence)));
+    let mut wrong = id(0);
+    wrong.term += 1;
+    assert!(s.apply_next_created(wrong).is_err());
+    wrong = id(0);
+    wrong.leader_node += 1;
+    assert!(s.apply_next_created(wrong).is_err());
+    assert_eq!(s.apply_next_created(id(0)).expect("first"), success(&c, 0));
+    complete(&s);
+    assert_eq!(s.apply_next_created(id(0)).expect("retry"), success(&c, 0));
+    complete(&s);
+    assert!(s.apply_next_created(wrong).is_err());
+    assert_eq!(s.apply_next_created(id(1)).expect("duplicate log"), success(&c, 0));
+    assert_eq!(s.last_applied().expect("last"), Some(id(1)));
+    assert_eq!(
+        s.created_apply_result(1).expect("result"),
+        Some(CreatedApplyResult {
+            log_id: id(1),
+            result: success(&c, 0)
+        })
+    );
+    let tx = s.database.begin_read().expect("read");
+    assert_eq!(tx.open_table(EVENTS).expect("events").len().expect("len"), 1);
+}
+#[test]
+fn created_head_version_event_and_dedup_reopen_together() {
+    let (dir, mut s) = fixture();
+    let c = command();
+    assert_eq!(apply(&mut s, 0, &c), success(&c, 0));
+    drop(s);
+    let mut s = reopen(&dir.path().join("db"));
+    complete(&s);
+    assert_eq!(apply(&mut s, 1, &c), success(&c, 0));
+    let mut changed = c.clone();
+    changed.prepared.content_digest = [6; 32];
+    assert_eq!(
+        apply(&mut s, 2, &changed),
+        CreatedResult::OperationConflict,
+        "U11 full canonical operation conflict"
+    );
+    assert_eq!(
+        s.created_operation(c.operation_id)
+            .expect("original")
+            .expect("operation")
+            .result,
+        success(&c, 0)
+    );
+    let mut stale = c.clone();
+    stale.operation_id = [6; 16];
+    stale.prepared.version_id = [6; 16];
+    assert_eq!(
+        apply(&mut s, 3, &stale),
+        CreatedResult::HeadMismatch {
+            actual_head: Some(c.prepared.version_id)
+        }
+    );
+    let mut reuse = c.clone();
+    reuse.operation_id = [7; 16];
+    reuse.expected_head = Some(c.prepared.version_id);
+    assert_eq!(apply(&mut s, 4, &reuse), CreatedResult::VersionConflict);
+    let mut replacement = stale.clone();
+    replacement.operation_id = [10; 16];
+    replacement.expected_head = Some(c.prepared.version_id);
+    assert_eq!(apply(&mut s, 5, &replacement), success(&replacement, 5));
+    assert_eq!(s.created_head(&c.object).expect("head"), Some(replacement.prepared.version_id));
+    assert_eq!(
+        s.created_version(&c.object, c.prepared.version_id)
+            .expect("old version")
+            .expect("version")
+            .decision_log_id,
+        id(0)
+    );
+    assert!(s.created_event(c.operation_id).expect("old event").is_some());
+    assert!(s.created_event(replacement.operation_id).expect("new event").is_some());
+    assert!(s.created_event(stale.operation_id).expect("rejection event").is_none());
+    assert_eq!(
+        apply(&mut s, 6, &stale),
+        CreatedResult::HeadMismatch {
+            actual_head: Some(c.prepared.version_id)
+        }
+    );
+    drop(s);
+    let s = reopen(&dir.path().join("db"));
+    assert_eq!(s.last_applied().expect("last"), Some(id(6)));
+    let tx = s.database.begin_read().expect("read");
+    assert_eq!(tx.open_table(EVENTS).expect("events").len().expect("len"), 2);
+}
+#[test]
+fn created_binding_is_pinned_and_not_caller_selected() {
+    let (dir, mut s) = fixture();
+    let mut bad = command();
+    bad.binding_revision = 10;
+    assert_eq!(apply(&mut s, 0, &bad), CreatedResult::BindingMismatch);
+    assert!(s.created_head(&bad.object).expect("head").is_none());
+    assert!(s.created_event(bad.operation_id).expect("event").is_none());
+    assert_eq!(
+        s.created_operation(bad.operation_id)
+            .expect("operation")
+            .expect("saved")
+            .result,
+        CreatedResult::BindingMismatch
+    );
+    s.pin_capture_binding(binding()).expect("idempotent pin");
+    let mut changed = binding();
+    changed.target_id = [10; 16];
+    assert!(matches!(s.pin_capture_binding(changed), Err(StoreError::BindingConflict)));
+    drop(s);
+    let s = reopen(&dir.path().join("db"));
+    assert_eq!(s.capture_binding().expect("binding"), Some(binding()));
+}
+#[test]
+fn created_schema_compatibility_and_corruption_are_checked() {
+    let dir = tempfile::tempdir().expect("dir");
+    let path = dir.path().join("db");
+    let s = Store::initialize(&path, identity()).expect("schema1");
+    assert_eq!(s.capture_binding().expect("unmigrated"), None);
+    drop(s);
+    let mut s = reopen(&path);
+    assert_eq!(s.last_applied().expect("no apply"), None);
+    append(&mut s, 0, &command());
+    s.save_committed(id(0)).expect("schema2");
+    drop(s);
+    let mut s = reopen(&path);
+    assert_eq!(s.capture_binding().expect("schema2 unchanged"), None);
+    s.pin_capture_binding(binding()).expect("explicit upgrade");
+    drop(s);
+    let mut s = reopen(&path);
+    assert_eq!(s.capture_binding().expect("schema3"), Some(binding()));
+    s.apply_next_created(id(0)).expect("apply");
+    drop(s);
+    let db = Database::open(&path).expect("raw db");
+    let tx = db.begin_write().expect("write corruption");
+    tx.open_table(EVENTS)
+        .expect("events")
+        .remove(command().operation_id.as_slice())
+        .expect("remove reference");
+    tx.commit().expect("commit corruption");
+    drop(db);
+    let result = Store::open(&path, identity());
+    assert!(
+        matches!(result, Err(StoreError::CorruptRecord)),
+        "U11 missing event rejected before expect"
+    );
+    let (dir, mut s) = fixture();
+    assert_eq!(s.read_committed().expect("empty committed"), None);
+    append(&mut s, 0, &command());
+    s.save_committed(id(0)).expect("committed schema3");
+    assert_eq!(s.capture_binding().expect("no downgrade"), Some(binding()));
+    drop(s);
+    let path = dir.path().join("db");
+    let db = Database::open(&path).expect("raw db");
+    let tx = db.begin_write().expect("write");
+    let mut marker = vec![1];
+    marker.extend(encode_id(id(0)));
+    tx.open_table(META)
+        .expect("meta")
+        .insert("last_applied", marker.as_slice())
+        .expect("corrupt applied");
+    tx.commit().expect("commit");
+    drop(db);
+    assert!(
+        matches!(Store::open(&path, identity()), Err(StoreError::CorruptRecord)),
+        "U11 last applied missing result rejected"
+    );
+    let (dir, s) = fixture();
+    drop(s);
+    let path = dir.path().join("db");
+    let db = Database::open(&path).expect("raw pinned db");
+    let tx = db.begin_write().expect("missing binding write");
+    tx.open_table(META)
+        .expect("meta")
+        .remove("capture_binding")
+        .expect("remove binding");
+    tx.commit().expect("binding corruption commit");
+    drop(db);
+    assert!(
+        matches!(Store::open(&path, identity()), Err(StoreError::CorruptRecord)),
+        "U11 missing pinned binding rejected"
+    );
+
+    let (dir, mut s) = fixture();
+    let c = command();
+    assert_eq!(apply(&mut s, 0, &c), success(&c, 0));
+    assert_eq!(apply(&mut s, 1, &c), success(&c, 0));
+    drop(s);
+    let path = dir.path().join("db");
+    let db = Database::open(&path).expect("raw duplicate db");
+    let tx = db.begin_write().expect("retarget decision write");
+    let mut key = created::object_bytes(&c.object).expect("object key");
+    key.extend(c.prepared.version_id);
+    let mut version = Vec::new();
+    created::put_prepared(&mut version, &c.prepared);
+    version.extend(encode_id(id(1)));
+    version.extend(binding().revision.to_be_bytes());
+    tx.open_table(VERSIONS)
+        .expect("versions")
+        .insert(key.as_slice(), version.as_slice())
+        .expect("retarget version");
+    let mut event = created::object_bytes(&c.object).expect("event object");
+    created::put_prepared(&mut event, &c.prepared);
+    event.extend(encode_id(id(1)));
+    event.extend(created::binding_bytes(&binding()).expect("binding bytes"));
+    tx.open_table(EVENTS)
+        .expect("events")
+        .insert(c.operation_id.as_slice(), event.as_slice())
+        .expect("retarget event");
+    let canonical = c.encode().expect("canonical");
+    let mut operation = u64::try_from(canonical.len()).expect("length").to_be_bytes().to_vec();
+    operation.extend(canonical);
+    operation.extend(created::result_bytes(&success(&c, 1)));
+    tx.open_table(OPERATIONS)
+        .expect("operations")
+        .insert(c.operation_id.as_slice(), operation.as_slice())
+        .expect("retarget original response");
+    for index in 0..=1 {
+        let mut result = encode_id(id(index));
+        result.extend(created::result_bytes(&success(&c, 1)));
+        tx.open_table(APPLY_RESULTS)
+            .expect("results")
+            .insert(index, result.as_slice())
+            .expect("retarget duplicate response");
+    }
+    tx.commit().expect("coherent retarget commit");
+    drop(db);
+    assert!(
+        matches!(Store::open(&path, identity()), Err(StoreError::CorruptRecord)),
+        "U11 original decision cannot retarget duplicate log"
+    );
+}
+#[test]
+fn created_process_child() {
+    if let Some(path) = std::env::var_os("U11_CHILD_DB") {
+        let mut s = Store::initialize(path, identity()).expect("child init");
+        s.pin_capture_binding(binding()).expect("child pin");
+        assert_eq!(apply(&mut s, 0, &command()), success(&command(), 0));
+        complete(&s);
+        println!("U11_ACK");
+        std::io::stdout().flush().expect("flush ACK");
+        loop {
+            std::thread::park()
+        }
+    }
+    let (_dir, mut s) = fixture();
+    assert_eq!(apply(&mut s, 0, &command()), success(&command(), 0));
+    complete(&s);
+}
+#[test]
+fn created_acknowledged_apply_survives_child_kill() {
+    let dir = tempfile::tempdir().expect("dir");
+    let path = dir.path().join("db");
+    let mut child = Command::new(std::env::current_exe().expect("test executable"))
+        .args(["--exact", "application_tests::created_process_child", "--nocapture"])
+        .env("U11_CHILD_DB", &path)
+        .stdout(Stdio::piped())
+        .stderr(Stdio::inherit())
+        .spawn()
+        .expect("spawn");
+    let stdout = child.stdout.take().expect("stdout");
+    let (tx, rx) = mpsc::channel();
+    let reader = std::thread::spawn(move || {
+        for line in BufReader::new(stdout).lines() {
+            let line = line.expect("child line");
+            if line.ends_with("U11_ACK") {
+                let _ = tx.send(());
+                break;
+            }
+        }
+    });
+    let acknowledged = rx.recv_timeout(Duration::from_secs(20));
+    child.kill().expect("kill child");
+    child.wait().expect("wait child");
+    reader.join().expect("reader cleanup");
+    assert!(acknowledged.is_ok(), "U11 bounded child ACK missing");
+    complete(&reopen(&path));
+}
+#[test]
+fn created_power_cut_before_and_after_apply_flush() {
+    let (dir, mut s) = fixture();
+    append(&mut s, 0, &command());
+    s.save_committed(id(0)).expect("committed");
+    drop(s);
+    let path = dir.path().join("db");
+    let backend = CrashBackend::new(&path);
+    let mut s = Store::from_database(backend.database(), identity()).expect("backend");
+    assert_eq!(s.apply_next_created(id(0)).expect("apply ACK"), success(&command(), 0));
+    backend.cut();
+    drop(s);
+    complete(&reopen(&path));
+    for sync in 1..=2 {
+        for after in [false, true] {
+            let (dir, mut s) = fixture();
+            append(&mut s, 0, &command());
+            s.save_committed(id(0)).expect("committed");
+            drop(s);
+            let path = dir.path().join("db");
+            let backend = CrashBackend::new(&path);
+            let mut s = Store::from_database(backend.database(), identity()).expect("backend");
+            let (entered, resume) = backend.pause_sync(sync, after);
+            let writer = std::thread::spawn(move || {
+                let result = s.apply_next_created(id(0));
+                (s, result)
+            });
+            entered.wait();
+            backend.cut();
+            resume.wait();
+            let (s, result) = writer.join().expect("writer");
+            drop(s);
+            let reopened = Store::open(&path, identity());
+            assert!(
+                reopened.is_ok(),
+                "U11 apply cut must reopen whole old/new state sync={sync} after={after}: {:?}",
+                reopened.as_ref().err()
+            );
+            let s = reopened.expect("asserted reopen");
+            if s.last_applied().expect("last").is_some() {
+                complete(&s)
+            } else {
+                assert!(result.is_err());
+                assert_eq!(s.created_head(&command().object).expect("old head"), None);
+                assert_eq!(
+                    s.created_version(&command().object, command().prepared.version_id)
+                        .expect("old version"),
+                    None
+                );
+                assert_eq!(s.created_event(command().operation_id).expect("old event"), None);
+                assert_eq!(s.created_operation(command().operation_id).expect("old op"), None);
+                assert_eq!(s.created_apply_result(0).expect("old result"), None)
+            }
+            eprintln!("U11 cut boundary sync={sync} after={after}");
+        }
+    }
+}

crates/capture-authority/src/committed.rs

@@ -0,0 +1,63 @@
+use crate::{LOGS, LogId, META, Store, StoreError, decode_id, decode_log, encode_id, number};
+use redb::{ReadableDatabase, ReadableTable};
+
+pub(crate) fn validate_committed(
+    meta: &impl ReadableTable<&'static str, &'static [u8]>,
+    logs: &impl ReadableTable<u64, &'static [u8]>,
+) -> Result<Option<LogId>, StoreError> {
+    let schema = number(meta.get("schema")?.ok_or(StoreError::CorruptRecord)?.value())?;
+    let marker = meta.get("committed")?;
+    match (schema, marker) {
+        (1 | 3 | 4, None) => {
+            if matches!(schema, 3 | 4) && meta.get("last_applied")?.is_none_or(|v| v.value() != [0]) {
+                return Err(StoreError::CorruptRecord);
+            }
+            Ok(None)
+        }
+        (2 | 3 | 4, Some(marker)) => {
+            let id = decode_id(marker.value())?;
+            let record = logs.get(id.index)?.ok_or(StoreError::CorruptRecord)?;
+            if decode_log(record.value())?.id != id {
+                return Err(StoreError::CorruptRecord);
+            }
+            Ok(Some(id))
+        }
+        (1 | 2, _) => Err(StoreError::CorruptRecord),
+        _ => Err(StoreError::UnsupportedSchema),
+    }
+}
+
+impl Store {
+    /// Read the exact committed boundary, validating it against the stored log.
+    pub fn read_committed(&self) -> Result<Option<LogId>, StoreError> {
+        let tx = self.database.begin_read()?;
+        validate_committed(&tx.open_table(META)?, &tx.open_table(LOGS)?)
+    }
+
+    /// Persist a monotonic committed boundary supplied by the authority owner.
+    pub fn save_committed(&mut self, id: LogId) -> Result<(), StoreError> {
+        let tx = self.write_transaction()?;
+        {
+            let mut meta = tx.open_table(META)?;
+            let logs = tx.open_table(LOGS)?;
+            let previous = validate_committed(&meta, &logs)?;
+            let record = logs.get(id.index)?.ok_or(StoreError::InvalidCommittedLog)?;
+            if decode_log(record.value())?.id != id {
+                return Err(StoreError::InvalidCommittedLog);
+            }
+            if let Some(previous) = previous {
+                if id.index < previous.index {
+                    return Err(StoreError::CommittedRegression);
+                }
+                if id == previous {
+                    return Ok(());
+                }
+            }
+            meta.insert("committed", encode_id(id).as_slice())?;
+            if previous.is_none() && !matches!(number(meta.get("schema")?.ok_or(StoreError::CorruptRecord)?.value())?, 3 | 4) {
+                meta.insert("schema", 2u64.to_be_bytes().as_slice())?;
+            }
+        }
+        self.commit(tx)
+    }
+}

crates/capture-authority/src/created.rs

@@ -0,0 +1,242 @@
+use crate::{LogId, StoreError, decode_id, encode_id};
+
+pub const MAX_OBJECT_KEY_LENGTH: usize = 4096;
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct ObjectIdentity {
+    pub bucket_incarnation: [u8; 16],
+    pub key: Vec<u8>,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct PreparedIdentity {
+    pub version_id: [u8; 16],
+    pub preparation_id: [u8; 16],
+    pub content_digest: [u8; 32],
+    pub content_length: u64,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct DecideCreated {
+    pub operation_id: [u8; 16],
+    pub object: ObjectIdentity,
+    pub expected_head: Option<[u8; 16]>,
+    pub prepared: PreparedIdentity,
+    pub binding_revision: u64,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct CaptureBinding {
+    pub revision: u64,
+    pub binding_id: [u8; 16],
+    pub target_id: [u8; 16],
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub enum CreatedResult {
+    Created {
+        version_id: [u8; 16],
+        event_id: [u8; 16],
+        decision_log_id: LogId,
+    },
+    HeadMismatch {
+        actual_head: Option<[u8; 16]>,
+    },
+    OperationConflict,
+    BindingMismatch,
+    VersionConflict,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct CreatedVersion {
+    pub prepared: PreparedIdentity,
+    pub decision_log_id: LogId,
+    pub binding_revision: u64,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct CreatedEvent {
+    pub object: ObjectIdentity,
+    pub prepared: PreparedIdentity,
+    pub decision_log_id: LogId,
+    pub binding: CaptureBinding,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct CreatedOperation {
+    pub canonical_command: Vec<u8>,
+    pub result: CreatedResult,
+}
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct CreatedApplyResult {
+    pub log_id: LogId,
+    pub result: CreatedResult,
+}
+
+pub(crate) struct Decoder<'a> {
+    bytes: &'a [u8],
+}
+impl<'a> Decoder<'a> {
+    pub(crate) fn new(bytes: &'a [u8]) -> Self {
+        Self { bytes }
+    }
+    pub(crate) fn take<const N: usize>(&mut self) -> Result<[u8; N], StoreError> {
+        let a = self
+            .bytes
+            .get(..N)
+            .ok_or(StoreError::CorruptRecord)?
+            .try_into()
+            .map_err(|_| StoreError::CorruptRecord)?;
+        self.bytes = &self.bytes[N..];
+        Ok(a)
+    }
+    pub(crate) fn u64(&mut self) -> Result<u64, StoreError> {
+        Ok(u64::from_be_bytes(self.take()?))
+    }
+    pub(crate) fn uuid(&mut self) -> Result<[u8; 16], StoreError> {
+        let id = self.take()?;
+        if id == [0; 16] {
+            return Err(StoreError::CorruptRecord);
+        }
+        Ok(id)
+    }
+    pub(crate) fn option(&mut self) -> Result<Option<[u8; 16]>, StoreError> {
+        match self.take::<1>()?[0] {
+            0 => Ok(None),
+            1 => Ok(Some(self.uuid()?)),
+            _ => Err(StoreError::CorruptRecord),
+        }
+    }
+    pub(crate) fn object(&mut self) -> Result<ObjectIdentity, StoreError> {
+        let bucket_incarnation = self.uuid()?;
+        let len = usize::try_from(u32::from_be_bytes(self.take()?)).map_err(|_| StoreError::CorruptRecord)?;
+        if len > MAX_OBJECT_KEY_LENGTH {
+            return Err(StoreError::CorruptRecord);
+        }
+        let key = self.bytes.get(..len).ok_or(StoreError::CorruptRecord)?.to_vec();
+        self.bytes = &self.bytes[len..];
+        Ok(ObjectIdentity { bucket_incarnation, key })
+    }
+    pub(crate) fn prepared(&mut self) -> Result<PreparedIdentity, StoreError> {
+        Ok(PreparedIdentity {
+            version_id: self.uuid()?,
+            preparation_id: self.uuid()?,
+            content_digest: self.take()?,
+            content_length: self.u64()?,
+        })
+    }
+    pub(crate) fn id(&mut self) -> Result<LogId, StoreError> {
+        decode_id(&self.take::<24>()?)
+    }
+    pub(crate) fn binding(&mut self) -> Result<CaptureBinding, StoreError> {
+        Ok(CaptureBinding {
+            revision: self.u64()?,
+            binding_id: self.uuid()?,
+            target_id: self.uuid()?,
+        })
+    }
+    pub(crate) fn finish(self) -> Result<(), StoreError> {
+        if self.bytes.is_empty() {
+            Ok(())
+        } else {
+            Err(StoreError::CorruptRecord)
+        }
+    }
+    pub(crate) fn result(&mut self) -> Result<CreatedResult, StoreError> {
+        match self.take::<1>()?[0] {
+            0 => Ok(CreatedResult::Created {
+                version_id: self.uuid()?,
+                event_id: self.uuid()?,
+                decision_log_id: self.id()?,
+            }),
+            1 => Ok(CreatedResult::HeadMismatch {
+                actual_head: self.option()?,
+            }),
+            2 => Ok(CreatedResult::OperationConflict),
+            3 => Ok(CreatedResult::BindingMismatch),
+            4 => Ok(CreatedResult::VersionConflict),
+            _ => Err(StoreError::CorruptRecord),
+        }
+    }
+}
+pub(crate) fn put_option(b: &mut Vec<u8>, id: Option<[u8; 16]>) {
+    b.push(u8::from(id.is_some()));
+    if let Some(id) = id {
+        b.extend(id)
+    }
+}
+pub(crate) fn object_bytes(o: &ObjectIdentity) -> Result<Vec<u8>, StoreError> {
+    if o.bucket_incarnation == [0; 16] || o.key.len() > MAX_OBJECT_KEY_LENGTH {
+        return Err(StoreError::CorruptRecord);
+    }
+    let capacity = 20usize.checked_add(o.key.len()).ok_or(StoreError::CorruptRecord)?;
+    let mut b = Vec::with_capacity(capacity);
+    b.extend(o.bucket_incarnation);
+    b.extend(
+        u32::try_from(o.key.len())
+            .map_err(|_| StoreError::CorruptRecord)?
+            .to_be_bytes(),
+    );
+    b.extend(&o.key);
+    Ok(b)
+}
+pub(crate) fn put_prepared(b: &mut Vec<u8>, p: &PreparedIdentity) {
+    b.extend(p.version_id);
+    b.extend(p.preparation_id);
+    b.extend(p.content_digest);
+    b.extend(p.content_length.to_be_bytes())
+}
+pub(crate) fn binding_bytes(v: &CaptureBinding) -> Result<Vec<u8>, StoreError> {
+    let mut b = v.revision.to_be_bytes().to_vec();
+    b.extend(v.binding_id);
+    b.extend(v.target_id);
+    let mut d = Decoder::new(&b);
+    d.binding()?;
+    d.finish()?;
+    Ok(b)
+}
+pub(crate) fn result_bytes(r: &CreatedResult) -> Vec<u8> {
+    let mut b = Vec::new();
+    match r {
+        CreatedResult::Created {
+            version_id,
+            event_id,
+            decision_log_id,
+        } => {
+            b.push(0);
+            b.extend(version_id);
+            b.extend(event_id);
+            b.extend(encode_id(*decision_log_id))
+        }
+        CreatedResult::HeadMismatch { actual_head } => {
+            b.push(1);
+            put_option(&mut b, *actual_head)
+        }
+        CreatedResult::OperationConflict => b.push(2),
+        CreatedResult::BindingMismatch => b.push(3),
+        CreatedResult::VersionConflict => b.push(4),
+    }
+    b
+}
+impl DecideCreated {
+    pub fn encode(&self) -> Result<Vec<u8>, StoreError> {
+        let object = object_bytes(&self.object)?;
+        let capacity = object.len().checked_add(134).ok_or(StoreError::CorruptRecord)?;
+        let mut b = Vec::with_capacity(capacity);
+        b.extend(b"CRTD\x01");
+        b.extend(self.operation_id);
+        b.extend(object);
+        put_option(&mut b, self.expected_head);
+        put_prepared(&mut b, &self.prepared);
+        b.extend(self.binding_revision.to_be_bytes());
+        Self::decode(&b)?;
+        Ok(b)
+    }
+    pub fn decode(bytes: &[u8]) -> Result<Self, StoreError> {
+        let mut d = Decoder::new(bytes);
+        if d.take::<5>()? != *b"CRTD\x01" {
+            return Err(StoreError::CorruptRecord);
+        }
+        let c = Self {
+            operation_id: d.uuid()?,
+            object: d.object()?,
+            expected_head: d.option()?,
+            prepared: d.prepared()?,
+            binding_revision: d.u64()?,
+        };
+        d.finish()?;
+        Ok(c)
+    }
+}

crates/capture-authority/src/native_snapshot.rs

@@ -0,0 +1,26 @@
+//! Owned consistent native application snapshot view.
+use crate::{StoreError, snapshot};
+use std::io::Write;
+
+#[derive(Debug, thiserror::Error)]
+pub enum NativeSnapshotError {
+    #[error(transparent)]
+    Store(#[from] StoreError),
+    #[error(transparent)]
+    Snapshot(#[from] snapshot::SnapshotError),
+    #[error("native snapshot Store lock poisoned")]
+    Poisoned,
+    #[error("native snapshot worker failed")]
+    Worker(#[source] tokio::task::JoinError),
+    #[error("native snapshot view unavailable")]
+    Unavailable,
+}
+
+pub struct NativeSnapshotView {
+    _private: (),
+}
+impl NativeSnapshotView {
+    pub fn stream<W: Write>(self, _writer: W, _limits: snapshot::Limits) -> Result<W, NativeSnapshotError> {
+        Err(NativeSnapshotError::Unavailable)
+    }
+}

crates/capture-authority/src/node.rs

@@ -0,0 +1,371 @@
+//! Validated capture peer metadata and its independent CNOD v1 codec.
+use serde::{Deserialize, Deserializer, Serialize, Serializer, ser::SerializeStruct};
+
+#[derive(Clone, Debug, Default, PartialEq, Eq)]
+pub struct CaptureNode {
+    rpc_uri: String,
+}
+
+#[derive(Debug, thiserror::Error)]
+pub enum CaptureNodeError {
+    #[error("invalid RPC URI")]
+    InvalidRpcUri,
+    #[error("invalid RPC URI")]
+    InvalidRpcUriParse(#[source] url::ParseError),
+    #[error("RPC URI too long")]
+    RpcUriTooLong,
+    #[error("invalid node magic")]
+    InvalidNodeMagic,
+    #[error("unsupported node version")]
+    UnsupportedNodeVersion,
+    #[error("invalid node length")]
+    InvalidNodeLength,
+    #[error("truncated node record")]
+    TruncatedNodeRecord,
+    #[error("invalid node UTF-8")]
+    InvalidNodeUtf8(#[source] std::str::Utf8Error),
+    #[error("trailing node bytes")]
+    TrailingNodeBytes,
+}
+impl CaptureNodeError {
+    pub fn code(&self) -> &'static str {
+        match self {
+            Self::InvalidRpcUri | Self::InvalidRpcUriParse(_) => "invalid_rpc_uri",
+            Self::RpcUriTooLong => "rpc_uri_too_long",
+            Self::InvalidNodeMagic => "invalid_node_magic",
+            Self::UnsupportedNodeVersion => "unsupported_node_version",
+            Self::InvalidNodeLength => "invalid_node_length",
+            Self::TruncatedNodeRecord => "truncated_node_record",
+            Self::InvalidNodeUtf8(_) => "invalid_node_utf8",
+            Self::TrailingNodeBytes => "trailing_node_bytes",
+        }
+    }
+}
+impl CaptureNode {
+    pub fn new(rpc_uri: String) -> Result<Self, CaptureNodeError> {
+        let node = Self { rpc_uri };
+        node.validate()?;
+        Ok(node)
+    }
+
+    pub fn rpc_uri(&self) -> &str {
+        &self.rpc_uri
+    }
+
+    pub fn validate(&self) -> Result<(), CaptureNodeError> {
+        let uri = &self.rpc_uri;
+        if uri.len() > 4096 {
+            return Err(CaptureNodeError::RpcUriTooLong);
+        }
+        if uri.is_empty() || uri.chars().any(|c| c.is_whitespace() || c.is_control() || c == '\\') {
+            return Err(CaptureNodeError::InvalidRpcUri);
+        }
+        let rest = uri
+            .strip_prefix("http://")
+            .or_else(|| uri.strip_prefix("https://"))
+            .ok_or(CaptureNodeError::InvalidRpcUri)?;
+        if uri.contains('?') || uri.contains('#') {
+            return Err(CaptureNodeError::InvalidRpcUri);
+        }
+        let (authority, path) = rest.split_once('/').map_or((rest, ""), |(a, p)| (a, p));
+        // An empty suffix after the first slash is the only allowed path.
+        if !path.is_empty() || authority.contains('@') {
+            return Err(CaptureNodeError::InvalidRpcUri);
+        }
+        let (host, port) = if authority.starts_with('[') {
+            let end = authority.find(']').ok_or(CaptureNodeError::InvalidRpcUri)?;
+            let port = authority
+                .get(end + 1..)
+                .and_then(|suffix| suffix.strip_prefix(':'))
+                .ok_or(CaptureNodeError::InvalidRpcUri)?;
+            (&authority[..=end], port)
+        } else {
+            authority.split_once(':').ok_or(CaptureNodeError::InvalidRpcUri)?
+        };
+        if host.is_empty() || port.is_empty() || !port.bytes().all(|b| b.is_ascii_digit()) {
+            return Err(CaptureNodeError::InvalidRpcUri);
+        }
+        let port = port.parse::<u16>().map_err(|_| CaptureNodeError::InvalidRpcUri)?;
+        if port == 0 {
+            return Err(CaptureNodeError::InvalidRpcUri);
+        }
+        let parsed_url = url::Url::parse(uri).map_err(CaptureNodeError::InvalidRpcUriParse)?;
+        if parsed_url.host().is_none() {
+            return Err(CaptureNodeError::InvalidRpcUri);
+        }
+        Ok(())
+    }
+
+    pub fn encode(&self) -> Result<Vec<u8>, CaptureNodeError> {
+        self.validate()?;
+        let length = u32::try_from(self.rpc_uri.len()).map_err(|_| CaptureNodeError::InvalidNodeLength)?;
+        let mut bytes = b"CNOD\x01".to_vec();
+        bytes.extend_from_slice(&length.to_be_bytes());
+        bytes.extend_from_slice(self.rpc_uri.as_bytes());
+        Ok(bytes)
+    }
+
+    pub fn decode(bytes: &[u8]) -> Result<Self, CaptureNodeError> {
+        if bytes.len() < 9 {
+            return Err(CaptureNodeError::TruncatedNodeRecord);
+        }
+        if &bytes[..4] != b"CNOD" {
+            return Err(CaptureNodeError::InvalidNodeMagic);
+        }
+        if bytes[4] != 1 {
+            return Err(CaptureNodeError::UnsupportedNodeVersion);
+        }
+        let length = u32::from_be_bytes(bytes[5..9].try_into().map_err(|_| CaptureNodeError::InvalidNodeLength)?);
+        if !(1..=4096).contains(&length) {
+            return Err(CaptureNodeError::InvalidNodeLength);
+        }
+        let length = usize::try_from(length).map_err(|_| CaptureNodeError::InvalidNodeLength)?;
+        let end = 9_usize.checked_add(length).ok_or(CaptureNodeError::InvalidNodeLength)?;
+        if bytes.len() < end {
+            return Err(CaptureNodeError::TruncatedNodeRecord);
+        }
+        if bytes.len() != end {
+            return Err(CaptureNodeError::TrailingNodeBytes);
+        }
+        let uri = std::str::from_utf8(&bytes[9..end]).map_err(CaptureNodeError::InvalidNodeUtf8)?;
+        Self::new(uri.to_owned())
+    }
+}
+impl Serialize for CaptureNode {
+    fn serialize<S: Serializer>(&self, serializer: S) -> Result<S::Ok, S::Error> {
+        self.validate().map_err(serde::ser::Error::custom)?;
+        let mut state = serializer.serialize_struct("CaptureNode", 1)?;
+        state.serialize_field("rpc_uri", &self.rpc_uri)?;
+        state.end()
+    }
+}
+impl<'de> Deserialize<'de> for CaptureNode {
+    fn deserialize<D: Deserializer<'de>>(deserializer: D) -> Result<Self, D::Error> {
+        #[derive(Deserialize)]
+        #[serde(deny_unknown_fields)]
+        struct Fields {
+            rpc_uri: String,
+        }
+        let fields = Fields::deserialize(deserializer).map_err(|_| serde::de::Error::custom("invalid capture node fields"))?;
+        Self::new(fields.rpc_uri).map_err(serde::de::Error::custom)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::error::Error;
+
+    fn record(payload: &[u8]) -> Vec<u8> {
+        let mut bytes = b"CNOD\x01".to_vec();
+        bytes.extend_from_slice(&u32::try_from(payload.len()).expect("test payload length").to_be_bytes());
+        bytes.extend_from_slice(payload);
+        bytes
+    }
+    fn reject_uri(uri: &str) {
+        let result = CaptureNode::new(uri.to_owned());
+        assert!(result.is_err(), "URI should reject: {uri:?}");
+        assert_eq!(result.expect_err("rejected URI").code(), "invalid_rpc_uri");
+    }
+    fn reject_record(bytes: &[u8], code: &str) {
+        let result = CaptureNode::decode(bytes);
+        assert!(result.is_err(), "record should reject with {code}");
+        assert_eq!(result.expect_err("rejected record").code(), code);
+    }
+    #[test]
+    fn actual_openraft_node_bound() {
+        fn require_node<T: openraft::Node>() {}
+        require_node::<CaptureNode>();
+        assert_eq!(CaptureNode::default().rpc_uri(), "");
+    }
+    #[test]
+    fn default_and_empty_rejected_at_every_boundary() {
+        let node = CaptureNode::default();
+        assert_eq!(node.rpc_uri(), "");
+        reject_uri("");
+        assert_eq!(node.validate().expect_err("empty validation").code(), "invalid_rpc_uri");
+        assert!(serde_json::to_string(&node).is_err());
+        assert!(serde_json::from_str::<CaptureNode>(r#"{"rpc_uri":""}"#).is_err());
+        assert_eq!(node.encode().expect_err("empty encoding").code(), "invalid_rpc_uri");
+        reject_record(&record(b""), "invalid_node_length");
+    }
+    #[test]
+    fn accepted_uris_preserve_exact_bytes() {
+        for uri in [
+            "http://node:80",
+            "https://node:443",
+            "http://node:1",
+            "https://node:65535/",
+            "http://127.0.0.1:9000",
+            "https://[::1]:443/",
+            "http://[2001:db8::1]:1",
+            "http://NoDe:00080/",
+        ] {
+            let result = CaptureNode::new(uri.to_owned());
+            assert!(result.is_ok(), "valid URI construction: {uri}");
+            let node = result.expect("valid URI");
+            assert_eq!(node.rpc_uri(), uri, "preserve explicit default port and raw URI");
+            assert_eq!(node.encode().expect("encode URI"), record(uri.as_bytes()));
+            assert_eq!(CaptureNode::decode(&record(uri.as_bytes())).expect("decode URI"), node);
+        }
+    }
+    #[test]
+    fn golden_cnod_bytes() {
+        let node = CaptureNode {
+            rpc_uri: "http://node:80".to_owned(),
+        };
+        let golden = b"CNOD\x01\x00\x00\x00\x0ehttp://node:80";
+        assert_eq!(node.encode().ok().as_deref(), Some(golden.as_slice()), "golden CNOD bytes");
+    }
+    #[test]
+    fn valid_cnod_decode() {
+        let golden = b"CNOD\x01\x00\x00\x00\x0ehttp://node:80";
+        let decoded = CaptureNode::decode(golden);
+        assert!(decoded.is_ok(), "golden decode succeeds");
+        assert_eq!(decoded.expect("golden decoded node").rpc_uri(), "http://node:80");
+    }
+    #[test]
+    fn serde_shape_roundtrip_and_strict_fields() {
+        let node = CaptureNode::new("https://node:443/".to_owned()).expect("serde node");
+        let json = r#"{"rpc_uri":"https://node:443/"}"#;
+        assert_eq!(serde_json::to_string(&node).expect("serialize node"), json);
+        assert_eq!(serde_json::from_str::<CaptureNode>(json).expect("deserialize node"), node);
+        for json in [
+            r#"{}"#,
+            r#"{"rpc_uri":1}"#,
+            r#"{"rpc_uri":null}"#,
+            r#"{"rpc_uri":"http://node:1","extra":0}"#,
+            r#"{"rpc_uri":"http://node:1","rpc_uri":"http://node:2"}"#,
+        ] {
+            assert!(serde_json::from_str::<CaptureNode>(json).is_err(), "strict fields: {json}");
+        }
+    }
+    #[test]
+    fn credentials_and_raw_uri_shapes_reject() {
+        for uri in [
+            "http://user@node:1",
+            "http://user:password@node:1",
+            "http://@node:1",
+            "HTTP://node:1",
+            "Https://node:1",
+            "http://node",
+            "http://node:",
+            "http://node:0",
+            "http://node:-1",
+            "http://node:+1",
+            "http://node:65536",
+            "http://node:999999999999999999999",
+            "http://[::1:1",
+            "http://[::1]extra:1",
+            "http://[invalid]:1",
+            "http://::1:1",
+            "http://:1",
+            "/disk/path",
+            "file:///disk/path",
+            "http://node:1/./",
+            "http://node:1/../",
+            "http://node:1/%2e/",
+            "http://node:1/%2E%2e/",
+            "http://node:1/path",
+            "http://node:1\\",
+            " http://node:1",
+            "http://node:1\n",
+            "http://no\tde:1",
+            "http://node:1\u{00a0}",
+            "http://node:1\u{0085}",
+            "http://node:1\0",
+        ] {
+            reject_uri(uri);
+        }
+    }
+    #[test]
+    fn query_and_fragment_markers_reject() {
+        for uri in [
+            "http://node:1?",
+            "http://node:1?x=y",
+            "http://node:1/?",
+            "http://node:1#",
+            "http://node:1#fragment",
+            "http://node:1/#",
+        ] {
+            reject_uri(uri);
+        }
+    }
+    #[test]
+    fn utf8_byte_length_boundaries() {
+        for length in [4095, 4096] {
+            let prefix = "http://node:";
+            let uri = format!("{prefix}{}1", "0".repeat(length - prefix.len() - 1));
+            assert_eq!(uri.len(), length);
+            assert!(url::Url::parse(&uri).is_ok(), "boundary fixture accepted by actual URL parser");
+            let node = CaptureNode::new(uri.clone()).expect("boundary node");
+            assert_eq!(node.rpc_uri(), uri);
+            assert_eq!(
+                CaptureNode::decode(&node.encode().expect("boundary encoding")).expect("boundary decoding"),
+                node
+            );
+        }
+        let uri = format!("http://node:{}1", "0".repeat(4097 - "http://node:".len() - 1));
+        assert_eq!(CaptureNode::new(uri).expect_err("oversize URI").code(), "rpc_uri_too_long");
+        let unicode = format!("http://{}:1", "é".repeat(2044));
+        assert!(unicode.chars().count() < 4096);
+        assert!(unicode.len() > 4096);
+        assert_eq!(CaptureNode::new(unicode).expect_err("UTF-8 byte limit").code(), "rpc_uri_too_long");
+    }
+    #[test]
+    fn cnod_headers_and_lengths_reject_exact_codes() {
+        let valid = record(b"http://node:80");
+        for end in 0..9 {
+            reject_record(&valid[..end], "truncated_node_record");
+        }
+        let mut bad = valid.clone();
+        bad[0] = b'X';
+        reject_record(&bad, "invalid_node_magic");
+        bad = valid.clone();
+        bad[4] = 2;
+        reject_record(&bad, "unsupported_node_version");
+        for length in [0_u32, 4097, u32::MAX] {
+            bad = b"CNOD\x01".to_vec();
+            bad.extend_from_slice(&length.to_be_bytes());
+            reject_record(&bad, "invalid_node_length");
+        }
+        for end in 9..valid.len() {
+            reject_record(&valid[..end], "truncated_node_record");
+        }
+        reject_record(&record(&[0xff]), "invalid_node_utf8");
+        reject_record(&record(&[0xc3, 0x28]), "invalid_node_utf8");
+        reject_record(&record(b"invalid URI"), "invalid_rpc_uri");
+    }
+    #[test]
+    fn trailing_bytes_reject() {
+        for tail in [&b"x"[..], &b"xyz"[..]] {
+            let mut bytes = record(b"http://node:80");
+            bytes.extend_from_slice(tail);
+            reject_record(&bytes, "trailing_node_bytes");
+        }
+    }
+    #[test]
+    fn errors_and_sources_redact_input() {
+        let sentinel = "http://SECRET_USER:SECRET_PASSWORD@node:1";
+        let error = CaptureNode::new(sentinel.to_owned()).expect_err("credentials reject");
+        let mut current: &dyn Error = &error;
+        loop {
+            let rendered = format!("{current} {current:?}");
+            assert!(!rendered.contains("SECRET"));
+            match current.source() {
+                Some(source) => current = source,
+                None => break,
+            }
+        }
+        let json = format!(r#"{{"rpc_uri":"{sentinel}"}}"#);
+        let error = serde_json::from_str::<CaptureNode>(&json).expect_err("serde credentials reject");
+        assert!(!format!("{error} {error:?}").contains("SECRET"));
+        let json = format!(r#"{{"{sentinel}":0}}"#);
+        let error = serde_json::from_str::<CaptureNode>(&json).expect_err("unknown fields reject");
+        assert!(!format!("{error} {error:?}").contains("SECRET"));
+        let error = CaptureNode::new("http://[bad]:1".to_owned()).expect_err("parser rejection");
+        assert!(error.source().is_some(), "retain URL parser source");
+        let error = CaptureNode::decode(&record(&[0xff])).expect_err("UTF-8 rejection");
+        assert!(error.source().is_some(), "retain UTF-8 source");
+    }
+}

crates/capture-authority/src/truncate.rs

@@ -0,0 +1,43 @@
+use crate::{LOGS, META, Store, StoreError, decode_id, decode_log, encode_id};
+use redb::ReadableTable;
+
+impl Store {
+    /// Atomically remove every log entry at or after `from`.
+    pub fn truncate_suffix(&mut self, from: u64) -> Result<(), StoreError> {
+        let tx = self.write_transaction()?;
+        {
+            let mut meta = tx.open_table(META)?;
+            let logs = tx.open_table(LOGS)?;
+            if crate::committed::validate_committed(&meta, &logs)?.is_some_and(|id| from <= id.index) {
+                return Err(StoreError::CommittedLogProtected);
+            }
+            drop(logs);
+            let tail = meta.get("tail")?.map(|v| decode_id(v.value())).transpose()?;
+            if tail.is_none_or(|id| from > id.index) {
+                return Ok(());
+            }
+            let mut logs = tx.open_table(LOGS)?;
+            let new_tail = if from == 0 {
+                None
+            } else {
+                let index = from - 1;
+                let record = logs.get(index)?.ok_or(StoreError::InvalidLogSequence)?;
+                let id = decode_log(record.value())?.id;
+                if id.index != index {
+                    return Err(StoreError::CorruptRecord);
+                }
+                Some(id)
+            };
+            logs.retain(|index, _| index < from)?;
+            match new_tail {
+                Some(id) => {
+                    meta.insert("tail", encode_id(id).as_slice())?;
+                }
+                None => {
+                    meta.remove("tail")?;
+                }
+            }
+        }
+        self.commit(tx)
+    }
+}

crates/common/Cargo.toml

@@ -29,13 +29,11 @@ workspace = true
 
 [features]
 default = []
-test-util = []
 hotpath = ["hotpath/hotpath", "hotpath/tokio"]
 hotpath-alloc = ["hotpath", "hotpath/hotpath-alloc"]
 hotpath-cpu = ["hotpath", "hotpath/hotpath-cpu"]
 
 [dependencies]
-arc-swap.workspace = true
 hotpath.workspace = true
 tokio = { workspace = true, features = ["fs", "rt-multi-thread"] }
 tonic = { workspace = true, features = ["gzip", "deflate"] }

crates/common/src/globals.rs

@@ -12,10 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use arc_swap::ArcSwap;
 use std::collections::HashMap;
+use std::sync::LazyLock;
 use std::sync::atomic::{AtomicU64, Ordering};
-use std::sync::{Arc, LazyLock};
 use tokio::sync::RwLock;
 use tonic::transport::Channel;
 
@@ -28,42 +27,6 @@ pub static GLOBAL_ROOT_CERT: LazyLock<RwLock<Option<Vec<u8>>>> = LazyLock::new(|
 pub static GLOBAL_MTLS_IDENTITY: LazyLock<RwLock<Option<MtlsIdentityPem>>> = LazyLock::new(|| RwLock::new(None));
 pub static GLOBAL_OUTBOUND_TLS_GENERATION: LazyLock<AtomicU64> = LazyLock::new(|| AtomicU64::new(0));
 
-#[derive(Clone, Default)]
-pub struct OutboundTlsSnapshot {
-    pub generation: u64,
-    pub root_ca_pem: Option<Vec<u8>>,
-    pub mtls_identity: Option<MtlsIdentityPem>,
-}
-
-static GLOBAL_OUTBOUND_TLS_STATE: LazyLock<ArcSwap<OutboundTlsSnapshot>> =
-    LazyLock::new(|| ArcSwap::from_pointee(OutboundTlsSnapshot::default()));
-
-pub fn get_global_outbound_tls_snapshot() -> Arc<OutboundTlsSnapshot> {
-    GLOBAL_OUTBOUND_TLS_STATE.load_full()
-}
-
-/// Publish identity, trust and generation together; the individual globals are legacy views.
-pub async fn publish_global_outbound_tls_snapshot(snapshot: OutboundTlsSnapshot) -> u64 {
-    GLOBAL_OUTBOUND_TLS_STATE.rcu(|current| {
-        if current.generation != 0
-            && current.root_ca_pem == snapshot.root_ca_pem
-            && current.mtls_identity == snapshot.mtls_identity
-        {
-            return Arc::clone(current);
-        }
-        Arc::new(OutboundTlsSnapshot {
-            generation: snapshot.generation.max(current.generation.saturating_add(1)),
-            root_ca_pem: snapshot.root_ca_pem.clone(),
-            mtls_identity: snapshot.mtls_identity.clone(),
-        })
-    });
-    let accepted = get_global_outbound_tls_snapshot();
-    *GLOBAL_ROOT_CERT.write().await = accepted.root_ca_pem.clone();
-    *GLOBAL_MTLS_IDENTITY.write().await = accepted.mtls_identity.clone();
-    GLOBAL_OUTBOUND_TLS_GENERATION.store(accepted.generation, Ordering::Relaxed);
-    accepted.generation
-}
-
 /// Log level to use when reporting cached gRPC connection eviction.
 #[derive(Clone, Copy, Debug, Eq, PartialEq)]
 pub enum ConnectionEvictionLogLevel {
@@ -97,45 +60,6 @@ pub fn try_get_global_local_node_name() -> Option<String> {
         .filter(|name| !name.is_empty())
 }
 
-/// Run a synchronous test callback while publishing a name under its writer lock.
-/// Call from a blocking worker; returning or unwinding releases the writer.
-#[cfg(any(test, feature = "test-util"))]
-#[doc(hidden)]
-pub fn with_local_node_name_write_lock_for_test<T>(name: &str, work: impl FnOnce() -> T) -> T {
-    let mut writer = GLOBAL_LOCAL_NODE_NAME.blocking_write();
-    *writer = name.to_string();
-    let result = work();
-    drop(writer);
-    result
-}
-
-#[cfg(test)]
-mod node_identity_tests {
-    use super::*;
-
-    #[tokio::test]
-    async fn local_node_name_nonblocking_read_recovers_after_writer_release() {
-        let previous = get_global_local_node_name().await;
-        let (unavailable, published, panicked) = tokio::task::spawn_blocking(|| {
-            let unavailable = with_local_node_name_write_lock_for_test("published-node", try_get_global_local_node_name);
-            let published = try_get_global_local_node_name();
-            let panicked = std::panic::catch_unwind(|| {
-                with_local_node_name_write_lock_for_test("unwound-node", || panic!("identity callback failed"));
-            });
-            (unavailable, published, panicked)
-        })
-        .await
-        .expect("identity writer worker must complete");
-        let after_unwind = try_get_global_local_node_name();
-        set_global_local_node_name(&previous).await;
-
-        assert!(unavailable.is_none(), "a held writer must make the nonblocking getter unavailable");
-        assert_eq!(published.as_deref(), Some("published-node"));
-        assert!(panicked.is_err(), "the callback panic must propagate after releasing the lock");
-        assert_eq!(after_unwind.as_deref(), Some("unwound-node"));
-    }
-}
-
 /// Set the global RustFS address used for gRPC connections.
 ///
 /// # Arguments
@@ -161,27 +85,17 @@ pub async fn get_global_addr() -> String {
 /// # Arguments
 /// * `cert` - A vector of bytes representing the PEM-encoded root CA certificate.
 pub async fn set_global_root_cert(cert: Vec<u8>) {
-    GLOBAL_OUTBOUND_TLS_STATE.rcu(|current| {
-        let mut next = (**current).clone();
-        next.root_ca_pem = Some(cert.clone());
-        Arc::new(next)
-    });
     *GLOBAL_ROOT_CERT.write().await = Some(cert);
 }
 
 /// Clear the global root CA certificate for outbound gRPC clients.
 pub async fn clear_global_root_cert() {
-    GLOBAL_OUTBOUND_TLS_STATE.rcu(|current| {
-        let mut next = (**current).clone();
-        next.root_ca_pem = None;
-        Arc::new(next)
-    });
     *GLOBAL_ROOT_CERT.write().await = None;
 }
 
 /// Get the global root CA certificate for outbound gRPC clients.
 pub async fn get_global_root_cert() -> Option<Vec<u8>> {
-    get_global_outbound_tls_snapshot().root_ca_pem.clone()
+    GLOBAL_ROOT_CERT.read().await.clone()
 }
 
 /// Set the global mTLS identity (cert+key PEM) for outbound gRPC clients.
@@ -191,32 +105,22 @@ pub async fn get_global_root_cert() -> Option<Vec<u8>> {
 /// # Arguments
 /// * `identity` - An optional MtlsIdentityPem struct containing the cert and key PEM.
 pub async fn set_global_mtls_identity(identity: Option<MtlsIdentityPem>) {
-    GLOBAL_OUTBOUND_TLS_STATE.rcu(|current| {
-        let mut next = (**current).clone();
-        next.mtls_identity = identity.clone();
-        Arc::new(next)
-    });
     *GLOBAL_MTLS_IDENTITY.write().await = identity;
 }
 
 /// Get the global mTLS identity for outbound gRPC clients.
 pub async fn get_global_mtls_identity() -> Option<MtlsIdentityPem> {
-    get_global_outbound_tls_snapshot().mtls_identity.clone()
+    GLOBAL_MTLS_IDENTITY.read().await.clone()
 }
 
 /// Set the global outbound TLS generation.
 pub fn set_global_outbound_tls_generation(generation: u64) {
-    GLOBAL_OUTBOUND_TLS_STATE.rcu(|current| {
-        let mut next = (**current).clone();
-        next.generation = generation;
-        Arc::new(next)
-    });
     GLOBAL_OUTBOUND_TLS_GENERATION.store(generation, Ordering::Relaxed);
 }
 
 /// Get the global outbound TLS generation.
 pub fn get_global_outbound_tls_generation() -> u64 {
-    get_global_outbound_tls_snapshot().generation
+    GLOBAL_OUTBOUND_TLS_GENERATION.load(Ordering::Relaxed)
 }
 
 /// Evict a stale/dead connection from the global connection cache.
@@ -290,77 +194,8 @@ pub async fn clear_all_connections() {
 ///
 /// When present, gRPC clients will present this identity to servers requesting/requiring mTLS.
 /// When absent, clients proceed with standard server-authenticated TLS.
-#[derive(Clone, PartialEq, Eq)]
+#[derive(Clone, Debug)]
 pub struct MtlsIdentityPem {
     pub cert_pem: Vec<u8>,
     pub key_pem: Vec<u8>,
 }
-
-impl std::fmt::Debug for MtlsIdentityPem {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("MtlsIdentityPem")
-            .field("cert_bytes", &self.cert_pem.len())
-            .field("key_pem", &"[REDACTED]")
-            .finish()
-    }
-}
-
-#[cfg(test)]
-mod tls_snapshot_tests {
-    use super::*;
-
-    #[tokio::test]
-    async fn outbound_snapshot_publication_is_coherent_and_skips_unchanged_material() {
-        let start = get_global_outbound_tls_generation().saturating_add(1);
-        let snapshot = OutboundTlsSnapshot {
-            generation: start,
-            root_ca_pem: Some(vec![1]),
-            mtls_identity: Some(MtlsIdentityPem {
-                cert_pem: vec![1],
-                key_pem: vec![1],
-            }),
-        };
-        let first = publish_global_outbound_tls_snapshot(snapshot.clone()).await;
-        assert_eq!(
-            publish_global_outbound_tls_snapshot(OutboundTlsSnapshot {
-                generation: first.saturating_add(1),
-                ..snapshot
-            })
-            .await,
-            first
-        );
-        let writer = tokio::spawn(async move {
-            for tag in 2_u8..=255 {
-                publish_global_outbound_tls_snapshot(OutboundTlsSnapshot {
-                    generation: start.saturating_add(u64::from(tag)),
-                    root_ca_pem: Some(vec![tag]),
-                    mtls_identity: Some(MtlsIdentityPem {
-                        cert_pem: vec![tag],
-                        key_pem: vec![tag],
-                    }),
-                })
-                .await;
-                tokio::task::yield_now().await;
-            }
-        });
-        while !writer.is_finished() {
-            let state = get_global_outbound_tls_snapshot();
-            let identity = state.mtls_identity.as_ref().expect("identity in complete snapshot");
-            assert_eq!(state.root_ca_pem.as_ref().expect("roots in complete snapshot"), &identity.cert_pem);
-            assert_eq!(identity.cert_pem, identity.key_pem);
-            tokio::task::yield_now().await;
-        }
-        writer.await.expect("writer completes");
-    }
-
-    #[test]
-    fn mtls_identity_debug_redacts_private_key() {
-        let identity = MtlsIdentityPem {
-            cert_pem: b"public certificate".to_vec(),
-            key_pem: b"PRIVATE-KEY-MARKER".to_vec(),
-        };
-        let debug = format!("{identity:?}");
-        assert!(!debug.contains("PRIVATE-KEY-MARKER"));
-        assert!(debug.contains("[REDACTED]"));
-    }
-}

crates/common/src/mrf_channel.rs

@@ -172,41 +172,11 @@ pub struct MrfUnverifiedLegacyEvent {
     pub anchor: MrfDurableRepairAnchor,
 }
 
-#[derive(Clone, Copy)]
-struct ResolvedMrfRepairIdentity {
-    version_id: Option<[u8; 16]>,
-    scope: Option<MrfScope>,
-    delete_marker_purge: Option<MrfDeleteMarkerPurgeIdentity>,
-    lease: MrfIngressLease,
-    bucket_incarnation_id: Uuid,
-}
-
 impl MrfDurableRepairAnchor {
     /// Build a dischargeable anchor only when the caller supplies the storage
     /// incarnation and the original ingress lease. Legacy replay records lack
     /// both pieces and therefore remain fail-closed.
     pub fn from_intent(intent: &MrfIntent, bucket_incarnation_id: Uuid) -> Option<Self> {
-        let identity = Self::resolve_identity(intent, bucket_incarnation_id)?;
-        Some(Self {
-            kind: intent.kind,
-            bucket: intent.bucket.clone(),
-            object: intent.object.clone(),
-            version_id: identity.version_id,
-            scope: identity.scope,
-            delete_marker_purge: identity.delete_marker_purge,
-            lease: identity.lease,
-            bucket_incarnation_id: identity.bucket_incarnation_id,
-        })
-    }
-
-    /// Check whether the intent can form an anchor for this exact source,
-    /// without retaining its bucket/object handles. This checks identity shape,
-    /// not storage verification or retirement authority.
-    pub fn is_eligible_for_source(intent: &MrfIntent, source: Uuid) -> bool {
-        Self::resolve_identity(intent, source).is_some_and(|identity| identity.bucket_incarnation_id == source)
-    }
-
-    fn resolve_identity(intent: &MrfIntent, bucket_incarnation_id: Uuid) -> Option<ResolvedMrfRepairIdentity> {
         let lease = intent.lease?;
         let (version_id, scope) = canonical_identity(intent.kind, intent.version_id, intent.scope);
         let delete_marker_purge = canonical_delete_marker_purge_identity(intent)?;
@@ -216,7 +186,10 @@ impl MrfDurableRepairAnchor {
         if bucket_incarnation_id.is_nil() {
             return None;
         }
-        Some(ResolvedMrfRepairIdentity {
+        Some(Self {
+            kind: intent.kind,
+            bucket: intent.bucket.clone(),
+            object: intent.object.clone(),
             version_id,
             scope,
             delete_marker_purge,
@@ -260,32 +233,24 @@ pub fn consume_recorded_verified_mrf_repair_events_for(bucket: &str, anchors: &m
     let Ok(mut events) = registry.lock() else {
         return 0;
     };
-    consume_verified_mrf_repair_events_from_queue(bucket, anchors, &mut events)
-}
-
-fn consume_verified_mrf_repair_events_from_queue(
-    bucket: &str,
-    anchors: &mut Vec<MrfDurableRepairAnchor>,
-    events: &mut std::collections::VecDeque<MrfVerifiedRepairEvent>,
-) -> usize {
     let before = anchors.len();
-    let previous_event_count = events.len();
-    events.retain(|event| {
+    let mut retained = std::collections::VecDeque::with_capacity(events.len());
+    while let Some(event) = events.pop_front() {
         if event.bucket.as_ref() != bucket {
-            return true;
+            retained.push_back(event);
+            continue;
         }
         let mut matched = false;
         anchors.retain(|anchor| {
-            let proven = anchor.is_proven_by(event);
+            let proven = anchor.is_proven_by(&event);
             matched |= proven;
             !proven
         });
-        !matched
-    });
-    // The old queue rebuild reserved the pre-consumption event count. Reuse
-    // that buffer when possible without extending its capacity lifetime:
-    // an emptied queue releases its backing storage on the next consumption.
-    events.shrink_to(previous_event_count);
+        if !matched {
+            retained.push_back(event);
+        }
+    }
+    *events = retained;
     before.saturating_sub(anchors.len())
 }
 
@@ -856,21 +821,6 @@ static MRF_VERIFIED_REPAIR_EVENTS: OnceLock<std::sync::Mutex<std::collections::V
 static MRF_UNVERIFIED_LEGACY_EVENTS: OnceLock<std::sync::Mutex<std::collections::VecDeque<MrfUnverifiedLegacyEvent>>> =
     OnceLock::new();
 
-/// Whether durable repair events may need consumption. This is only a work
-/// hint: either queue can receive an event after its empty observation, and a
-/// bounded queue can evict that event before the next consumer invocation.
-/// An unavailable lock is unknown, so callers must keep their normal path.
-/// An allocated empty verified queue also needs the normal consumer pass to
-/// release its buffer; this hint must not extend that buffer's lifetime.
-pub fn mrf_durable_repair_events_may_be_pending() -> bool {
-    MRF_VERIFIED_REPAIR_EVENTS
-        .get()
-        .is_some_and(|registry| registry.try_lock().map_or(true, |events| events.capacity() != 0))
-        || MRF_UNVERIFIED_LEGACY_EVENTS
-            .get()
-            .is_some_and(|registry| registry.try_lock().map_or(true, |events| !events.is_empty()))
-}
-
 /// Record a legacy notification for compatibility. This is not an
 /// acknowledgement of storage verification or durable repair completion.
 pub fn note_mrf_repaired(bucket: &str, object: &str, version_id: Option<[u8; 16]>) {
@@ -985,15 +935,6 @@ pub fn take_mrf_unverified_legacy_events_for(bucket: &str) -> Vec<MrfUnverifiedL
     taken
 }
 
-#[cfg(test)]
-mod event_presence_tests;
-
-#[cfg(test)]
-mod proof_queue_tests;
-
-#[cfg(test)]
-mod repair_identity_tests;
-
 #[cfg(test)]
 mod tests {
     use super::*;

crates/common/src/mrf_channel/event_presence_tests.rs

@@ -1,140 +0,0 @@
-// Copyright 2026 RustFS Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use super::*;
-
-fn isolated(name: &str) -> bool {
-    const CHILD: &str = "RUSTFS_MRF_EVENT_PRESENCE_TEST";
-    let test = format!("mrf_channel::event_presence_tests::{name}");
-    if let Ok(current) = std::env::var(CHILD) {
-        assert_eq!(current, test, "only the selected registry test may run in the child");
-        return false;
-    }
-    let output = std::process::Command::new(std::env::current_exe().expect("test executable"))
-        .args([&test, "--exact", "--test-threads=1"])
-        .env(CHILD, &test)
-        .output()
-        .expect("run isolated global registry test");
-    let stdout = String::from_utf8_lossy(&output.stdout);
-    assert!(
-        output.status.success() && stdout.contains("1 passed; 0 failed"),
-        "registry child must execute exactly once: {stdout}{}",
-        String::from_utf8_lossy(&output.stderr)
-    );
-    true
-}
-
-#[test]
-fn event_presence_covers_both_queues_and_contention_without_consuming() {
-    if isolated("event_presence_covers_both_queues_and_contention_without_consuming") {
-        return;
-    }
-    assert!(!mrf_durable_repair_events_may_be_pending());
-    let anchor = MrfDurableRepairAnchor {
-        kind: MrfKind::PartialWrite,
-        bucket: Arc::from("presence-test"),
-        object: Arc::from("object"),
-        version_id: Some([7; 16]),
-        scope: Some(MrfScope {
-            pool_index: 0,
-            set_index: 0,
-        }),
-        delete_marker_purge: None,
-        lease: MrfIngressLease::new(21),
-        bucket_incarnation_id: Uuid::new_v4(),
-    };
-    let event = MrfVerifiedRepairEvent {
-        kind: anchor.kind,
-        bucket: anchor.bucket.clone(),
-        object: anchor.object.clone(),
-        version_id: anchor.version_id,
-        scope: anchor.scope,
-        delete_marker_purge: None,
-        lease: Some(anchor.lease),
-        bucket_incarnation_id: anchor.bucket_incarnation_id,
-        disposition: MrfVerifiedRepairDisposition::VerifiedHealthy,
-    };
-    note_mrf_verified_repair(event.clone());
-    assert!(mrf_durable_repair_events_may_be_pending());
-    assert!(mrf_durable_repair_events_may_be_pending(), "the hint must not consume a proof");
-    assert_eq!(take_mrf_verified_repair_events_for(&anchor.bucket), vec![event]);
-    assert!(
-        mrf_durable_repair_events_may_be_pending(),
-        "the empty allocated verified buffer still needs its original maintenance pass"
-    );
-    let mut retained = vec![anchor.clone()];
-    assert_eq!(consume_recorded_verified_mrf_repair_events_for(&anchor.bucket, &mut retained), 0);
-    assert_eq!(retained, vec![anchor.clone()], "buffer maintenance cannot discharge an anchor");
-    assert_eq!(
-        MRF_VERIFIED_REPAIR_EVENTS
-            .get()
-            .expect("verified registry")
-            .lock()
-            .expect("verified storage")
-            .capacity(),
-        0,
-        "the actual consumer releases empty proof storage before subsequent idle passes can skip it"
-    );
-    assert!(!mrf_durable_repair_events_may_be_pending());
-    let legacy = MrfUnverifiedLegacyEvent { anchor };
-    note_mrf_unverified_legacy(legacy.clone());
-    assert!(mrf_durable_repair_events_may_be_pending(), "legacy-only work must not be skipped");
-    assert_eq!(take_mrf_unverified_legacy_events_for(&legacy.anchor.bucket), vec![legacy]);
-    assert!(!mrf_durable_repair_events_may_be_pending());
-    let verified = MRF_VERIFIED_REPAIR_EVENTS.get().expect("verified registry initialized");
-    let guard = verified.lock().expect("hold verified registry");
-    assert!(mrf_durable_repair_events_may_be_pending(), "contended verified is unknown");
-    drop(guard);
-    let legacy = MRF_UNVERIFIED_LEGACY_EVENTS.get().expect("legacy registry initialized");
-    let guard = legacy.lock().expect("hold legacy registry");
-    assert!(mrf_durable_repair_events_may_be_pending(), "contended legacy is unknown");
-    drop(guard);
-    assert!(!mrf_durable_repair_events_may_be_pending());
-}
-
-#[test]
-fn event_presence_keeps_poisoned_verified_unknown() {
-    if isolated("event_presence_keeps_poisoned_verified_unknown") {
-        return;
-    }
-    assert!(!mrf_durable_repair_events_may_be_pending());
-    let registry = MRF_VERIFIED_REPAIR_EVENTS.get_or_init(|| Mutex::new(Default::default()));
-    assert!(
-        std::panic::catch_unwind(|| {
-            let _guard = registry.lock().expect("unpoisoned registry");
-            panic!("poison the isolated verified registry");
-        })
-        .is_err()
-    );
-    assert!(mrf_durable_repair_events_may_be_pending());
-    assert!(registry.is_poisoned(), "the hint must not clear or recover poisoned state");
-}
-
-#[test]
-fn event_presence_keeps_poisoned_legacy_unknown() {
-    if isolated("event_presence_keeps_poisoned_legacy_unknown") {
-        return;
-    }
-    assert!(!mrf_durable_repair_events_may_be_pending());
-    let registry = MRF_UNVERIFIED_LEGACY_EVENTS.get_or_init(|| Mutex::new(Default::default()));
-    assert!(
-        std::panic::catch_unwind(|| {
-            let _guard = registry.lock().expect("unpoisoned registry");
-            panic!("poison the isolated legacy registry");
-        })
-        .is_err()
-    );
-    assert!(mrf_durable_repair_events_may_be_pending());
-    assert!(registry.is_poisoned(), "the hint must not clear or recover poisoned state");
-}

crates/common/src/mrf_channel/proof_queue_tests.rs

@@ -1,304 +0,0 @@
-// Copyright 2026 RustFS Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use super::*;
-use std::collections::VecDeque;
-
-fn proof_fixture() -> (MrfDurableRepairAnchor, MrfVerifiedRepairEvent) {
-    let anchor = MrfDurableRepairAnchor {
-        kind: MrfKind::PartialWrite,
-        bucket: Arc::from("proof-queue"),
-        object: Arc::from("matched"),
-        version_id: Some([7; 16]),
-        scope: Some(MrfScope {
-            pool_index: 1,
-            set_index: 2,
-        }),
-        delete_marker_purge: None,
-        lease: MrfIngressLease::new(21),
-        bucket_incarnation_id: Uuid::new_v4(),
-    };
-    let event = MrfVerifiedRepairEvent {
-        kind: anchor.kind,
-        bucket: anchor.bucket.clone(),
-        object: anchor.object.clone(),
-        version_id: anchor.version_id,
-        scope: anchor.scope,
-        delete_marker_purge: anchor.delete_marker_purge,
-        lease: Some(anchor.lease),
-        bucket_incarnation_id: anchor.bucket_incarnation_id,
-        disposition: MrfVerifiedRepairDisposition::Repaired,
-    };
-    (anchor, event)
-}
-
-#[test]
-fn proof_queue_reuses_unmatched_storage_in_contiguous_and_wrapped_layouts() {
-    for wrapped in [false, true] {
-        for foreign in [false, true] {
-            let (anchor, mut event) = proof_fixture();
-            if foreign {
-                event.bucket = Arc::from("another-bucket");
-            } else {
-                event.lease = None;
-            }
-            let mut events = VecDeque::with_capacity(8);
-            events.extend(std::iter::repeat_n(event, events.capacity()));
-            if wrapped {
-                let first = events.pop_front().expect("full local proof queue");
-                events.push_back(first);
-                assert!(!events.as_slices().1.is_empty(), "exercise a wrapped producer queue");
-            }
-            let capacity = events.capacity();
-            let storage = events
-                .iter()
-                .map(std::ptr::from_ref)
-                .min()
-                .expect("nonempty local proof storage");
-            let expected = events.clone();
-            let mut anchors = vec![anchor.clone()];
-            for _ in 0..4 {
-                assert_eq!(
-                    consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-                    0
-                );
-                assert_eq!(events, expected);
-                assert_eq!(anchors, vec![anchor.clone()]);
-                assert_eq!(events.capacity(), capacity);
-                assert_eq!(
-                    events.iter().map(std::ptr::from_ref).min(),
-                    Some(storage),
-                    "a full unmatched queue must reuse its buffer instead of rebuilding it for every bucket"
-                );
-            }
-        }
-    }
-}
-
-#[test]
-fn proof_queue_consumes_all_exact_anchor_copies_and_preserves_duplicate_proof_order() {
-    let (anchor, event) = proof_fixture();
-    let waiting = MrfDurableRepairAnchor {
-        object: Arc::from("waiting"),
-        ..anchor.clone()
-    };
-    let foreign = MrfVerifiedRepairEvent {
-        bucket: Arc::from("foreign"),
-        ..event.clone()
-    };
-    let unleased = MrfVerifiedRepairEvent {
-        lease: None,
-        ..event.clone()
-    };
-    let future = MrfVerifiedRepairEvent {
-        object: Arc::from("future"),
-        ..event.clone()
-    };
-    let mut events = VecDeque::from(vec![
-        foreign.clone(),
-        event.clone(),
-        unleased.clone(),
-        event.clone(),
-        future.clone(),
-    ]);
-    let first = events.pop_front().expect("foreign proof starts the queue");
-    events.push_back(first);
-    assert!(!events.as_slices().1.is_empty(), "match and retain proofs across the wrap boundary");
-    let mut anchors = vec![anchor.clone(), waiting.clone(), anchor.clone()];
-    assert_eq!(
-        consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-        2
-    );
-    assert_eq!(anchors, vec![waiting]);
-    assert_eq!(
-        events,
-        VecDeque::from(vec![unleased, event, future, foreign]),
-        "the later duplicate proof remains available after the first proof consumes every exact anchor"
-    );
-    let retained = events.clone();
-    assert_eq!(
-        consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-        0
-    );
-    assert_eq!(events, retained, "a later consumer must still be able to match the retained proofs");
-}
-
-#[test]
-fn proof_queue_keeps_every_incomplete_or_mismatched_identity() {
-    let (anchor, event) = proof_fixture();
-    let different_purge = MrfDeleteMarkerPurgeIdentity {
-        bucket_incarnation_id: anchor.bucket_incarnation_id,
-        marker_incarnation_id: anchor.bucket_incarnation_id,
-        marker_identity: [3; 32],
-    };
-    for rejected in [
-        MrfVerifiedRepairEvent {
-            kind: MrfKind::DecodeFailure,
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            bucket: Arc::from("foreign"),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            object: Arc::from("other"),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            version_id: None,
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            version_id: Some([8; 16]),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            scope: None,
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            scope: Some(MrfScope {
-                pool_index: 2,
-                set_index: 2,
-            }),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            scope: Some(MrfScope {
-                pool_index: 1,
-                set_index: 3,
-            }),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            lease: None,
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            lease: Some(MrfIngressLease::new(22)),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            bucket_incarnation_id: Uuid::new_v4(),
-            ..event.clone()
-        },
-        MrfVerifiedRepairEvent {
-            delete_marker_purge: Some(different_purge),
-            ..event
-        },
-    ] {
-        let mut anchors = vec![anchor.clone()];
-        let mut events = VecDeque::from([rejected.clone()]);
-        assert_eq!(
-            consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-            0
-        );
-        assert_eq!(anchors, vec![anchor.clone()]);
-        assert_eq!(events, VecDeque::from([rejected]));
-    }
-}
-
-#[test]
-fn proof_queue_requires_the_complete_delete_marker_purge_identity() {
-    let (mut anchor, mut event) = proof_fixture();
-    let purge = MrfDeleteMarkerPurgeIdentity {
-        bucket_incarnation_id: anchor.bucket_incarnation_id,
-        marker_incarnation_id: anchor.bucket_incarnation_id,
-        marker_identity: [3; 32],
-    };
-    anchor.kind = MrfKind::DeleteMarkerPurge;
-    anchor.delete_marker_purge = Some(purge);
-    event.kind = anchor.kind;
-    event.delete_marker_purge = anchor.delete_marker_purge;
-    for rejected in [
-        None,
-        Some(MrfDeleteMarkerPurgeIdentity {
-            bucket_incarnation_id: Uuid::new_v4(),
-            ..purge
-        }),
-        Some(MrfDeleteMarkerPurgeIdentity {
-            marker_incarnation_id: Uuid::new_v4(),
-            ..purge
-        }),
-        Some(MrfDeleteMarkerPurgeIdentity {
-            marker_identity: [4; 32],
-            ..purge
-        }),
-    ] {
-        let rejected = MrfVerifiedRepairEvent {
-            delete_marker_purge: rejected,
-            ..event.clone()
-        };
-        let mut anchors = vec![anchor.clone()];
-        let mut events = VecDeque::from([rejected.clone(), event.clone()]);
-        assert_eq!(
-            consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-            1
-        );
-        assert!(anchors.is_empty());
-        assert_eq!(events, VecDeque::from([rejected]));
-    }
-}
-
-#[test]
-fn proof_queue_capacity_follows_the_preconsumption_length_and_releases_on_the_next_empty_call() {
-    let (anchor, event) = proof_fixture();
-    let future_events = (0..3)
-        .map(|index| MrfVerifiedRepairEvent {
-            object: Arc::from(format!("future-{index}")),
-            ..event.clone()
-        })
-        .collect::<Vec<_>>();
-    let mut events = VecDeque::with_capacity(64);
-    events.push_back(event);
-    events.extend(future_events.iter().cloned());
-    let previous_event_count = events.len();
-    let mut anchors = vec![anchor.clone()];
-    assert_eq!(
-        consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-        1
-    );
-    assert_eq!(events.iter().cloned().collect::<Vec<_>>(), future_events);
-    assert_eq!(
-        events.capacity(),
-        VecDeque::<MrfVerifiedRepairEvent>::with_capacity(previous_event_count).capacity(),
-        "preserve the former queue rebuild's capacity lifetime instead of retaining the producer high water"
-    );
-    let previous_event_count = events.len();
-    let mut anchors = future_events
-        .iter()
-        .map(|event| MrfDurableRepairAnchor {
-            object: event.object.clone(),
-            ..anchor.clone()
-        })
-        .collect();
-    assert_eq!(
-        consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-        3
-    );
-    assert!(events.is_empty());
-    assert_eq!(
-        events.capacity(),
-        VecDeque::<MrfVerifiedRepairEvent>::with_capacity(previous_event_count).capacity()
-    );
-    assert_eq!(
-        consume_verified_mrf_repair_events_from_queue(&anchor.bucket, &mut anchors, &mut events),
-        0
-    );
-    assert_eq!(
-        events.capacity(),
-        0,
-        "the next empty consumption releases backing storage just as a rebuild did"
-    );
-}

crates/common/src/mrf_channel/repair_identity_tests.rs

@@ -1,256 +0,0 @@
-// Copyright 2026 RustFS Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use super::*;
-
-// Compatibility reference from a47649fb60c0492cc0fb02bff28439c5e357f8e4.
-// Keep this independent of the shared resolver and its canonical helpers.
-fn original_anchor(intent: &MrfIntent, bucket_incarnation_id: Uuid) -> Option<MrfDurableRepairAnchor> {
-    let lease = intent.lease?;
-    let version_id = intent.version_id.filter(|bytes| *bytes != [0; 16]);
-    let (version_id, scope) = match intent.kind {
-        MrfKind::MetadataCorruption => (None, None),
-        MrfKind::DecodeFailure | MrfKind::PartialWrite | MrfKind::DeleteMarkerPurge => (version_id, intent.scope),
-    };
-    let delete_marker_purge = match intent.kind {
-        MrfKind::DeleteMarkerPurge => intent
-            .delete_marker_purge
-            .as_ref()
-            .filter(|payload| {
-                !payload.bucket_incarnation_id.is_nil()
-                    && payload.marker_incarnation_id == payload.bucket_incarnation_id
-                    && !payload.marker.is_empty()
-                    && payload.marker.len() <= MRF_MAX_DELETE_MARKER_BYTES
-            })
-            .map(|payload| {
-                Some(MrfDeleteMarkerPurgeIdentity {
-                    bucket_incarnation_id: payload.bucket_incarnation_id,
-                    marker_incarnation_id: payload.marker_incarnation_id,
-                    marker_identity: payload.marker_identity,
-                })
-            })?,
-        _ if intent.delete_marker_purge.is_none() => None,
-        _ => return None,
-    };
-    let bucket_incarnation_id = delete_marker_purge
-        .map(|identity| identity.bucket_incarnation_id)
-        .unwrap_or(bucket_incarnation_id);
-    if bucket_incarnation_id.is_nil() {
-        return None;
-    }
-    Some(MrfDurableRepairAnchor {
-        kind: intent.kind,
-        bucket: intent.bucket.clone(),
-        object: intent.object.clone(),
-        version_id,
-        scope,
-        delete_marker_purge,
-        lease,
-        bucket_incarnation_id,
-    })
-}
-
-#[test]
-fn shared_repair_identity_matches_the_original_constructor_and_source_filter() {
-    let source = Uuid::from_u128(1);
-    let foreign = Uuid::from_u128(2);
-    let purge = MrfDeleteMarkerPurge {
-        bucket_incarnation_id: source,
-        marker_incarnation_id: source,
-        marker_identity: [9; 32],
-        marker: Arc::from([7_u8]),
-    };
-    let payloads = [
-        None,
-        Some(purge.clone()),
-        Some(MrfDeleteMarkerPurge {
-            bucket_incarnation_id: foreign,
-            marker_incarnation_id: foreign,
-            ..purge.clone()
-        }),
-        Some(MrfDeleteMarkerPurge {
-            marker: Arc::from([]),
-            ..purge.clone()
-        }),
-        Some(MrfDeleteMarkerPurge {
-            marker: Arc::from(vec![7; MRF_MAX_DELETE_MARKER_BYTES]),
-            ..purge.clone()
-        }),
-        Some(MrfDeleteMarkerPurge {
-            marker: Arc::from(vec![7; MRF_MAX_DELETE_MARKER_BYTES + 1]),
-            ..purge.clone()
-        }),
-        Some(MrfDeleteMarkerPurge {
-            marker_incarnation_id: foreign,
-            ..purge.clone()
-        }),
-        Some(MrfDeleteMarkerPurge {
-            bucket_incarnation_id: Uuid::nil(),
-            marker_incarnation_id: Uuid::nil(),
-            ..purge
-        }),
-    ];
-    let mut cases = 0;
-    for kind in [
-        MrfKind::DecodeFailure,
-        MrfKind::MetadataCorruption,
-        MrfKind::PartialWrite,
-        MrfKind::DeleteMarkerPurge,
-    ] {
-        for lease in [None, Some(MrfIngressLease::new(0)), Some(MrfIngressLease::new(7))] {
-            for version_id in [None, Some([0; 16]), Some([7; 16])] {
-                for scope in [
-                    None,
-                    Some(MrfScope {
-                        pool_index: u32::MAX,
-                        set_index: 3,
-                    }),
-                ] {
-                    for input_source in [Uuid::nil(), source, foreign] {
-                        for (payload_index, payload) in payloads.iter().enumerate() {
-                            let intent = MrfIntent {
-                                bucket: Arc::from("source-shape"),
-                                object: Arc::from("object"),
-                                version_id,
-                                kind,
-                                delete_marker_purge: payload.clone(),
-                                scope,
-                                lease,
-                                enqueued_at_ms: 0,
-                                attempts: 0,
-                            };
-                            let expected = original_anchor(&intent, input_source);
-                            assert_eq!(
-                                MrfDurableRepairAnchor::from_intent(&intent, input_source),
-                                expected,
-                                "kind={kind:?}, lease={lease:?}, version={version_id:?}, scope={scope:?}, source={input_source}, payload={payload_index}"
-                            );
-                            assert_eq!(
-                                MrfDurableRepairAnchor::is_eligible_for_source(&intent, input_source),
-                                expected
-                                    .as_ref()
-                                    .is_some_and(|anchor| anchor.bucket_incarnation_id == input_source),
-                                "source eligibility must preserve the exact source: kind={kind:?}, lease={lease:?}, version={version_id:?}, scope={scope:?}, source={input_source}, payload={payload_index}"
-                            );
-                            cases += 1;
-                        }
-                    }
-                }
-            }
-        }
-    }
-    assert_eq!(cases, 1728, "exercise the complete explicit identity-shape matrix");
-}
-
-#[test]
-fn repair_identity_preserves_literal_purge_source_lease_and_metadata_boundaries() {
-    let source = Uuid::from_u128(1);
-    let purge_source = Uuid::from_u128(2);
-    let mut intent = MrfIntent {
-        bucket: Arc::from("literal-bucket"),
-        object: Arc::from("marker"),
-        version_id: None,
-        kind: MrfKind::DeleteMarkerPurge,
-        delete_marker_purge: Some(MrfDeleteMarkerPurge {
-            bucket_incarnation_id: purge_source,
-            marker_incarnation_id: purge_source,
-            marker_identity: [9; 32],
-            marker: Arc::from([7_u8]),
-        }),
-        scope: None,
-        lease: Some(MrfIngressLease::new(17)),
-        enqueued_at_ms: 0,
-        attempts: 0,
-    };
-    assert!(
-        !valid_intent_payload(&intent),
-        "constructor compatibility is broader than durable admission"
-    );
-    assert_eq!(
-        MrfDurableRepairAnchor::from_intent(&intent, Uuid::nil()),
-        Some(MrfDurableRepairAnchor {
-            kind: MrfKind::DeleteMarkerPurge,
-            bucket: Arc::from("literal-bucket"),
-            object: Arc::from("marker"),
-            version_id: None,
-            scope: None,
-            delete_marker_purge: Some(MrfDeleteMarkerPurgeIdentity {
-                bucket_incarnation_id: purge_source,
-                marker_incarnation_id: purge_source,
-                marker_identity: [9; 32],
-            }),
-            lease: MrfIngressLease::new(17),
-            bucket_incarnation_id: purge_source,
-        })
-    );
-    assert!(
-        !MrfDurableRepairAnchor::is_eligible_for_source(&intent, source),
-        "a purge cannot borrow a conflicting entry source"
-    );
-    assert!(!MrfDurableRepairAnchor::is_eligible_for_source(&intent, Uuid::nil()));
-    assert!(MrfDurableRepairAnchor::is_eligible_for_source(&intent, purge_source));
-    intent.lease = None;
-    assert!(MrfDurableRepairAnchor::from_intent(&intent, purge_source).is_none());
-    assert!(!MrfDurableRepairAnchor::is_eligible_for_source(&intent, purge_source));
-    intent.lease = Some(MrfIngressLease::new(17));
-    intent
-        .delete_marker_purge
-        .as_mut()
-        .expect("purge fixture")
-        .marker_incarnation_id = Uuid::nil();
-    assert!(MrfDurableRepairAnchor::from_intent(&intent, purge_source).is_none());
-    intent
-        .delete_marker_purge
-        .as_mut()
-        .expect("purge fixture")
-        .bucket_incarnation_id = Uuid::nil();
-    assert!(
-        !MrfDurableRepairAnchor::is_eligible_for_source(&intent, source),
-        "nil marker identity cannot borrow a non-nil source"
-    );
-    intent.kind = MrfKind::PartialWrite;
-    assert!(
-        MrfDurableRepairAnchor::from_intent(&intent, source).is_none(),
-        "ordinary intents cannot carry a purge payload"
-    );
-    intent.delete_marker_purge = None;
-    intent.kind = MrfKind::MetadataCorruption;
-    intent.version_id = Some([7; 16]);
-    intent.scope = Some(MrfScope {
-        pool_index: 8,
-        set_index: 9,
-    });
-    assert_eq!(
-        MrfDurableRepairAnchor::from_intent(&intent, source),
-        Some(MrfDurableRepairAnchor {
-            kind: MrfKind::MetadataCorruption,
-            bucket: Arc::from("literal-bucket"),
-            object: Arc::from("marker"),
-            version_id: None,
-            scope: None,
-            delete_marker_purge: None,
-            lease: MrfIngressLease::new(17),
-            bucket_incarnation_id: source,
-        })
-    );
-    assert!(MrfDurableRepairAnchor::is_eligible_for_source(&intent, source));
-    assert!(MrfDurableRepairAnchor::from_intent(&intent, Uuid::nil()).is_none());
-    intent.lease = Some(MrfIngressLease::new(0));
-    assert_eq!(
-        MrfDurableRepairAnchor::from_intent(&intent, source)
-            .expect("preserve the existing opaque lease shape")
-            .lease,
-        MrfIngressLease::new(0)
-    );
-}

crates/concurrency/src/adaptive.rs

@@ -1,957 +0,0 @@
-// Copyright 2026 RustFS Team
-// Licensed under the Apache License, Version 2.0.
-
-//! Mutable admission capacity without replacing or revoking held permits.
-
-use std::num::NonZeroU16;
-use std::sync::Arc;
-use std::sync::atomic::{AtomicU64, AtomicUsize, Ordering};
-use std::time::Duration;
-use tokio::sync::{Notify, Semaphore};
-use tokio::time::Instant;
-
-/// A resource owner publishes a capacity; running units retain their permits
-/// after a reduction. Lease slots are reserved once, with no per-unit Arc
-/// allocation. Tokio wait queues may lazily initialize platform mutex storage
-/// on first contention. It does not discover devices, start
-/// workers, or decide resource policy. Cancellation drops the acquire future.
-///
-/// Uncontended admission is atomic. Parked work takes a FIFO admission turn,
-/// so repeated fast acquisitions cannot starve an older waiter. Source/age
-/// priority still belongs to the caller's scheduler, before joining this queue.
-pub struct AdaptiveAdmission {
-    // One CAS observes both capacity and active count, so a concurrent shrink
-    // cannot admit a new unit using an already superseded capacity.
-    state: AtomicU64,
-    origin: Instant,
-    valid_until_millis: AtomicU64,
-    // Only the FIFO turn holder subscribes. A stored notification is harmless:
-    // capacity and freshness are always checked again after waking.
-    changed: Notify,
-    queued: AtomicUsize,
-    // FIFO admission turn, not a mutable-state lock. No storage lock or
-    // admitted permit may be held while joining. This private semaphore is
-    // never closed; release and resize do not acquire it.
-    turn: Semaphore,
-    maximum: u32,
-    slots: Box<[PermitSlots]>,
-    next_chunk: AtomicUsize,
-}
-
-struct PermitSlots {
-    free: AtomicU64,
-    references: [AtomicUsize; 64],
-}
-
-struct QueuedAdmission<'a>(&'a AdaptiveAdmission);
-
-impl Drop for QueuedAdmission<'_> {
-    fn drop(&mut self) {
-        self.0.queued.fetch_sub(1, Ordering::AcqRel);
-    }
-}
-
-/// One active unit. Clones share that unit without allocating; the final drop
-/// returns it to the same owner, including after resize or waiter cancellation.
-#[must_use = "hold the permit for the admitted unit's complete lifetime"]
-pub struct AdaptivePermit {
-    owner: Arc<AdaptiveAdmission>,
-    slot: usize,
-}
-
-impl Clone for AdaptivePermit {
-    fn clone(&self) -> Self {
-        // Every slot reference owns a strong owner reference. Clone the Arc
-        // first, so its overflow protection also bounds the slot refcount.
-        let owner = Arc::clone(&self.owner);
-        owner.slots[self.slot / 64].references[self.slot % 64].fetch_add(1, Ordering::Relaxed);
-        Self { owner, slot: self.slot }
-    }
-}
-
-impl std::fmt::Debug for AdaptivePermit {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("AdaptivePermit")
-            .field("slot", &self.slot)
-            .field("owner", &self.owner.snapshot())
-            .finish()
-    }
-}
-
-/// One coherent view of configured capacity and actual active units.
-#[derive(Debug, Clone, Copy, PartialEq, Eq)]
-pub struct AdaptiveAdmissionSnapshot {
-    /// Current cap on new admissions. Zero pauses all new units.
-    pub limit: u32,
-    /// Existing units, which may temporarily exceed a reduced limit.
-    pub active: u32,
-}
-
-/// Node scheduling capacity, separate from physical I/O and buffer ownership.
-/// Its non-cloneable units end with the async operation; a native tail must
-/// retain its own physical-resource unit, never this scheduling unit.
-pub struct NodeAdmission {
-    capacity: Arc<AdaptiveAdmission>,
-    completed: AtomicU64,
-}
-
-/// One async operation's scheduling unit. Dropping a cancelled or failed
-/// operation releases capacity without claiming completed work.
-#[must_use = "retain the unit until the async operation ends"]
-pub struct NodePermit {
-    _capacity: AdaptivePermit,
-    owner: Arc<NodeAdmission>,
-}
-
-impl NodeAdmission {
-    /// Construct one node owner with no capacity until a resource sample arrives.
-    pub fn new(maximum: NonZeroU16) -> Arc<Self> {
-        Arc::new(Self {
-            capacity: AdaptiveAdmission::new(maximum, 0),
-            completed: AtomicU64::new(0),
-        })
-    }
-
-    /// Publish feedback with this owner's actual completed-work counter.
-    pub fn update(&self, controller: &mut FeedbackController, observation: Option<FeedbackObservation>) -> FeedbackDecision {
-        let observation = observation.map(|mut observation| {
-            observation.completed_units = self.completed();
-            observation
-        });
-        controller.update(&self.capacity, observation)
-    }
-
-    /// Current leased scheduling cap and held async units.
-    pub fn snapshot(&self) -> AdaptiveAdmissionSnapshot {
-        self.capacity.snapshot()
-    }
-
-    /// Completed operations, or unknown after the counter exhausts its range.
-    pub fn completed(&self) -> Option<u64> {
-        let value = self.completed.load(Ordering::Acquire);
-        (value != u64::MAX).then_some(value)
-    }
-
-    /// Try without waiting, including when the caller already owns a lock.
-    pub fn try_acquire(self: &Arc<Self>) -> Option<NodePermit> {
-        self.capacity.try_acquire().map(|capacity| NodePermit {
-            _capacity: capacity,
-            owner: self.clone(),
-        })
-    }
-
-    /// Wait in FIFO order. Callers must provide a deadline/cancellation and
-    /// acquire before storage locks, not while retaining another node unit.
-    pub async fn acquire(self: &Arc<Self>) -> NodePermit {
-        NodePermit {
-            _capacity: self.capacity.acquire().await,
-            owner: self.clone(),
-        }
-    }
-}
-
-impl NodePermit {
-    /// Count a completed operation exactly once, then release its node unit.
-    /// This counter is scheduling progress, not an authoritative heal receipt.
-    pub fn complete(self) {
-        let _ = self
-            .owner
-            .completed
-            .try_update(Ordering::AcqRel, Ordering::Acquire, |value| value.checked_add(1));
-    }
-}
-
-/// Confidence-qualified pressure for one complete sampling window.
-#[derive(Debug, Clone, Copy, PartialEq, Eq)]
-pub enum ResourcePressure {
-    /// Missing, stale, reset, or insufficient observations; never evidence of idle.
-    Unknown,
-    /// A resource or foreground-protection threshold was reached.
-    Busy,
-    /// All required observations were fresh and below their low watermarks.
-    Stable,
-}
-
-/// Resource-owner input to the controller, after applying operator and kernel
-/// caps. This describes one owner; callers must not duplicate a device's budget
-/// for aliases, pools, or remote requesters.
-#[derive(Debug, Clone, Copy)]
-pub struct FeedbackObservation {
-    /// Identity of the resource/telemetry generation. Replacement resets probes.
-    pub generation: u64,
-    /// Upper bound imposed by resources and explicit configuration; zero blocks.
-    pub max_concurrency: u32,
-    /// Complete-window pressure, including the confidence of its sources.
-    pub pressure: ResourcePressure,
-    /// Cumulative completed work in this generation, not just scheduled work.
-    pub completed_units: Option<u64>,
-}
-
-/// Why the latest concurrency limit was selected.
-#[derive(Debug, Clone, Copy, PartialEq, Eq)]
-pub enum FeedbackReason {
-    /// The resource cap is unavailable; no new unit can be bounded safely.
-    Unavailable,
-    /// A changed owner or reset work counter requires a new baseline.
-    NewBaseline,
-    /// The resource or operator cap reduced the permitted concurrency.
-    CapReduced,
-    /// Pressure caused an immediate multiplicative reduction.
-    Pressure,
-    /// Incomplete telemetry retains at most one maintenance unit.
-    LowConfidence,
-    /// Waiting for consecutive stable windows, work progress and cooldown.
-    Recovering,
-    /// One additional unit is being probed after a stable recovery period.
-    Probe,
-}
-
-/// One controller decision, suitable for metrics without dynamic labels.
-#[derive(Debug, Clone, Copy, PartialEq, Eq)]
-pub struct FeedbackDecision {
-    /// New admissions use this cap; existing permits remain valid.
-    pub limit: u32,
-    /// Decision cause, distinct from a proof of work completion or storage health.
-    pub reason: FeedbackReason,
-}
-
-/// Single-owner additive probe/multiplicative decrease controller. It adjusts
-/// only concurrency. Every source must be fresh before declaring `Stable`.
-///
-/// The sampler runs every second. A gap over two seconds resets recovery; eight
-/// consecutive stable seconds and observed completed work allow one extra unit.
-/// Pressure or lower caps apply immediately. This policy is not a performance
-/// claim; its parameters require the same workload/foreground-SLO acceptance.
-pub struct FeedbackController {
-    previous: Option<(Instant, FeedbackObservation)>,
-    stable_since: Option<Instant>,
-    last_change: Instant,
-    limit: u32,
-}
-
-impl Default for FeedbackController {
-    fn default() -> Self {
-        Self::new()
-    }
-}
-
-impl FeedbackController {
-    /// Start without a resource observation or an admission grant.
-    pub fn new() -> Self {
-        Self {
-            previous: None,
-            stable_since: None,
-            last_change: Instant::now(),
-            limit: 0,
-        }
-    }
-
-    /// Consume one sample and publish a three-second lease to the same gate.
-    /// If this function stops being called, admission expires independently.
-    pub fn update(&mut self, gate: &AdaptiveAdmission, observation: Option<FeedbackObservation>) -> FeedbackDecision {
-        let observation = observation.map(|mut observation| {
-            observation.max_concurrency = observation.max_concurrency.min(gate.maximum);
-            observation
-        });
-        let decision = self.decide(Instant::now(), observation);
-        gate.set_sampled_limit(decision.limit, Duration::from_secs(3));
-        decision
-    }
-
-    fn decide(&mut self, now: Instant, observation: Option<FeedbackObservation>) -> FeedbackDecision {
-        let Some(observation) = observation else {
-            self.previous = None;
-            self.stable_since = None;
-            self.last_change = now;
-            self.limit = 0;
-            return FeedbackDecision {
-                limit: 0,
-                reason: FeedbackReason::Unavailable,
-            };
-        };
-        let previous = self.previous.replace((now, observation));
-        let contiguous = previous.is_some_and(|(at, old)| {
-            now.saturating_duration_since(at) <= Duration::from_secs(2)
-                && old.generation == observation.generation
-                && matches!((old.completed_units, observation.completed_units), (Some(a), Some(b)) if b >= a)
-        });
-        let advanced = previous
-            .is_some_and(|(_, old)| matches!((old.completed_units, observation.completed_units), (Some(a), Some(b)) if b > a));
-        let old_limit = self.limit;
-        let reason = if observation.max_concurrency == 0 {
-            self.limit = 0;
-            self.stable_since = None;
-            FeedbackReason::CapReduced
-        } else if observation.pressure == ResourcePressure::Busy {
-            self.limit = (self.limit / 2).max(1).min(observation.max_concurrency);
-            self.stable_since = None;
-            // Repeated busy windows restart cooldown, even at the minimum cap.
-            self.last_change = now;
-            FeedbackReason::Pressure
-        } else if observation.max_concurrency < self.limit {
-            self.limit = observation.max_concurrency;
-            self.stable_since = None;
-            FeedbackReason::CapReduced
-        } else if observation.pressure == ResourcePressure::Unknown || observation.completed_units.is_none() {
-            self.limit = 1.min(observation.max_concurrency);
-            self.stable_since = None;
-            FeedbackReason::LowConfidence
-        } else if !contiguous || self.limit == 0 {
-            self.limit = 1.min(observation.max_concurrency);
-            self.stable_since = None;
-            self.last_change = now;
-            FeedbackReason::NewBaseline
-        } else {
-            let since = *self.stable_since.get_or_insert(now);
-            if advanced
-                && now.saturating_duration_since(since) >= Duration::from_secs(8)
-                && now.saturating_duration_since(self.last_change) >= Duration::from_secs(8)
-                && self.limit < observation.max_concurrency
-            {
-                self.limit += 1;
-                FeedbackReason::Probe
-            } else {
-                FeedbackReason::Recovering
-            }
-        };
-        if old_limit != self.limit {
-            self.last_change = now;
-        }
-        FeedbackDecision {
-            limit: self.limit,
-            reason,
-        }
-    }
-}
-
-impl AdaptiveAdmission {
-    /// Construct a bounded owner, with zero initial capacity allowed for unknown
-    /// input. The immutable maximum reserves at most 65,535 lease slots once;
-    /// acquiring or cloning a permit does not allocate a per-unit Arc.
-    pub fn new(maximum: NonZeroU16, limit: u32) -> Arc<Self> {
-        let maximum_u32 = u32::from(maximum.get());
-        let maximum = usize::from(maximum.get());
-        let slots = (0..maximum.div_ceil(64))
-            .map(|chunk| {
-                let count = (maximum - chunk * 64).min(64);
-                PermitSlots {
-                    free: AtomicU64::new(if count == 64 { u64::MAX } else { (1_u64 << count) - 1 }),
-                    references: std::array::from_fn(|_| AtomicUsize::new(0)),
-                }
-            })
-            .collect::<Vec<_>>()
-            .into_boxed_slice();
-        Arc::new(Self {
-            state: AtomicU64::new(u64::from(limit.min(maximum_u32)) << 32),
-            origin: Instant::now(),
-            valid_until_millis: AtomicU64::new(u64::MAX),
-            changed: Notify::new(),
-            queued: AtomicUsize::new(0),
-            turn: Semaphore::new(1),
-            maximum: maximum_u32,
-            slots,
-            next_chunk: AtomicUsize::new(0),
-        })
-    }
-
-    /// Publish a cap bounded by the immutable maximum, without losing held work
-    /// or changing owner identity.
-    pub fn set_limit(&self, limit: u32) {
-        let limit = limit.min(self.maximum);
-        let _ = self.state.try_update(Ordering::AcqRel, Ordering::Acquire, |state| {
-            Some((u64::from(limit) << 32) | (state & u64::from(u32::MAX)))
-        });
-        self.changed.notify_one();
-    }
-
-    /// Publish a cap with a finite observation lifetime. Expired observations
-    /// block new admissions even if the sampling worker has stopped. Existing
-    /// permits are retained. A single resource owner must serialize publishers.
-    ///
-    /// Capacity is published before freshness, so a concurrent reader can
-    /// temporarily reject a fresh update but cannot use an increased lifetime
-    /// with the previous, potentially larger capacity.
-    pub fn set_sampled_limit(&self, limit: u32, valid_for: Duration) {
-        self.set_limit(limit);
-        let deadline = self
-            .elapsed_millis()
-            .saturating_add(u64::try_from(valid_for.as_millis()).unwrap_or(u64::MAX));
-        self.valid_until_millis.store(deadline, Ordering::Release);
-        self.changed.notify_one();
-    }
-
-    fn elapsed_millis(&self) -> u64 {
-        u64::try_from(self.origin.elapsed().as_millis()).unwrap_or(u64::MAX)
-    }
-
-    fn fresh(&self) -> bool {
-        self.elapsed_millis() < self.valid_until_millis.load(Ordering::Acquire)
-    }
-
-    /// Read the cap and active count from the same atomic observation.
-    pub fn snapshot(&self) -> AdaptiveAdmissionSnapshot {
-        // Acquire freshness first: a newly published lease must not be paired
-        // with capacity loaded before its preceding reduction.
-        let fresh = self.fresh();
-        let state = self.state.load(Ordering::Acquire);
-        AdaptiveAdmissionSnapshot {
-            limit: if fresh {
-                u32::try_from(state >> 32).unwrap_or(u32::MAX)
-            } else {
-                0
-            },
-            active: u32::try_from(state & u64::from(u32::MAX)).unwrap_or(u32::MAX),
-        }
-    }
-
-    /// Admit immediately if capacity remains, without a per-unit allocation.
-    /// The uncontended path uses atomics; refunding a racing acquisition may
-    /// wake the FIFO head through Tokio's notification queue.
-    pub fn try_acquire(self: &Arc<Self>) -> Option<AdaptivePermit> {
-        if self.queued.load(Ordering::Acquire) != 0 {
-            return None;
-        }
-        let permit = self.acquire_capacity()?;
-        // A waiter can register between the first queue check and the capacity
-        // CAS. Refund before exposing the permit rather than stealing its turn.
-        (self.queued.load(Ordering::Acquire) == 0).then_some(permit)
-    }
-
-    fn acquire_capacity(self: &Arc<Self>) -> Option<AdaptivePermit> {
-        if !self.fresh() {
-            return None;
-        }
-        let state = self.state.load(Ordering::Acquire);
-        if state & u64::from(u32::MAX) >= state >> 32 {
-            return None;
-        }
-        let slot = self.reserve_slot()?;
-        if self
-            .state
-            .try_update(Ordering::AcqRel, Ordering::Acquire, |state| {
-                let active = state & u64::from(u32::MAX);
-                (active < state >> 32).then(|| state + 1)
-            })
-            .is_err()
-        {
-            self.release_slot(slot);
-            self.changed.notify_one();
-            return None;
-        }
-        self.slots[slot / 64].references[slot % 64].store(1, Ordering::Release);
-        let permit = AdaptivePermit {
-            owner: self.clone(),
-            slot,
-        };
-        // The capacity CAS can retry across an observation expiry. Recheck
-        // before giving ownership to the caller; dropping refunds this unit.
-        self.fresh().then_some(permit)
-    }
-
-    fn reserve_slot(&self) -> Option<usize> {
-        let first = self.next_chunk.fetch_add(1, Ordering::Relaxed) % self.slots.len();
-        for offset in 0..self.slots.len() {
-            let chunk_index = (first + offset) % self.slots.len();
-            let free = &self.slots[chunk_index].free;
-            if let Ok(bits) = free.try_update(Ordering::AcqRel, Ordering::Acquire, |bits| (bits != 0).then(|| bits & (bits - 1)))
-            {
-                // trailing_zeros is in 0..64 because the zero word is rejected.
-                return Some(chunk_index * 64 + bits.trailing_zeros() as usize);
-            }
-        }
-        None
-    }
-
-    fn release_slot(&self, slot: usize) {
-        self.slots[slot / 64].free.fetch_or(1_u64 << (slot % 64), Ordering::Release);
-    }
-
-    /// Wait for capacity. The future owns no permit until it returns; dropping
-    /// a pending future leaves capacity untouched. Wrap in the caller's existing
-    /// cancellation/deadline select before acquiring any storage commit locks.
-    pub async fn acquire(self: &Arc<Self>) -> AdaptivePermit {
-        if let Some(permit) = self.try_acquire() {
-            return permit;
-        }
-        self.queued.fetch_add(1, Ordering::AcqRel);
-        let _queued = QueuedAdmission(self);
-        // Retain the successful permit inside its Result. No code can close
-        // this private semaphore, so acquisition cannot return an error.
-        let _turn = self.turn.acquire().await;
-        loop {
-            let notified = self.changed.notified();
-            tokio::pin!(notified);
-            // Register before checking capacity to avoid a lost release/resize.
-            notified.as_mut().enable();
-            if let Some(permit) = self.acquire_capacity() {
-                return permit;
-            }
-            notified.await;
-        }
-    }
-}
-
-impl Drop for AdaptivePermit {
-    fn drop(&mut self) {
-        if self.owner.slots[self.slot / 64].references[self.slot % 64].fetch_sub(1, Ordering::AcqRel) != 1 {
-            return;
-        }
-        // Publish the free slot before exposing another admission capacity.
-        self.owner.release_slot(self.slot);
-        self.owner.state.fetch_sub(1, Ordering::AcqRel);
-        self.owner.changed.notify_one();
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[tokio::test]
-    async fn node_progress_requires_explicit_completion_and_never_wraps() {
-        let node = NodeAdmission::new(NonZeroU16::new(2).expect("two slots"));
-        node.capacity.set_limit(2);
-        drop(node.acquire().await);
-        assert_eq!(node.completed(), Some(0), "cancellation is not progress");
-        node.acquire().await.complete();
-        assert_eq!(node.completed(), Some(1));
-        node.completed.store(u64::MAX - 1, Ordering::Release);
-        node.acquire().await.complete();
-        node.acquire().await.complete();
-        assert_eq!(node.completed(), None, "exhausted progress stays unknown");
-        assert_eq!(node.snapshot().active, 0);
-    }
-
-    #[tokio::test]
-    async fn cancelled_node_work_does_not_reserve_capacity_for_unrelated_native_tail() {
-        let node = NodeAdmission::new(NonZeroU16::new(1).expect("one slot"));
-        node.capacity.set_limit(1);
-        let physical = AdaptiveAdmission::new(NonZeroU16::new(1).expect("one disk unit"), 1);
-        let scheduling = node.acquire().await;
-        let native_tail = physical.acquire().await;
-        drop(scheduling);
-        assert_eq!(node.completed(), Some(0));
-        assert!(node.try_acquire().is_some(), "healthy resources can still be scheduled");
-        assert!(physical.try_acquire().is_none(), "the stalled disk remains charged");
-        drop(native_tail);
-        assert!(physical.try_acquire().is_some());
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn node_feedback_expires_without_revoking_held_work_or_fabricating_progress() {
-        let node = NodeAdmission::new(NonZeroU16::new(2).expect("two slots"));
-        let mut controller = FeedbackController::new();
-        let input = FeedbackObservation {
-            generation: 1,
-            max_concurrency: 2,
-            pressure: ResourcePressure::Stable,
-            completed_units: Some(99),
-        };
-        node.update(&mut controller, Some(input));
-        let held = node.acquire().await;
-        assert_eq!(node.completed(), Some(0));
-        tokio::time::advance(Duration::from_secs(4)).await;
-        assert_eq!(node.snapshot(), AdaptiveAdmissionSnapshot { limit: 0, active: 1 });
-        assert!(node.try_acquire().is_none());
-        held.complete();
-        assert_eq!(node.completed(), Some(1));
-        assert_eq!(node.snapshot().active, 0);
-    }
-
-    #[tokio::test]
-    async fn shared_unit_stays_admitted_until_native_worker_releases_its_last_clone() {
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(1).expect("one unit"), 1);
-        let outer = gate.acquire().await;
-        let native = outer.clone();
-        let (started_tx, started_rx) = tokio::sync::oneshot::channel();
-        let (release_tx, release_rx) = tokio::sync::oneshot::channel();
-        let worker = tokio::task::spawn_blocking(move || {
-            started_tx.send(()).expect("worker started");
-            release_rx.blocking_recv().expect("release native worker");
-            drop(native);
-        });
-        started_rx.await.expect("native worker is live");
-        drop(outer);
-        assert_eq!(gate.snapshot().active, 1);
-        assert!(gate.try_acquire().is_none());
-        gate.set_limit(0);
-        assert_eq!(gate.snapshot().active, 1);
-        release_tx.send(()).expect("release tail");
-        worker.await.expect("native tail completed");
-        assert_eq!(gate.snapshot().active, 0);
-        gate.set_limit(1);
-        assert!(gate.try_acquire().is_some());
-    }
-
-    #[tokio::test]
-    async fn bounded_slots_cross_word_boundaries_and_recycle_only_after_final_drop() {
-        let maximum = NonZeroU16::new(65).expect("partial second chunk");
-        let gate = AdaptiveAdmission::new(maximum, u32::MAX);
-        assert_eq!(gate.snapshot().limit, 65);
-        for _ in 0..4 {
-            let units: Vec<_> = (0..65).map(|_| gate.try_acquire().expect("free slot")).collect();
-            let tails = units.clone();
-            assert_eq!(gate.snapshot().active, 65);
-            assert!(gate.try_acquire().is_none());
-            drop(units);
-            assert_eq!(gate.snapshot().active, 65);
-            gate.set_limit(0);
-            for (remaining, tail) in (0..65).rev().zip(tails) {
-                drop(tail);
-                assert_eq!(gate.snapshot().active, remaining);
-            }
-            gate.set_limit(u32::MAX);
-            assert_eq!(gate.snapshot().limit, 65);
-        }
-    }
-
-    #[tokio::test(flavor = "multi_thread", worker_threads = 4)]
-    async fn shared_units_survive_concurrent_clone_drop_reuse_and_resize() {
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded owner"), 8);
-        let mut workers = Vec::new();
-        for _ in 0..16 {
-            let gate = gate.clone();
-            workers.push(tokio::spawn(async move {
-                for _ in 0..64 {
-                    let root = gate.acquire().await;
-                    let one = root.clone();
-                    let two = one.clone();
-                    drop(root);
-                    tokio::task::yield_now().await;
-                    drop(two);
-                    tokio::task::yield_now().await;
-                    drop(one);
-                }
-            }));
-        }
-        for limit in [0, 1, 8, 2, 0, 8] {
-            gate.set_limit(limit);
-            tokio::task::yield_now().await;
-        }
-        tokio::time::timeout(Duration::from_secs(5), async {
-            for worker in workers {
-                worker.await.expect("lease worker");
-            }
-        })
-        .await
-        .expect("all lease workers complete");
-        assert_eq!(gate.snapshot(), AdaptiveAdmissionSnapshot { limit: 8, active: 0 });
-        let full: Vec<_> = (0..8).map(|_| gate.try_acquire().expect("all slots returned")).collect();
-        assert!(gate.try_acquire().is_none());
-        drop(full);
-    }
-
-    #[tokio::test]
-    async fn queued_work_cannot_be_starved_by_new_fast_acquisitions() {
-        use std::task::{Context, Poll, Waker};
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 1);
-        let held = gate.acquire().await;
-        let mut first = Box::pin(gate.acquire());
-        let mut second = Box::pin(gate.acquire());
-        let mut context = Context::from_waker(Waker::noop());
-        assert!(first.as_mut().poll(&mut context).is_pending());
-        assert!(second.as_mut().poll(&mut context).is_pending());
-        drop(held);
-        assert!(gate.try_acquire().is_none(), "new work must not steal a parked worker's turn");
-        assert!(second.as_mut().poll(&mut context).is_pending(), "later work cannot pass the queue head");
-        let Poll::Ready(first_permit) = first.as_mut().poll(&mut context) else {
-            panic!("queue head must run")
-        };
-        drop(first_permit);
-        let Poll::Ready(second_permit) = second.as_mut().poll(&mut context) else {
-            panic!("second waiter must run")
-        };
-        drop(second_permit);
-        assert!(gate.try_acquire().is_some(), "drained queue restores the atomic fast path");
-    }
-
-    #[tokio::test]
-    async fn cancelling_queued_work_preserves_fifo_capacity_and_fast_path() {
-        use std::task::{Context, Poll, Waker};
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        let mut head = Box::pin(gate.acquire());
-        let mut middle = Box::pin(gate.acquire());
-        let mut tail = Box::pin(gate.acquire());
-        let mut context = Context::from_waker(Waker::noop());
-        assert!(head.as_mut().poll(&mut context).is_pending());
-        assert!(middle.as_mut().poll(&mut context).is_pending());
-        assert!(tail.as_mut().poll(&mut context).is_pending());
-        drop(middle);
-        drop(head);
-        gate.set_limit(1);
-        assert!(
-            gate.try_acquire().is_none(),
-            "cancelled head must not expose the surviving waiter's capacity"
-        );
-        let Poll::Ready(permit) = tail.as_mut().poll(&mut context) else {
-            panic!("surviving waiter must run")
-        };
-        drop(permit);
-        assert!(gate.try_acquire().is_some());
-
-        gate.set_limit(0);
-        let mut cancelled = Box::pin(gate.acquire());
-        assert!(cancelled.as_mut().poll(&mut context).is_pending());
-        drop(cancelled);
-        gate.set_limit(1);
-        assert!(
-            gate.try_acquire().is_some(),
-            "cancelling the final waiter must not strand an admission turn"
-        );
-    }
-
-    #[tokio::test]
-    async fn cancelling_a_notified_head_hands_capacity_to_the_next_waiter() {
-        use std::task::{Context, Poll, Waker};
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(2).expect("two units"), 0);
-        let mut head = Box::pin(gate.acquire());
-        let mut next = Box::pin(gate.acquire());
-        let mut context = Context::from_waker(Waker::noop());
-        assert!(head.as_mut().poll(&mut context).is_pending());
-        assert!(next.as_mut().poll(&mut context).is_pending());
-        gate.set_limit(1);
-        // Cancel after notification, before the head consumes its capacity.
-        drop(head);
-        let Poll::Ready(permit) = next.as_mut().poll(&mut context) else {
-            panic!("successor must observe capacity without another notification")
-        };
-        assert_eq!(gate.snapshot(), AdaptiveAdmissionSnapshot { limit: 1, active: 1 });
-        drop(permit);
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn one_resize_serves_multiple_fifo_turns_but_stored_wakes_cannot_bypass_expiry() {
-        use std::task::{Context, Poll, Waker};
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(2).expect("two units"), 0);
-        let mut first = Box::pin(gate.acquire());
-        let mut second = Box::pin(gate.acquire());
-        let mut context = Context::from_waker(Waker::noop());
-        assert!(first.as_mut().poll(&mut context).is_pending());
-        assert!(second.as_mut().poll(&mut context).is_pending());
-        gate.set_sampled_limit(2, Duration::from_secs(1));
-        let Poll::Ready(one) = first.as_mut().poll(&mut context) else {
-            panic!("first turn must observe resize")
-        };
-        let Poll::Ready(two) = second.as_mut().poll(&mut context) else {
-            panic!("second turn must observe resize without releasing the first unit")
-        };
-        tokio::time::advance(Duration::from_secs(1)).await;
-        drop(one);
-        drop(two);
-        let mut expired = Box::pin(gate.acquire());
-        assert!(expired.as_mut().poll(&mut context).is_pending(), "stored wake is not a fresh lease");
-        gate.set_sampled_limit(1, Duration::from_secs(1));
-        let Poll::Ready(renewed) = expired.as_mut().poll(&mut context) else {
-            panic!("fresh observation must wake expired admission")
-        };
-        drop(renewed);
-        assert_eq!(gate.snapshot().active, 0);
-    }
-    use std::future::Future;
-    use std::task::Poll;
-
-    #[tokio::test]
-    async fn shrink_preserves_all_held_units_and_blocks_until_below_new_limit() {
-        let owner = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 3);
-        let first = owner.acquire().await;
-        let second = owner.acquire().await;
-        let third = owner.acquire().await;
-        owner.set_limit(1);
-        assert_eq!(owner.snapshot(), AdaptiveAdmissionSnapshot { limit: 1, active: 3 });
-        drop(first);
-        assert!(owner.try_acquire().is_none());
-        drop(second);
-        assert!(owner.try_acquire().is_none());
-        drop(third);
-        let fourth = owner.try_acquire().expect("last old permit returned to original owner");
-        assert_eq!(owner.snapshot(), AdaptiveAdmissionSnapshot { limit: 1, active: 1 });
-        drop(fourth);
-        assert_eq!(owner.snapshot().active, 0);
-    }
-
-    #[tokio::test]
-    async fn zero_capacity_and_cancelled_waiter_do_not_leak_or_bypass() {
-        let owner = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        let mut waiting = Box::pin(owner.acquire());
-        std::future::poll_fn(|cx| {
-            assert!(waiting.as_mut().poll(cx).is_pending());
-            Poll::Ready(())
-        })
-        .await;
-        drop(waiting);
-        assert_eq!(owner.snapshot(), AdaptiveAdmissionSnapshot { limit: 0, active: 0 });
-        assert!(owner.try_acquire().is_none());
-        owner.set_limit(1);
-        let permit = owner.acquire().await;
-        assert!(owner.try_acquire().is_none());
-        drop(permit);
-        assert_eq!(owner.snapshot().active, 0);
-    }
-
-    #[tokio::test]
-    async fn resize_wakes_a_registered_waiter_without_an_unrelated_release() {
-        let owner = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        let mut waiting = Box::pin(owner.acquire());
-        std::future::poll_fn(|cx| {
-            assert!(waiting.as_mut().poll(cx).is_pending());
-            Poll::Ready(())
-        })
-        .await;
-        owner.set_limit(1);
-        let permit = tokio::time::timeout(std::time::Duration::from_secs(1), waiting)
-            .await
-            .expect("resize notification");
-        assert_eq!(owner.snapshot().active, 1);
-        drop(permit);
-    }
-
-    #[tokio::test(flavor = "multi_thread", worker_threads = 4)]
-    async fn concurrent_resize_release_and_acquire_keep_the_owner_count_exact() {
-        let owner = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 8);
-        let mut tasks = Vec::new();
-        for _ in 0..16 {
-            let owner = owner.clone();
-            tasks.push(tokio::spawn(async move {
-                for _ in 0..100 {
-                    let permit = owner.acquire().await;
-                    tokio::task::yield_now().await;
-                    drop(permit);
-                }
-            }));
-        }
-        for limit in [1, 0, 4, 2, 8, 0, u32::MAX, 1] {
-            owner.set_limit(limit);
-            tokio::task::yield_now().await;
-        }
-        owner.set_limit(8);
-        tokio::time::timeout(std::time::Duration::from_secs(5), async {
-            for task in tasks {
-                task.await.expect("admission task");
-            }
-        })
-        .await
-        .expect("all waiters make progress");
-        assert_eq!(owner.snapshot(), AdaptiveAdmissionSnapshot { limit: 8, active: 0 });
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn expired_observation_blocks_new_work_without_revoking_held_permits() {
-        let owner = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        owner.set_sampled_limit(2, Duration::from_secs(2));
-        let held = owner.acquire().await;
-        tokio::time::advance(Duration::from_secs(2)).await;
-        assert!(owner.try_acquire().is_none());
-        assert_eq!(owner.snapshot(), AdaptiveAdmissionSnapshot { limit: 0, active: 1 });
-        drop(held);
-        assert_eq!(owner.snapshot().active, 0);
-        owner.set_sampled_limit(1, Duration::from_secs(2));
-        let fresh = owner.try_acquire().expect("fresh observation reopens same owner");
-        owner.set_sampled_limit(8, Duration::ZERO);
-        assert!(owner.try_acquire().is_none());
-        drop(fresh);
-        assert_eq!(owner.snapshot().active, 0);
-    }
-
-    fn healthy(completed: u64) -> FeedbackObservation {
-        FeedbackObservation {
-            generation: 1,
-            max_concurrency: 8,
-            pressure: ResourcePressure::Stable,
-            completed_units: Some(completed),
-        }
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn controller_grows_one_unit_per_cooldown_and_cuts_pressure_immediately() {
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        let mut controller = FeedbackController::new();
-        assert_eq!(controller.update(&gate, Some(healthy(0))).limit, 1);
-        for completed in 1..=25 {
-            tokio::time::advance(Duration::from_secs(1)).await;
-            let limit = controller.update(&gate, Some(healthy(completed))).limit;
-            assert_eq!(
-                limit,
-                if completed < 9 {
-                    1
-                } else if completed < 17 {
-                    2
-                } else if completed < 25 {
-                    3
-                } else {
-                    4
-                }
-            );
-        }
-        let held = gate.try_acquire().expect("existing work");
-        let mut pressure = healthy(26);
-        pressure.pressure = ResourcePressure::Busy;
-        assert_eq!(
-            controller.update(&gate, Some(pressure)),
-            FeedbackDecision {
-                limit: 2,
-                reason: FeedbackReason::Pressure
-            }
-        );
-        pressure.max_concurrency = 0;
-        assert_eq!(controller.update(&gate, Some(pressure)).limit, 0);
-        assert_eq!(gate.snapshot().active, 1);
-        assert!(gate.try_acquire().is_none());
-        drop(held);
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn controller_missing_low_sample_stalled_or_changed_sources_never_expand() {
-        for kind in 0..5 {
-            let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-            let mut controller = FeedbackController::new();
-            for tick in 0..32 {
-                let mut observation = healthy(tick);
-                match kind {
-                    0 => observation.pressure = ResourcePressure::Unknown,
-                    1 => observation.completed_units = None,
-                    2 => observation.completed_units = Some(0),
-                    3 => observation.generation = tick,
-                    _ => {}
-                }
-                tokio::time::advance(Duration::from_secs(if kind == 4 { 3 } else { 1 })).await;
-                assert!(controller.update(&gate, Some(observation)).limit <= 1, "kind {kind}");
-            }
-            assert_eq!(
-                controller.update(&gate, None),
-                FeedbackDecision {
-                    limit: 0,
-                    reason: FeedbackReason::Unavailable
-                }
-            );
-            assert!(gate.try_acquire().is_none());
-        }
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn controller_resumes_after_zero_cap_without_requiring_impossible_progress() {
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        let mut controller = FeedbackController::new();
-        let mut sample = healthy(0);
-        sample.max_concurrency = 0;
-        assert_eq!(controller.update(&gate, Some(sample)).limit, 0);
-        tokio::time::advance(Duration::from_secs(1)).await;
-        sample.max_concurrency = 8;
-        assert_eq!(controller.update(&gate, Some(sample)).limit, 1);
-        assert!(gate.try_acquire().is_some(), "minimum progress restarts inside the restored cap");
-    }
-
-    #[tokio::test(start_paused = true)]
-    async fn controller_applies_pressure_reduction_and_lower_cap_in_the_same_window() {
-        let gate = AdaptiveAdmission::new(NonZeroU16::new(8).expect("bounded test owner"), 0);
-        let mut controller = FeedbackController::new();
-        for completed in 0..=25 {
-            controller.update(&gate, Some(healthy(completed)));
-            tokio::time::advance(Duration::from_secs(1)).await;
-        }
-        assert_eq!(gate.snapshot().limit, 4);
-        let mut sample = healthy(26);
-        sample.max_concurrency = 3;
-        sample.pressure = ResourcePressure::Busy;
-        assert_eq!(controller.update(&gate, Some(sample)).limit, 2);
-    }
-}

crates/concurrency/src/foreground.rs

@@ -1,568 +0,0 @@
-// Copyright 2026 RustFS Team
-// Licensed under the Apache License, Version 2.0.
-
-//! Bounded foreground completion windows for background resource feedback.
-//! Recording uses fixed atomic storage, without an allocation, mutex or task
-//! per request. A single reader rotates banks and waits for the retired bank's
-//! writers before reading it. Delayed writers delay publication, not the
-//! recorded window boundary. Percentiles are intervals, never exact values.
-
-use std::sync::atomic::{AtomicBool, AtomicU64, AtomicUsize, Ordering};
-use std::time::{Duration, Instant};
-
-const SUB_BUCKETS: usize = 64;
-const FINITE_BUCKETS: usize = 27 * SUB_BUCKETS;
-const BUCKETS: usize = FINITE_BUCKETS + 1;
-const MAX_MICROS: u64 = u32::MAX as u64;
-
-/// Bounded foreground categories; no bucket, object or request identifiers.
-#[derive(Clone, Copy, Debug, PartialEq, Eq)]
-pub enum ForegroundClass {
-    /// Object body reads.
-    Read,
-    /// Object writes, including multipart and copy operations.
-    Write,
-    /// Metadata and listing operations.
-    Metadata,
-    /// Other operations, including rejection before typed dispatch.
-    Other,
-}
-
-impl ForegroundClass {
-    /// Stable order shared by window summaries.
-    pub const ALL: [Self; 4] = [Self::Read, Self::Write, Self::Metadata, Self::Other];
-
-    /// Fixed metric label.
-    pub const fn as_str(self) -> &'static str {
-        match self {
-            Self::Read => "read",
-            Self::Write => "write",
-            Self::Metadata => "metadata",
-            Self::Other => "other",
-        }
-    }
-
-    const fn index(self) -> usize {
-        match self {
-            Self::Read => 0,
-            Self::Write => 1,
-            Self::Metadata => 2,
-            Self::Other => 3,
-        }
-    }
-}
-
-/// Outcome at the observed response-body boundary, not a client ACK.
-#[derive(Clone, Copy, Debug, PartialEq, Eq)]
-pub enum CompletionOutcome {
-    /// Successful response body completed.
-    Success,
-    /// Error response, service error, or body-stream failure.
-    Failed,
-    /// The request/body was dropped before completion.
-    Cancelled,
-}
-
-/// Inclusive bounds containing the observed percentile.
-#[derive(Clone, Copy, Debug, PartialEq, Eq)]
-pub struct LatencyBounds {
-    /// Inclusive lower bound in nanoseconds.
-    pub lower_nanos: u64,
-    /// Inclusive upper bound; absent for the overflow bucket.
-    pub upper_nanos: Option<u64>,
-}
-
-/// One class in a complete observation window.
-#[derive(Clone, Copy, Debug, Default, PartialEq, Eq)]
-pub struct ClassCompletions {
-    /// All finishes, including errors and cancellation.
-    pub finished: u64,
-    /// Error responses, service errors and body-stream failures.
-    pub failed: u64,
-    /// Censored observations; these must not be treated as successful latency.
-    pub cancelled: u64,
-    /// P99 of non-cancelled finishes, when at least 100 samples exist.
-    /// This count permits computing the rank; it is not statistical confidence.
-    pub p99: Option<LatencyBounds>,
-}
-
-/// Clock readings bracketing an atomic event-window cut. A scheduler pause
-/// between these readings is visible, rather than disguised as precise timing.
-#[derive(Clone, Copy, Debug)]
-pub struct WindowBoundary {
-    /// Read immediately before the event cut.
-    pub earliest: Instant,
-    /// Read immediately after the event cut.
-    pub latest: Instant,
-}
-
-/// Complete windows retain their actual cut brackets, even if publication waits.
-#[derive(Clone, Copy, Debug)]
-pub struct ForegroundWindow {
-    /// Start of this event window.
-    pub start: WindowBoundary,
-    /// Cut of the retired bank, possibly older than publication time.
-    pub end: WindowBoundary,
-    /// Request starts in this window, including pre-dispatch rejection.
-    pub started: Option<u64>,
-    /// Requests still alive at the cut; absent after accounting overflow.
-    pub in_flight: Option<u64>,
-    /// Per-class completions in ForegroundClass::ALL order.
-    pub classes: [Option<ClassCompletions>; 4],
-    /// False after a counter overflow or an inconsistent event count.
-    pub valid: bool,
-}
-
-struct ClassBank {
-    latencies: [AtomicU64; BUCKETS],
-    failed: AtomicU64,
-    cancelled: AtomicU64,
-}
-
-impl ClassBank {
-    fn new() -> Self {
-        Self {
-            latencies: std::array::from_fn(|_| AtomicU64::new(0)),
-            failed: AtomicU64::new(0),
-            cancelled: AtomicU64::new(0),
-        }
-    }
-}
-
-struct Bank {
-    writers: AtomicUsize,
-    started: AtomicU64,
-    overflowed: AtomicBool,
-    classes: [ClassBank; 4],
-}
-
-impl Bank {
-    fn new() -> Self {
-        Self {
-            writers: AtomicUsize::new(0),
-            started: AtomicU64::new(0),
-            overflowed: AtomicBool::new(false),
-            classes: std::array::from_fn(|_| ClassBank::new()),
-        }
-    }
-
-    fn increment(&self, counter: &AtomicU64) {
-        if counter.fetch_add(1, Ordering::Relaxed) == u64::MAX {
-            self.overflowed.store(true, Ordering::Relaxed);
-        }
-    }
-}
-
-struct Writer<'a>(&'a Bank);
-
-impl Drop for Writer<'_> {
-    fn drop(&mut self) {
-        self.0.writers.fetch_sub(1, Ordering::SeqCst);
-    }
-}
-
-/// One process-wide recorder. Its fixed storage is allocated with the owner.
-/// A request holds only a reference, timestamps and bounded enum state.
-pub struct ForegroundRecorder {
-    created_at: Instant,
-    epoch: AtomicU64,
-    reader_taken: AtomicBool,
-    banks: [Bank; 2],
-}
-
-impl Default for ForegroundRecorder {
-    fn default() -> Self {
-        Self::new()
-    }
-}
-
-impl ForegroundRecorder {
-    /// Construct empty accounting for one lifetime of the owning process.
-    pub fn new() -> Self {
-        Self {
-            created_at: Instant::now(),
-            epoch: AtomicU64::new(0),
-            reader_taken: AtomicBool::new(false),
-            banks: std::array::from_fn(|_| Bank::new()),
-        }
-    }
-
-    fn writer(&self) -> Writer<'_> {
-        loop {
-            let epoch = self.epoch.load(Ordering::SeqCst);
-            let bank = &self.banks[usize::from(epoch & 1 != 0)];
-            bank.writers.fetch_add(1, Ordering::SeqCst);
-            let writer = Writer(bank);
-            // Sequential ordering joins registration, the reader's cut, and
-            // its zero-writer check. A late old-epoch registrant cannot write
-            // into a bank that the reader has begun draining or reusing.
-            if self.epoch.load(Ordering::SeqCst) == epoch {
-                return writer;
-            }
-        }
-    }
-
-    /// Begin at the external service-call boundary. Drop records cancellation
-    /// unless finish has already recorded a response-body outcome.
-    pub fn start(&self) -> ForegroundObservation<'_> {
-        let started_at = Instant::now();
-        let writer = self.writer();
-        writer.0.increment(&writer.0.started);
-        ForegroundObservation {
-            recorder: self,
-            started_at: Some(started_at),
-            class: ForegroundClass::Other,
-        }
-    }
-
-    fn finish(&self, class: ForegroundClass, elapsed: Duration, outcome: CompletionOutcome) {
-        let writer = self.writer();
-        let bank = writer.0;
-        let class = &bank.classes[class.index()];
-        if outcome == CompletionOutcome::Cancelled {
-            bank.increment(&class.cancelled);
-            return;
-        }
-        let bucket = latency_bucket(elapsed);
-        bank.increment(&class.latencies[bucket]);
-        if outcome == CompletionOutcome::Failed {
-            bank.increment(&class.failed);
-        }
-    }
-
-    /// Claim the sole reader. It cannot be replaced after being dropped:
-    /// resetting its carried in-flight count would falsify later windows.
-    pub fn take_reader(&self) -> Option<ForegroundReader<'_>> {
-        self.reader_taken
-            .compare_exchange(false, true, Ordering::AcqRel, Ordering::Acquire)
-            .ok()?;
-        Some(ForegroundReader {
-            recorder: self,
-            start: WindowBoundary {
-                earliest: self.created_at,
-                latest: self.created_at,
-            },
-            retired: None,
-            in_flight: Some(0),
-        })
-    }
-}
-
-/// Non-cloneable request ownership, moved from the service future to its body.
-pub struct ForegroundObservation<'a> {
-    recorder: &'a ForegroundRecorder,
-    started_at: Option<Instant>,
-    class: ForegroundClass,
-}
-
-impl ForegroundObservation<'_> {
-    /// Set the first dispatched operation's bounded category.
-    pub fn set_class(&mut self, class: ForegroundClass) {
-        self.class = class;
-    }
-
-    /// Finish once. Later calls or Drop do not produce another observation.
-    pub fn finish(&mut self, outcome: CompletionOutcome) {
-        if let Some(started_at) = self.started_at.take() {
-            self.recorder.finish(self.class, started_at.elapsed(), outcome);
-        }
-    }
-}
-
-impl Drop for ForegroundObservation<'_> {
-    fn drop(&mut self) {
-        self.finish(CompletionOutcome::Cancelled);
-    }
-}
-
-/// Single-reader state; never acquired by a request path.
-pub struct ForegroundReader<'a> {
-    recorder: &'a ForegroundRecorder,
-    start: WindowBoundary,
-    retired: Option<(usize, WindowBoundary)>,
-    in_flight: Option<u64>,
-}
-
-impl ForegroundReader<'_> {
-    /// Cut a new window, or finish publishing an earlier cut. A live retired
-    /// writer returns None without blocking, discarding data or moving its cut.
-    pub fn sample(&mut self) -> Option<ForegroundWindow> {
-        if self.retired.is_none() {
-            let epoch = self.recorder.epoch.load(Ordering::SeqCst);
-            let next = epoch.checked_add(1)?;
-            let earliest = Instant::now();
-            self.recorder.epoch.store(next, Ordering::SeqCst);
-            let cut = WindowBoundary {
-                earliest,
-                latest: Instant::now(),
-            };
-            self.retired = Some((usize::from(epoch & 1 != 0), cut));
-        }
-        let (index, end) = self.retired?;
-        let bank = &self.recorder.banks[index];
-        if bank.writers.load(Ordering::SeqCst) != 0 {
-            return None;
-        }
-
-        let started = bank.started.swap(0, Ordering::Relaxed);
-        let overflowed = bank.overflowed.swap(false, Ordering::Relaxed);
-        let classes = std::array::from_fn(|index| {
-            let class = &bank.classes[index];
-            let failed = class.failed.swap(0, Ordering::Relaxed);
-            let cancelled = class.cancelled.swap(0, Ordering::Relaxed);
-            let buckets = std::array::from_fn::<_, BUCKETS, _>(|i| class.latencies[i].swap(0, Ordering::Relaxed));
-            let non_cancelled = buckets.iter().try_fold(0_u64, |sum, count| sum.checked_add(*count));
-            let finished = non_cancelled.and_then(|total| total.checked_add(cancelled));
-            let finished = finished.filter(|_| !overflowed && non_cancelled.is_some_and(|total| failed <= total))?;
-            Some(ClassCompletions {
-                finished,
-                failed,
-                cancelled,
-                p99: non_cancelled
-                    .filter(|total| *total >= 100)
-                    .and_then(|total| percentile_99(&buckets, total)),
-            })
-        });
-        let started = (!overflowed).then_some(started);
-        let finished = classes
-            .iter()
-            .try_fold(0_u64, |sum, class| sum.checked_add(class.as_ref()?.finished));
-        self.in_flight = self
-            .in_flight
-            .zip(finished)
-            .and_then(|(active, finished)| active.checked_add(started?)?.checked_sub(finished));
-        let window = ForegroundWindow {
-            start: self.start,
-            end,
-            started,
-            in_flight: self.in_flight,
-            classes,
-            valid: self.in_flight.is_some(),
-        };
-        self.start = end;
-        self.retired = None;
-        Some(window)
-    }
-}
-
-fn latency_bucket(elapsed: Duration) -> usize {
-    let micros = elapsed.as_nanos().div_ceil(1_000);
-    let Ok(micros) = u64::try_from(micros) else {
-        return FINITE_BUCKETS;
-    };
-    if micros > MAX_MICROS {
-        return FINITE_BUCKETS;
-    }
-    if micros < SUB_BUCKETS as u64 {
-        return usize::try_from(micros).unwrap_or(FINITE_BUCKETS);
-    }
-    let shift = 63 - micros.leading_zeros() - 6;
-    usize::try_from((u64::from(shift) + 1) * SUB_BUCKETS as u64 + (micros >> shift) - SUB_BUCKETS as u64)
-        .unwrap_or(FINITE_BUCKETS)
-}
-
-fn bucket_bounds(index: usize) -> LatencyBounds {
-    if index == FINITE_BUCKETS {
-        return LatencyBounds {
-            lower_nanos: MAX_MICROS * 1_000 + 1,
-            upper_nanos: None,
-        };
-    }
-    let (lower, upper) = if index < SUB_BUCKETS {
-        (index as u64, index as u64)
-    } else {
-        let shift = index / SUB_BUCKETS - 1;
-        let lower = (SUB_BUCKETS + index % SUB_BUCKETS) as u64 * (1_u64 << shift);
-        (lower, lower + (1_u64 << shift) - 1)
-    };
-    LatencyBounds {
-        lower_nanos: lower.saturating_sub(1) * 1_000 + u64::from(lower != 0),
-        upper_nanos: Some(upper * 1_000),
-    }
-}
-
-fn percentile_99(buckets: &[u64; BUCKETS], total: u64) -> Option<LatencyBounds> {
-    let rank = (u128::from(total) * 99).div_ceil(100);
-    let mut cumulative = 0_u128;
-    for (index, count) in buckets.iter().enumerate() {
-        cumulative += u128::from(*count);
-        if cumulative >= rank {
-            return Some(bucket_bounds(index));
-        }
-    }
-    None
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use std::sync::Arc;
-
-    #[test]
-    fn percentile_intervals_contain_each_bucket_edge_and_neighbour() {
-        for index in 0..BUCKETS {
-            let bounds = bucket_bounds(index);
-            assert_eq!(latency_bucket(Duration::from_nanos(bounds.lower_nanos)), index);
-            if let Some(upper) = bounds.upper_nanos {
-                assert!(upper >= bounds.lower_nanos);
-                assert_eq!(latency_bucket(Duration::from_nanos(upper)), index);
-                assert_eq!(latency_bucket(Duration::from_nanos(upper + 1)), index + 1);
-            }
-        }
-        assert_eq!(latency_bucket(Duration::MAX), FINITE_BUCKETS);
-    }
-
-    #[test]
-    fn body_ownership_carries_in_flight_counts_across_windows() {
-        let recorder = ForegroundRecorder::new();
-        let mut reader = recorder.take_reader().expect("first reader");
-        let mut request = recorder.start();
-        request.set_class(ForegroundClass::Read);
-        let first = reader.sample().expect("cut with a live request");
-        assert!(first.valid);
-        assert_eq!(first.started, Some(1));
-        assert_eq!(first.in_flight, Some(1));
-        assert_eq!(first.classes[0].expect("valid read count").finished, 0);
-
-        request.finish(CompletionOutcome::Success);
-        request.finish(CompletionOutcome::Failed);
-        drop(request);
-        let second = reader.sample().expect("body finished");
-        assert!(second.valid);
-        assert_eq!(second.started, Some(0));
-        assert_eq!(second.in_flight, Some(0));
-        let read = second.classes[0].expect("valid read summary");
-        assert_eq!((read.finished, read.failed, read.cancelled, read.p99), (1, 0, 0, None));
-        assert_eq!(first.end.earliest, second.start.earliest);
-        assert_eq!(first.end.latest, second.start.latest);
-    }
-
-    #[test]
-    fn cancellations_are_counted_without_fabricating_success_latencies() {
-        let recorder = ForegroundRecorder::new();
-        let mut reader = recorder.take_reader().expect("first reader");
-        for _ in 0..150 {
-            let mut request = recorder.start();
-            request.set_class(ForegroundClass::Write);
-        }
-        let window = reader.sample().expect("cancelled requests");
-        let write = window.classes[1].expect("known write count");
-        assert!(window.valid);
-        assert_eq!(window.in_flight, Some(0));
-        assert_eq!((write.finished, write.cancelled, write.failed, write.p99), (150, 150, 0, None));
-    }
-
-    #[test]
-    fn delayed_retired_writer_preserves_the_original_cut_and_next_bank() {
-        let recorder = ForegroundRecorder::new();
-        let mut reader = recorder.take_reader().expect("first reader");
-        let writer = recorder.writer();
-        assert!(reader.sample().is_none(), "never read a live writer's bank");
-        let original_cut = reader.retired.expect("pending cut").1;
-        let request = recorder.start(); // a start after the cut belongs to the new bank
-        writer.0.increment(&writer.0.started);
-        drop(writer);
-        let previous = reader.sample().expect("retired writer drained");
-        assert_eq!(previous.end.earliest, original_cut.earliest);
-        assert_eq!(previous.end.latest, original_cut.latest);
-        assert_eq!(previous.started, Some(1));
-        assert_eq!(previous.in_flight, Some(1));
-        drop(request);
-        let current = reader.sample().expect("next bank");
-        assert_eq!(current.started, Some(1));
-        assert_eq!(current.in_flight, Some(1));
-        assert_eq!(current.classes[3].expect("unclassified cancellation").cancelled, 1);
-    }
-
-    #[test]
-    fn completed_percentile_uses_rank_and_never_turns_maximum_into_p99() {
-        let recorder = ForegroundRecorder::new();
-        let mut reader = recorder.take_reader().expect("first reader");
-        for i in 0..100 {
-            let writer = recorder.writer();
-            writer.0.increment(&writer.0.started);
-            drop(writer);
-            recorder.finish(
-                ForegroundClass::Read,
-                if i == 99 {
-                    Duration::from_secs(60)
-                } else {
-                    Duration::from_millis(2)
-                },
-                if i == 99 {
-                    CompletionOutcome::Failed
-                } else {
-                    CompletionOutcome::Success
-                },
-            );
-        }
-        let window = reader.sample().expect("complete window");
-        let read = window.classes[0].expect("read summary");
-        assert_eq!((read.finished, read.failed, read.cancelled), (100, 1, 0));
-        let p99 = read.p99.expect("100 non-cancelled samples");
-        assert!(p99.lower_nanos <= 2_000_000);
-        assert!(p99.upper_nanos.is_some_and(|upper| (2_000_000..3_000_000).contains(&upper)));
-    }
-
-    #[test]
-    fn overflow_stays_unknown_and_reader_cannot_reset_event_accounting() {
-        let recorder = ForegroundRecorder::new();
-        {
-            let mut reader = recorder.take_reader().expect("first reader");
-            recorder.banks[0].started.store(u64::MAX, Ordering::Relaxed);
-            drop(recorder.start());
-            let invalid = reader.sample().expect("overflow window");
-            assert!(!invalid.valid);
-            assert_eq!(invalid.started, None);
-            assert_eq!(invalid.in_flight, None);
-            assert_eq!(invalid.classes, [None; 4]);
-            let later = reader.sample().expect("later empty window");
-            assert!(!later.valid, "a lost carried count cannot recover by inventing zero active work");
-        }
-        assert!(recorder.take_reader().is_none(), "reader replacement would lose live request ownership");
-    }
-
-    #[test]
-    fn concurrent_rotation_preserves_every_start_finish_and_cancellation() {
-        let recorder = Arc::new(ForegroundRecorder::new());
-        let mut reader = recorder.take_reader().expect("first reader");
-        let workers: Vec<_> = (0..4)
-            .map(|_| {
-                let recorder = Arc::clone(&recorder);
-                std::thread::spawn(move || {
-                    for index in 0..10_000 {
-                        let mut request = recorder.start();
-                        request.set_class(ForegroundClass::Read);
-                        if index % 10 != 0 {
-                            request.finish(CompletionOutcome::Success);
-                        }
-                    }
-                })
-            })
-            .collect();
-        let (mut started, mut finished, mut cancelled) = (0, 0, 0);
-        let mut add_window = |window: ForegroundWindow| {
-            assert!(window.valid, "concurrent event cuts must stay balanced");
-            started += window.started.expect("valid start count");
-            let reads = window.classes[0].expect("valid read count");
-            finished += reads.finished;
-            cancelled += reads.cancelled;
-        };
-        while workers.iter().any(|worker| !worker.is_finished()) {
-            if let Some(window) = reader.sample() {
-                add_window(window);
-            }
-            std::thread::yield_now();
-        }
-        for worker in workers {
-            worker.join().expect("writer must finish");
-        }
-        add_window(reader.sample().expect("drain after writers join"));
-        // A previously retired window may have been published by the last
-        // sample; cut once more to include the current, still open bank.
-        let final_window = reader.sample().expect("final open bank");
-        assert_eq!(final_window.in_flight, Some(0));
-        add_window(final_window);
-        assert_eq!((started, finished, cancelled), (40_000, 40_000, 4_000));
-    }
-}

crates/concurrency/src/lib.rs

@@ -36,10 +36,8 @@
 // Re-exported io-core type used by downstream timeout implementations.
 pub use rustfs_io_core::OperationProgress;
 
-pub mod adaptive;
 mod backpressure;
 mod deadlock;
-pub mod foreground;
 mod queue;
 pub mod workers;
 pub mod workload;

crates/concurrency/src/workers.rs

@@ -49,28 +49,12 @@ impl Workers {
 
     /// Acquire a worker slot, waiting until one becomes available.
     pub async fn take(&self) {
-        struct AcquireCancelGuard<'a> {
-            workers: &'a Workers,
-            completed: bool,
-        }
-        impl Drop for AcquireCancelGuard<'_> {
-            fn drop(&mut self) {
-                if !self.completed {
-                    self.workers.in_use.send_modify(|_| {});
-                }
-            }
-        }
-        let mut guard = AcquireCancelGuard {
-            workers: self,
-            completed: false,
-        };
         let permit = match self.semaphore.acquire().await {
             Ok(permit) => permit,
             // The semaphore is owned by `self` and never closed.
             Err(_) => unreachable!("worker semaphore is never closed"),
         };
         permit.forget(); // returned manually via give()
-        guard.completed = true;
         self.in_use.send_modify(|held| *held += 1);
         trace!(
             event = "worker_slot.acquire",
@@ -98,7 +82,6 @@ impl Workers {
         });
         if released {
             self.semaphore.add_permits(1);
-            self.in_use.send_modify(|_| {});
         }
         trace!(
             event = "worker_slot.release",
@@ -116,9 +99,7 @@ impl Workers {
     pub async fn wait(&self) {
         let mut rx = self.in_use.subscribe();
         // The sender is owned by `self`, so it outlives this borrow.
-        let _ = rx
-            .wait_for(|&held| held == 0 && self.semaphore.available_permits() == self.limit)
-            .await;
+        let _ = rx.wait_for(|&held| held == 0).await;
         debug!(
             event = "worker_slot.wait",
             component = "concurrency",
@@ -133,9 +114,7 @@ impl Workers {
 
     /// Return the current number of available worker slots.
     pub async fn available(&self) -> usize {
-        self.limit
-            .saturating_sub(*self.in_use.borrow())
-            .min(self.semaphore.available_permits())
+        self.limit.saturating_sub(*self.in_use.borrow())
     }
 }
 
@@ -238,36 +217,4 @@ mod tests {
         }
         assert_eq!(workers.available().await, 0);
     }
-
-    #[tokio::test]
-    async fn test_wait_does_not_complete_while_permit_is_in_transit_to_queued_taker() {
-        use std::future::Future;
-        use std::pin::pin;
-        use std::task::{Context, Poll, Waker};
-
-        let workers = Workers::new(1).unwrap();
-        workers.take().await;
-
-        let mut taker = pin!(workers.take());
-        let mut cx = Context::from_waker(Waker::noop());
-        assert!(matches!(taker.as_mut().poll(&mut cx), Poll::Pending));
-
-        let mut waiter = pin!(workers.wait());
-        assert!(matches!(waiter.as_mut().poll(&mut cx), Poll::Pending));
-
-        // Hand off the slot to the queued taker before `taker` is polled again.
-        workers.give().await;
-        assert_eq!(workers.available().await, 0);
-        assert!(
-            matches!(waiter.as_mut().poll(&mut cx), Poll::Pending),
-            "wait() must not complete while a permit is in transit to a woken taker"
-        );
-
-        assert!(matches!(taker.as_mut().poll(&mut cx), Poll::Ready(())));
-        assert!(matches!(waiter.as_mut().poll(&mut cx), Poll::Pending));
-
-        workers.give().await;
-        assert!(matches!(waiter.as_mut().poll(&mut cx), Poll::Ready(())));
-        assert_eq!(workers.available().await, 1);
-    }
 }

crates/config/README.md

@@ -116,29 +116,14 @@ when restoring `/rustfs/console` for a custom OEM build.
 - `RUSTFS_DATA_SCANNER_START_DELAY_SECS` (deprecated alias for compatibility)
 - `RUSTFS_SCANNER_IDLE_MODE` (canonical)
 - `RUSTFS_SCANNER_CACHE_SAVE_TIMEOUT_SECS` (canonical)
-- `RUSTFS_SCANNER_RECOVERY_BACKOFF_CAP_SECS` (canonical)
 - `RUSTFS_SCANNER_CYCLE_MAX_DURATION_SECS` (canonical)
 - `RUSTFS_SCANNER_CYCLE_MAX_OBJECTS` (canonical)
 - `RUSTFS_SCANNER_CYCLE_MAX_DIRECTORIES` (canonical)
-- `RUSTFS_SCANNER_MAX_CONCURRENT_SET_SCANS` (canonical)
-- `RUSTFS_SCANNER_MAX_CONCURRENT_DISK_SCANS` (canonical)
-- `RUSTFS_SCANNER_WALKERS_PER_DISK` (canonical)
 
 Scanner cycle budget controls:
 
 - When `RUSTFS_SCANNER_CYCLE_MAX_DURATION_SECS` is unset, the finite default is 1800 seconds (30 minutes), matching the scanner benchmark guidance.
 - An explicit `0` preserves the compatibility behavior of an unbounded runtime budget. Object and directory budgets likewise remain unbounded when explicitly set to `0`.
-
-Scanner recovery retry controls:
-
-- `RUSTFS_SCANNER_RECOVERY_BACKOFF_CAP_SECS` caps the wait between automatic startup recovery retries. The persistent admin key is `scanner.recovery_backoff_cap`; the environment takes precedence over the persistent value, and the unset default remains 600 seconds (10 minutes). Values must be positive integer seconds within the platform timer range; `0`, malformed values, and overflow are rejected. A valid environment override does not hide an invalid persistent value from `admin config set` validation.
-- The cap bounds the existing exponential retry schedule. Updating the persistent value applies to the next computed retry; it does not reschedule a wait already in progress. Retryable observations continue retrying after the short retry budget. Admin scanner status reports the effective value and source at `runtime_config.recovery_backoff_cap_seconds`.
-
-Scanner walker controls:
-
-- `RUSTFS_SCANNER_WALKERS_PER_DISK` sets how many bucket walkers each local disk of an erasure set runs; the default `1` walks one bucket per disk at a time, which is the right value for rotational media. On flash, `2` to `4` lets one disk walk several buckets concurrently, which matters most for single-disk deployments with many buckets, where the walk is otherwise fully serial. Values outside `1` through `16` are rejected by `admin config set`; an out-of-range or non-numeric environment value is logged as a scanner runtime config error (`state = "startup_apply_failed"`) and the scanner keeps the default. The walkers of one disk share that disk's scanner admission, so peers still see one scan session per physical disk, and `RUSTFS_SCANNER_MAX_CONCURRENT_DISK_SCANS` still caps the total.
-- The same value is available as the persistent `scanner.walkers_per_disk` key and is reported under `runtime_config.walkers_per_disk` in the admin scanner status; see [Scanner Runtime Controls](../../docs/operations/scanner-runtime-controls.md).
-
 - A timed-out cycle cancels cooperative scanner work, then fences its leader epoch before releasing the lease. An uncooperative I/O operation is dropped after the bounded shutdown window; its cursor is not claimed to be durable and the scanner reports `recovery-required` when the worker cannot stop cooperatively, the cycle state was not confirmed durable, or epoch fencing cannot be persisted.
 
 ## Mmap read environment aliases
@@ -281,7 +266,7 @@ Drive timeout profile preset:
   - `warn` (default): log warning and continue startup.
   - `fail`: abort startup with an error.
 
-RustFS production guidance remains direct-attached local POSIX filesystems. Network-mounted filesystems (for example `nfs`, `cifs`, `smb2`, and `fuse.*`) are treated as unsupported by this startup guard. On Windows, endpoints on network drives (UNC shares such as `\\server\share` and mapped drive letters) are treated the same way, even though Windows reports the server's filesystem (usually `NTFS`) for them.
+RustFS production guidance remains direct-attached local POSIX filesystems. Network-mounted filesystems (for example `nfs`, `cifs`, `smb2`, and `fuse.*`) are treated as unsupported by this startup guard.
 
 ## 📄 License
 

crates/config/src/constants/console.rs

@@ -69,17 +69,15 @@ pub const ENV_CONSOLE_RATE_LIMIT_RPM: &str = "RUSTFS_CONSOLE_RATE_LIMIT_RPM";
 pub const DEFAULT_CONSOLE_RATE_LIMIT_RPM: u32 = 100;
 
 /// Set the console authentication timeout in seconds
-/// Bounds how long the console waits for a single authenticated HTTP request before it is
-/// considered timed out. This is a request timeout, not the session lifetime: the Console
-/// credential lifetime issued by the OIDC callback is `RUSTFS_CONSOLE_OIDC_SESSION_DURATION`.
+/// Specifies how long a console authentication session remains valid
 /// Default: 3600 seconds (1 hour)
 /// Minimum: 300 seconds (5 minutes)
 /// Maximum: 86400 seconds (24 hours)
 pub const ENV_CONSOLE_AUTH_TIMEOUT: &str = "RUSTFS_CONSOLE_AUTH_TIMEOUT";
 
 /// Default console authentication timeout in seconds
-/// This is the default timeout for console authentication requests.
-/// After this timeout, a pending console request is abandoned.
+/// This is the default timeout for console authentication sessions.
+/// After this timeout, users need to re-authenticate to access the console.
 /// Default value: 3600 seconds (1 hour)
 /// Environment variable: RUSTFS_CONSOLE_AUTH_TIMEOUT
 /// Command line argument: --console-auth-timeout
@@ -87,26 +85,6 @@ pub const ENV_CONSOLE_AUTH_TIMEOUT: &str = "RUSTFS_CONSOLE_AUTH_TIMEOUT";
 /// Example: --console-auth-timeout 3600
 pub const DEFAULT_CONSOLE_AUTH_TIMEOUT: u64 = 3600;
 
-/// Console credential lifetime, in seconds, for Console credentials issued by the OIDC callback.
-///
-/// Process-level, Console-scoped setting: it applies to every OIDC provider and deliberately
-/// lives outside the per-provider `RUSTFS_IDENTITY_OPENID_*` namespace, so it is never
-/// suffix-scanned into a provider id and has no `identity_openid` KVS equivalent. The same value
-/// also sets how long the Console logout token stays valid.
-pub const ENV_CONSOLE_OIDC_SESSION_DURATION: &str = "RUSTFS_CONSOLE_OIDC_SESSION_DURATION";
-
-/// Default Console credential lifetime in seconds, used when the environment variable is unset,
-/// not a positive integer, or zero. Matches the value this path hard-coded before it was
-/// parameterised.
-pub const DEFAULT_CONSOLE_OIDC_SESSION_DURATION_SECS: usize = 3600;
-
-/// Upper bound on the Console credential lifetime in seconds (7 days).
-///
-/// Values above this are clamped with a warning. The bound keeps the moka logout cache's
-/// `time_to_live` well under its 1000-year assertion, and stops an operator from minting an
-/// effectively permanent Console STS credential on every node.
-pub const CONSOLE_OIDC_SESSION_DURATION_MAX_SECS: usize = 604_800;
-
 /// Toggle update check
 /// It controls whether to check for newer versions of rustfs
 /// Default value: true

crates/config/src/constants/heal.rs

@@ -197,10 +197,6 @@ pub const ENV_HEAL_MRF_REPLAY_BATCH: &str = "RUSTFS_HEAL_MRF_REPLAY_BATCH";
 /// Default behavior keeps the MRF intent pipeline enabled.
 pub const DEFAULT_HEAL_MRF_ENABLE: bool = true;
 
-/// Opt in to lifecycle V2 transitions after qualifying every reader and the rollback binary.
-pub const ENV_MRF_LIFECYCLE_V2_ENABLE: &str = "RUSTFS_MRF_LIFECYCLE_V2_ENABLE";
-pub const DEFAULT_MRF_LIFECYCLE_V2_ENABLE: bool = false;
-
 /// Default MRF queue capacity (matches MinIO's 100k MRF list ceiling).
 pub const DEFAULT_HEAL_MRF_QUEUE_SIZE: usize = 100_000;