#5804 added the on-disk segment budget to check_bucket_and_object_names, but PUT validates through check_put_object_args, which has its own checks and never calls it. An over-NAME_MAX key therefore still reached the disk layer and came back to the client as ENAMETOOLONG → InternalError 500, exactly the behavior #5785 reported.
Caught by re-running the acceptance suite against the locked build 4b2d79f5d, which contains #5804: S3-003 still failed with a 512-byte key.
Multipart is unaffected — check_new_multipart_args and check_multipart_object_args both route through check_object_args → check_bucket_and_object_names, which already carries the budget.
Verification: new test pins the same boundaries on check_put_object_args (255 ok / 256 rejected, byte-based via CJK, multi-segment long keys ok, __XLDIR__ budget for directory keys); cargo test -p rustfs-ecstore --lib -- bucket::utils 17 passed; cargo clippy -p rustfs-ecstore --all-targets clean; make pre-commit green.
* fix(admin): align replication-reset responses with madmin ResyncTargetsInfo shape
The replication-reset and replication-reset-status responses serialized
their shell as "Targets" and per-target fields in PascalCase, while
madmin-go ResyncTargetsInfo/ResyncTarget expect the "target" shell key
and lowercase field tags (arn/resetid/resyncStatus/replicationCount/
completedReplicationSize/failedReplicationCount/failedReplicationSize).
Go json decoding is case-insensitive per field, but Targets vs target,
Status vs resyncStatus and the size/count key names cannot match, so
mc replicate resync decoded empty results.
Rename the serde tags to the exact madmin wire shape, keep the
ResetBeforeDate/Error RustFS extension keys (unknown keys are ignored
by Go decoders), pin the shape with a snapshot unit test, and update
the e2e client DTO to decode the madmin shape.
* fix(admin): stream bare madmin DiffInfo documents from replication diff
POST /v3/replication/diff returned a single enveloped object
({Entries, IsTruncated, ScannedVersions}) while madmin-go
BucketReplicationDiff decodes the body with a json.Decoder loop over
bare DiffInfo documents. The envelope decoded as exactly one DiffInfo
with an empty object, so mc replicate diff printed a phantom empty row
instead of the real backlog.
Emit one DiffInfo JSON document per line by default, using the exact
madmin json tags (object/versionId/rStatus/deletemarker/lastModified;
Size stays as a RustFS extension key that Go decoders ignore). The
enveloped shape moves to the opt-in ?aggregate=true RustFS extension,
which remains the only carrier of scan-coverage metadata; a truncated
default-mode scan is surfaced via a warn tracing event instead of
in-stream. Pin both shapes with unit tests and tighten the e2e helper
to reject any envelope in the stream.
* feat(replication): validate replication config structure before persisting
PutBucketReplication accepted structurally invalid configurations that
MinIO's replication.Config.Validate rejects: empty or oversized rule
lists, duplicate or negative rule priorities, over-long rule IDs,
filters carrying more than one of Prefix/Tag/And, and delete marker
replication enabled on tag-filtered rules. Such configs persisted
silently and later produced undefined routing (e.g. ambiguous priority
ties) instead of failing the PUT.
Add validate_replication_config_structure as a pure function in
rustfs-replication (limits documented as constants), surface it through
the ecstore api facade, and run it first in the PUT capability gate so
defects are named before any metadata write. Missing Priority counts as
zero for the uniqueness check, matching Go's zero-value semantics. The
self-target rejection deliberately stays at set-remote-target, where the
endpoint is known; a config can never reference a self-pointing ARN.
Document the rule-level Destination.StorageClass contract (use the
remote target's storage_class instead) and renumber the acceptance
matrix e2e to unique priorities, which MinIO would also require.
* test(replication): pin duplicated wire types with boundary reconciliation tests
rustfs-filemeta (xl.meta disk format) and rustfs-replication (MRF/resync
persistence format) deliberately each own ReplicationStatusType,
VersionPurgeStatusType and ReplicationState; the boundary converts
between them via as_str(), whose From<&str> impls fall back to Empty on
unknown tokens — a variant added on one side silently degrades to Empty
on the other.
Add reconciliation tests in replication_filemeta_boundary: exhaustive
matches with no wildcard arm on both sides of both enums (a new variant
fails compilation until the mapping is reconsidered), string-token
round-trip asserts (a token the other side does not recognize fails
instead of quietly becoming Empty), and a full-field ReplicationState
round-trip. Cross-reference the tests from both type definitions.
Struct drift was already compile-guarded by the exhaustive struct
literals in the conversion functions.
* docs(replication): define split completion criteria and milestone sequence
The ecstore replication split plan had no completion measure — the
boundary scaffolding risked ossifying because nothing said when the
migration counts as done. Record the criteria in the module inventory:
done means the Required Contracts table's 'Current dependency to
remove' column is empty; the end state moves pool/resyncer/state into
crates/replication, with the boundary micro-files dissolving as code
crosses the crate line (batch-merging them beforehand is explicitly
rejected — the guard scripts anchor on their file names, so merging is
churn with zero functional gain; only datatypes.rs can retire early).
Sequence the remaining work as M2 (resyncer pure decision logic, after
the oversized function splits) → M3 (worker runtime, highest risk,
last) → M4 (retire boundaries and guard entries). Refresh the stale
first-step text — the event sink / runtime contracts already landed —
and update the split-plan status table accordingly.
* fix(replication): align structural validator with MinIO semantics after adversarial review
Three interop corrections found by adversarial review of the new
structural validator, plus review fallout fixes:
- Delete-marker replication is now rejected only for a direct Filter.Tag,
not for tags inside Filter.And — MinIO's validator only inspects the
direct tag, and mc replicate add --tags "k1=v1&k2=v2" (delete-marker
replication on by default) puts multiple tags into And.Tags, so the
stricter check rejected mc-generated configs MinIO accepts.
- Rule ID length is measured in bytes (Go len semantics), not chars —
a 255-char multibyte ID must not round-trip into a config MinIO
rejects.
- An empty <Tag/> element (no key) counts as absent, matching MinIO's
Tag.IsEmpty(); console form serializers emit empty tags, which would
otherwise trip the exactly-one-of and delete-marker checks.
Also: repair the store-uninitialized PUT test whose empty-rules fixture
now (correctly) fails structural validation before reaching the store
lookup; pin the previously untested startTime madmin key in the
reset-status shape test; and signal a truncated default-mode diff scan
via the x-rustfs-replication-diff-truncated response header — the bare
madmin stream has no envelope, so a truncated scan was otherwise
indistinguishable from a complete healthy one (madmin/mc ignore unknown
headers).
* test(e2e): activate SSE-S3 replication contract and pin resync fail-closed path
The SSE-S3 replication contract e2e was ignored under backlog#1291
(silent plaintext replication); the fail-closed gate in
replication_target_boundary.rs closed that hole, so the ignore reason
expired. Un-ignore the test — it now pins the current fail-closed
contract (FAILED status, failure event, readable encrypted source,
stable absence of all target versions), verified green.
Add test_bucket_replication_sse_s3_resync_stays_fail_closed: drives the
existing-object resync path (PUT ?replication-reset) over a FAILED
SSE-S3 object and asserts the resync generation reaches a terminal
state without ever materializing a target version, with the
stays-absent window also spanning fast-scanner heal cycles. The new
start_bucket_replication_reset helper doubles as the madmin
ResyncTargetsInfo shape assertion (target[0].arn/resetid) for the
reset-start response.
Refresh the stale nextest count commentary (the module is at 20 fast +
36 nightly = 56 tests by cargo nextest list; the SSE-S3-ignored note no
longer holds).
fix(ecstore): reject over-NAME_MAX object key segments up front and classify irreconcilable parity as corrupt metadata
Two defects found during release acceptance and the backlog#1776 investigation:
Object keys with any path segment longer than 255 bytes could never be stored (each segment maps to one on-disk directory entry), but the failure surfaced only when the disk layer hit ENAMETOOLONG, which leaked to clients as InternalError 500 (rustfs#5785). Validate the on-disk segment budget in check_bucket_and_object_names so such keys fail deterministically as ObjectNameInvalid (4xx) before any I/O. Directory-object keys (trailing '/') account for the __XLDIR__ suffix their final segment carries on disk.
object_quorum_from_meta conflated two very different no-quorum situations (rustfs#5801): stray or foreign metadata whose parity values are garbage produced the same retryable-looking ErasureReadQuorum (503) as a genuine partial outage, so clients retried unrecoverable reads and monitoring could not tell corruption from capacity loss. Now (a) parity counts outside [0, total_shards] are treated as invalid entries instead of being clamped to i32::MAX, which could poison common_parity's occurrence counting, and (b) when a full read quorum of disks answers but their parity values cannot be reconciled, the error is FileCorrupt — heal-actionable and non-retryable — while too-few-healthy-replies keeps returning ErasureReadQuorum.
Verification: 4 new unit tests (segment budget boundaries incl. byte-vs-char and __XLDIR__ budget; garbage parity sanitization; corrupt-vs-quorum classification), metadata::tests + utils::tests 62/62, set_disk+bucket suites 1214 passed with the single pre-existing heal_queue_marks_missing_versioning_state_as_missed cross-test flake also failing on a clean tree (not introduced here), clippy clean, make pre-commit green.
* test(replication): accept madmin nanosecond healthCheckDuration payloads
Red-phase TDD tests for P0-7: mc 'replicate add' sends the madmin default
healthCheckDuration=60s as a Go time.Duration nanosecond integer
(60000000000), which RustFS currently rejects as an unsupported field and
would misread as seconds. Also pins the defensive seconds-or-nanos read
for persisted bucket-targets metadata and the capability contract listing
healthCheckDuration as writable.
Currently failing (red):
- remote_target_request_accepts_go_duration_wire_values
- remote_target_request_accepts_legacy_seconds_health_check
- remote_target_health_check_duration_is_declared_writable
- bucket_target_reads_go_nanosecond_durations_defensively
- runtime_capabilities_response_reports_missing_topology_before_storage_init
* fix(replication): accept remote target healthCheckDuration nanoseconds
mc 'replicate add' always sends the madmin default healthcheck-seconds=60
serialized as a Go time.Duration nanosecond integer (60000000000), so the
default mc link-creation path (and 'mc replicate update') failed with
InvalidRequest. Move healthCheckDuration from the unsupported to the
writable remote-target field list; the capability contract in the runtime
capabilities response follows the constants automatically.
Fix the unit mismatch in both directions:
- Request parsing and persisted bucket-targets reads decode the value
defensively: below 10^7 it is legacy RustFS seconds, otherwise Go
time.Duration nanoseconds (also covers MinIO-written metadata).
totalDowntime shares the same wire shape and gets the same handling.
- The list-remote-targets admin response re-encodes only these two fields
as nanoseconds via a dedicated serialization path, leaving the persisted
seconds-based wire format untouched for existing readers.
The per-target health-check interval is accepted for mc compatibility but
not yet applied; the heartbeat keeps its global env-configured interval,
and the explicit 'healthcheck' update op stays rejected. disableProxy,
edge, and edgeSyncBeforeExpiry remain explicitly rejected.
The fallback future embeds the whole snapshot loader, and every object write nests a quota check several futures deep, so inlining it grew each write's state machine by the loader's full size — the debug-build 2MiB worker-stack overflow class fixed for bucket-config writes in #5648. Box the fallback at its call site; the allocation only happens on the degraded path.
Upgrading from a pre-v2 release leaves only the legacy .usage.json snapshot, which has no completeness marker and is demoted to non-authoritative, so every write to a quota-enabled bucket failed closed with a retryable 503 until the scanner's first complete cycle persisted .usage.v2.json — a production outage on large namespaces (issue #5716).
Quota admission now degrades to the last persisted per-bucket size: normalize_loaded_data_usage returns the pre-discard bucket sizes, the TTL-bounded snapshot cache retains them (carried forward through failed refreshes), and QuotaChecker::get_real_time_usage falls back to that baseline when the authoritative caches miss. The baseline is static between snapshot loads, so hard-quota enforcement is advisory for the duration of the degraded window — strictly tighter than beta.11 (usage treated as 0) and strictly more available than a blanket 503. Buckets absent from every persisted snapshot still fail closed, and removing a bucket's usage from the backend purges the baseline so a recreated bucket cannot inherit the dead incarnation's size.
Refs #5716
Both per-target replication methods assign the optimistic Completed status to rinfo before building put options, and the Err branch of replication_put_object_options returned rinfo unchanged. Since #5633 made the source encryption classification case-insensitive, managed SSE sources are rejected at this gate, and the rejection was reported as successful replication: the source object was marked COMPLETED, ObjectReplicationComplete was emitted, and nothing existed on the target.
Set replication_status = Failed (and record the error in the replicate_object branch) so the composite status, the OperationFailedReplication event, and MRF retries reflect the fail-closed outcome. This restores the contract pinned by test_bucket_replication_sse_kms_failure_contract, which timed out in the e2e-replication-nightly runs on 2026-08-03 and 2026-08-04.
* fix: address rc.1 release blockers
* fix: route release guards through architecture boundaries
* fix: close remaining rc.1 regression gaps
* refactor: group multipart listing options
* fix: resolve rc.1 CI regressions
* fix(ecstore): keep bucket-config writes off the caller's stack
A bucket-config write nests incarnation resolution (which can drive legacy
migration and a peer fan-out), a full metadata load, and `save` — itself an
object PUT that pulls in the whole erasure write path. Every request that
mutates bucket config is already several futures deep, so inlining all of
that into one state machine overflows the 2MiB worker stack in debug builds.
Two CI lanes aborted with SIGABRT on this:
ILM Integration (serial)
rustfs app::lifecycle_transition_api_test::
compensation_driven_complete_multipart_upload_still_transitions
Test and Lint (swift)
rustfs-protocols::swift_metadata_persistence::
swift_metadata_writes_are_durable
Neither test file is touched by this branch and both lanes are green on
main. Stack-pointer probing showed ~780KiB consumed between
`metadata_sys::update` and the config read alone, with single hops of
363KiB (`update` -> `acquire_config_write_guard_for_incarnation`), 125KiB
and 105KiB.
Box the deep sub-futures on both read-modify-write paths (`update` /
`update_checked` and `update_config_with` / `update_config_with_checked`)
so each guard's own state machine stays small. Behaviour is unchanged;
`update` -> guard drops to 253KiB and both tests pass on the default stack.
* fix(lifecycle): unbreak restore under the bucket generation fence
The ILM lane aborted on a stack overflow before reaching these, so they
were never reported; with that fixed, four restore tests fail. All four
are green on main and none of their test files are touched by this branch.
1. RestoreObject and ListMultipartUploads hard-required
`opts.expected_bucket_incarnation_id`, but `apply_bucket_generation_guard`
deliberately leaves it unset when no guard extension is present — only the
S3 access layer installs one. Every direct caller therefore got
`InternalError: ... bucket generation guard is missing`. Resolve the
current generation instead, the way the copy path already does. The fence
is unaffected: RestoreObject still re-reads the incarnation from disk and
compares before admitting the restore, and the multipart listing is
filtered by the value it resolves.
2. `restore_expiry_snapshot_matches` (new on this branch) rejected every
restored-copy expiry whose `restore_expires` had not already elapsed.
Whether the restored copy is due to expire is the ILM evaluator's
decision, made when it emitted DeleteRestoredAction; re-deriving it in
the set layer only adds a way for a legitimate action to be rejected.
The stale-event risk it appears to guard is already covered by the
surrounding snapshot match — a re-restore rewrites `restore_expires`,
so a replayed event fails the equality check. Drop the clause; the
fifteen identity clauses are unchanged.
Fixed:
rustfs app::lifecycle_transition_api_test::
restore_object_usecase_accepts_exactly_one_of_two_concurrent_restores
restore_object_usecase_completes_suspended_null_version_in_place
restore_object_usecase_reports_ongoing_conflict
rustfs-scanner::lifecycle_integration_test serial_tests::
test_restore_chain_local_read_expiry_keeps_remote_and_allows_re_restore
Verification: the CI ILM lane filter now runs 53/53 green locally.
* chore: address review follow-ups on this branch
Four items from the adversarial review that were still open.
- Restore the assertion `test_bucket_replication_replayed_delete_marker_
preserves_source_mtime_without_source_restart` is named for. The branch
had replaced the backlog#867 mtime check with `assert_replication_
converged`, which any successful replication satisfies, and deleted the
two helpers it needed — so the regression the test exists to catch would
now pass. This matters here specifically because the branch changes the
flag feeding `replication_delete_remove_options` and routes replay
through a new file and ordering.
- Drop `read_config_no_lock_preserve_empty`: zero production callers (the
one real consumer calls the `_with_metadata` variant directly). Its test
stanza now exercises that variant, so the coverage moves to live code
rather than being deleted.
- Revert the `bytesize` bump. It is a no-op: `Cargo.lock` already pinned
2.7.0 before this branch and is untouched, so the caret range already
resolved there. Nothing in the diff uses the crate.
- Split the AGENTS.md "Adversarial Validation" policy change out of this
branch. The edit is defensible on its own, but it relaxes the review gate
that this branch has to pass, so it should land as its own PR reviewed on
its own merits rather than bundled with the change that benefits from it.
The reverted hunks are unchanged and ready to re-apply.
Not changed, deliberately: the missing-sidecar path still fails closed.
`missing_bucket_incarnation_sidecar_for_new_metadata_fails_closed` pins
that on purpose, and serving a non-authoritative Object Lock state would
be the wrong trade. The residual concern stands and is recorded in review
— a crash between the two writes in `persist_new_and_set` leaves the
bucket unloadable until DeleteBucket+CreateBucket, and the repair branches
in `migrate_legacy_metadata` and `make_bucket` are unreachable dead code
for that case. Resolving it needs the read path and the (transaction-lock
holding) repair path to be separated, which is more than a follow-up edit.
* test(ci): serialize the new bucket-incarnation tests
The five tests this branch adds around the incarnation / lifecycle fence
drive `init_bucket_metadata_sys` and `bucket_metadata_sys_of` — process-global
OnceLock state that `serial_test`'s `#[serial]` cannot protect across
nextest's process boundary — and they delete+recreate buckets, the shape that
raced into InsufficientWriteQuorum in backlog#937.
Add them to the `ecstore-serial-flaky` group in both the default and ci
profiles (nextest evaluates a named profile's own overrides list, so the
ci mirror is required). Preventive serialization only, no retries.
Not a full fix for the review comment: `bucket_delete_waits_for_config_
mutation_fence` still proves liveness with a fixed 200ms sleep plus
`assert!(!delete.is_finished())`. Turning that into readiness polling needs
a production-side signal to wait on — asserting "still blocked" is inherently
a negative. Serializing the group removes the parallel-load pressure that
makes the window fragile; the sleep itself is left for a follow-up.
* test(ecstore): pin that a drained bucket is actually deletable
`DeleteBucket`'s emptiness check is `has_xlmeta_files`, a raw scan of the
bucket directory on local disks — not an S3-level listing. So "the client
drained the bucket" and "the bucket is deletable" are two different
contracts, and only the first one was covered.
That gap is what the `S3 Implemented Tests` lane is failing on: 219 cases,
all `BucketNotEmpty` on `nuke_prefixed_buckets`, with every test body
passing. The first one is `test_versioning_obj_suspend_versions`, reported
by pytest as PASSED followed by ERROR at teardown.
Add the missing assertion for the unversioned path: PUT, client DELETE,
then assert no `xl.meta` survives and `DeleteBucket` succeeds. It passes —
which is itself a result: the plain delete path leaves no residue, so the
s3-tests failure is not there.
The versioning-suspended path is the remaining suspect (the client DELETE
leaves a null delete marker, and draining means purging it by
`versionId=null`). It is not covered here: `BucketVersioningSys` resolves
through the ambient `get_bucket_metadata_sys()` OnceLock, which this unit
env cannot set, so the bucket never actually reports as suspended. That
repro belongs at the e2e layer where a real server owns the versioning
state.
* fix(ecstore): let an explicit null-version delete purge its delete marker
Root cause of the `S3 Implemented Tests` lane: 219 cases, all
`BucketNotEmpty` on `nuke_prefixed_buckets`, every test body passing.
On a versioning-suspended bucket a client DELETE leaves a null delete
marker — correct S3 semantics, and an `xl.meta` on disk. Draining the
bucket therefore means purging that marker as `?versionId=null`, which is
what `nuke_bucket` does before `DeleteBucket`. That purge was rejected:
explicit null-version purge of the null delete marker must succeed,
got [Some(MethodNotAllowed)]
so the marker survived, and `DeleteBucket`'s emptiness check — a raw
`has_xlmeta_files` scan of the bucket directory, not an S3 listing — kept
reporting the bucket as non-empty.
The two sides of the version comparison in the batch delete loop are in
different namespaces. `goi.version_id` is the client-facing identity, where
`from_file_info` synthesizes `Some(Uuid::nil())` for a null version on a
versioned *or versioning-suspended* bucket. `version_id` is the storage
identity, where `delete_file_info_version_id` maps an explicit
`?versionId=null` to `None`. Comparing them raw makes the purge look like a
version mismatch, so `explicit_delete_marker` is false and the
`MethodNotAllowed` from the lookup is recorded as a delete failure.
This only became reachable on this branch: previously `check_opts` did not
carry `dobj.version_id`, so `set_disk_delete_creates_delete_marker` was
true, `object_lock_check_required` was false, and the lookup that produces
`MethodNotAllowed` never ran. Adding the version id to `check_opts` lit up
a comparison that was already wrong.
Normalize both sides through `delete_file_info_version_id`.
The regression test injects a real Suspended bucket-config snapshot — the
delete path reads versioned/suspended from that snapshot, not from `opts`,
so without it `from_file_info` never synthesizes the null version id and
the branch is not reached. Mutation-checked: restoring the raw comparison
fails the test with the exact `MethodNotAllowed` above.
* fix(app): drop the now-needless struct update
Reverting `crates/replication` to main removed the extra `MrfReplicateEntry`
fields, so this literal specifies every field again and `..Default::default()`
trips `clippy::needless_update` under `-D warnings`.
Caught by CI, not locally: I had run `cargo check --workspace --all-targets`,
which does not see clippy-only lints. Ran `cargo clippy --workspace
--all-targets -- -D warnings` here — clean.
* test(e2e): assert the fresh-volume classification
four_node_empty_legacy_volumes_start_as_fresh only started the cluster and
listed buckets — no assertion, so any classification path that still permits
startup left it green without proving the pre-created empty `.minio.sys`
directories were treated as fresh volumes.
Pin what that classification actually leaves behind: no buckets adopted into
the namespace, `.rustfs.sys/format.json` written on every drive, and the empty
legacy directory left untouched rather than migrated into.
* fix(bucket): apply the requested Object Lock to existing buckets
Site replication replays make-with-versioning against the destination,
carrying the source's `lockEnabled`. When the destination bucket already
exists it takes `force_create`, and the whole option-application block was
gated on `confirmed_missing` — so the call returned success while the replica
stayed unlocked. Replicated versions could then be deleted without the
retention the source enforces.
Object Lock enable is one-way, so applying it to an existing bucket is safe:
move it out of the creation-only gate, keeping `created` and versioning-only
options creation-scoped as before.
An existing authoritative bucket takes the `cache_bucket_metadata_in` branch,
which only caches, so the enable would have been dropped on restart. Persist
instead when the enable actually changed something.
Mutation-checked: restoring the creation-only gate fails the new
`force_create_enables_object_lock_on_an_existing_bucket` with "Object Lock
must be enabled on the existing bucket".
cargo nextest run -p rustfs-ecstore --lib: 3633 passed.
* fix(ecstore): box the generation-checked config mutation paths too
The earlier stack fix boxed `update` and `delete`, but an authorized
bucket-config mutation carrying an incarnation takes `update_if_incarnation`
/ `delete_if_incarnation` instead — which were still inlining the whole
resolve/load/save chain into an already-deep request future. Same overflow,
sibling path.
* fix(restore): keep the nil-version normalization the strip removed
Reverting the replication subsystem to main took `set_disk/replication.rs`
with it, but one line in that file was this branch's own fix rather than
replication work:
- self.version_id.filter(|v| !v.is_nil()) == fi.version_id.filter(|v| !v.is_nil())
+ self.version_id == fi.version_id
For a versioning-suspended object the expected version is `Some(Uuid::nil())`
while the read-back `FileInfo` carries `None`, so the raw compare reports
every suspended restore as "restored object changed before restore metadata
finalization" and the copy-back never commits. Same nil-vs-None mismatch as
the null delete-marker purge fixed earlier on this branch.
Caught by `Test and Lint (rio-v2)`, not by my local runs: the test lives in
`transition_commit_failure_tests`, gated behind `feature = "test-util"`, so
the 3633-test suite I had been running never included it. Re-ran with
`--features rio-v2,test-util`: 3722 passed.
* feat(replication): purge delete markers by the target's own version id
When a delete marker is replicated, the target assigns it a version id. The
purge that follows derived one from the *source* uuid instead, which is only
correct when the target mirrors source version ids. A generic S3 target does
not: the derived id addresses a version that does not exist there, so the
purge is a no-op and the replica keeps a marker the source has already
removed. Same failure class as #4401.
Record the id the target reports and address it directly on purge.
Data path, all of it driven by the object's internal metadata rather than the
`ReplicationState` wire form, which encodes positionally and cannot carry a
map:
- `rustfs-utils`: the `replication-delete-marker-version-<arn>` key family,
plus `strip_internal_prefix_preserving_case` — ARNs are case-sensitive and
the existing `strip_internal_prefix` lowercases.
- `ReplicationState` gains the map and a `..._corrupt` flag, both
`#[serde(skip)]`; `ReplicatedTargetInfo` carries the per-target id.
- `persist_target_delete_marker_versions` is merge-only. A delete arriving
over internode RPC has an empty map, so treating it as authoritative would
let a remote disk erase an id the local disk still holds.
- `delete_object_version` copies the map into `fi.metadata` before dispatch,
so the durable carrier crosses the wire even though the field does not.
- The keys are folded into the quorum hash through their normalized form:
the dual internal prefixes carrying one mapping share an identity, while a
genuine disagreement between disks still shows up as a quorum difference.
- `corrupt` (the prefixes disagreed) fails closed: skip the purge and warn
rather than guess an id and risk destroying a live version on the target.
Ported from the rc.1 branch, which cannot merge as a whole: its MRF replay
rewrite collides with #5659/#5671/#5672/#5673 and regressed
`MRF_PENDING_CAP`. main's MRF machinery is kept; only this capability moves
across. It touches no MRF code.
Two things did not survive the port, deliberately. The branch's
`missing_is_complete` purge regression does not exist here — it came from its
own HEAD-precheck rewrite, and main's simpler path never had it. And the
branch's `MrfReplicateEntry` ordering fields are MRF-redesign scope, left
behind.
Verification: cargo fmt --all --check, git diff --check,
cargo check --workspace --all-targets, and the suites for the four touched
crates — 4070 tests, 2 pre-existing failures unrelated to this change
(`system_resolver_negative_result_reaches_the_dns_allowlist`,
`test_resolve_domain_preserves_system_resolver_error_provenance`; both are
the sandbox DNS interception, they fail on a clean checkout too).
* fix(replication): keep the layer guard happy
scripts/check_architecture_migration_rules.sh matches on text, so the doc
comments naming `rustfs_filemeta::` read as a cross-layer dependency even
though nothing imports it. Reword them; the guard passes.
* fix(replication): make the target-version cap deterministic
Two defects in this PR, both found in review.
The cap was applied while iterating a `HashMap`, so *which* 1000 entries
survived depended on iteration order. Two disks decoding the same oversized
metadata could keep different subsets, hash differently, and lose quorum —
instead of both reporting the same corruption. Collect first, then truncate
in `BTreeMap` order, which is total and identical everywhere.
And `persist_target_delete_marker_versions` discarded the `corrupt` flag from
the RPC carrier, committing a delete-marker update that looked clean while the
exact remote marker identity was unknown. It now declines to merge a corrupt
carrier. Because the helper only ever inserts, declining leaves the durable
keys already on the object untouched, which is strictly safer than writing a
mapping we cannot trust.
Residual, stated rather than papered over: corruption confined to the RPC
carrier is not persisted as a sentinel, so a later reader of an object that
carried no durable keys still sees "legacy, no mapping" rather than "corrupt".
Persisting that would need a wire-format addition; the consumer already fails
closed on any corruption it can observe.
New test: `target_delete_marker_versions_cap_is_deterministic_across_decodes`
decodes the same 1050-entry map twice and asserts both the corrupt flag and
the retained subset agree.
* fix(replication): preserve multipart source mtime (#5669)
* fix(kms): repair unopenable ciphertext and cover the Vault backends (#5668)
* Add black-box behavior tests for KMS resilience and serialization
* fix(kms): repair unopenable ciphertext across backends
Black-box testing of the KMS crate surfaced several defects that make
encrypted data permanently unreadable.
Symmetric envelopes. The Local and Vault Transit backends returned raw
cipher output from `encrypt` while `decrypt` parsed a JSON envelope, so
anything sealed through the master-key path could never be opened again.
Local also discarded the AES-GCM nonce. Both now emit the same envelope
`decrypt` consumes, matching the Static backend.
Deterministic AAD. The object layer derived AEAD additional data by
serializing a `HashMap` directly. Iteration order differs per instance,
so a context rebuilt from storage produced different AAD bytes than the
one used to seal and the object stopped opening. Ordering by key removes
that dependency, matching the Static backend's existing `context_aad`.
Objects written with the default single-key context are unaffected,
since a one-entry map has only one serialization.
Cipher in the header projection. `metadata_to_headers` recorded the SSE
mode (`AES256` / `aws:kms`), which cannot represent ChaCha20-Poly1305,
so a ChaCha-sealed object came back claiming `aws:kms` and was opened
with the wrong cipher. The cipher now travels in
`x-rustfs-encryption-algorithm` — the header the storage layer already
reads but nothing ever wrote. Objects without it fall back as before.
Also: the Static backend ignored `key_spec` and always issued 256-bit
data keys; Local `list_keys` hardcoded `truncated: false`, ignored
`marker`, and paginated over unordered `read_dir`, so a paginating
client silently saw a partial key list; and Local and Vault KV2 reported
`key_id: "unknown"` from `decrypt` despite the envelope naming the
master key.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(kms): cover both Vault backends and key rotation
The behavior suite ran only against Local and Static, and its own harness
documented the gap: the Vault backends had no business-capability
coverage at all. Setting `RUSTFS_KMS_VAULT_TOKEN` now adds Vault KV2 and
Vault Transit to every `for_each_backend` spec against a live server.
That lane is what surfaced the Transit envelope defect fixed in the
previous commit.
`rotate` and `versioning` are advertised only by the Vault backends, so
until now every capability-gated branch for them took the
`UnsupportedCapability` side and the working half was never asserted — a
rotation that dropped prior key versions would have gone green. The new
`behavior_rotation.rs` pins that half: material sealed before a rotation
still opens after it, repeated rotations accumulate versions rather than
overwriting a single spare, and the history survives a restart.
Two test defects fixed. `objects_round_trip_across_sizes_and_algorithms`
asserted a 1-byte object differs from its own ciphertext, which collides
once every 256 runs; the assertion now applies only where a collision is
not realistic, and small objects stay covered by the tag check and the
decrypt round-trip. `test_from_env_selects_token_file` depended on
`RUSTFS_KMS_VAULT_TOKEN` being absent from the caller's environment and
now clears it explicitly.
The snapshots directory was also removed from `.gitignore`: insta
snapshots are the assertions themselves, so leaving them untracked gives
CI nothing to compare against. Only `.snap.new` scratch files are
ignored now.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(kms): adapt behavior suite to current key APIs
Rebasing onto main brought four API changes the suite predates.
`DeleteKeyRequest` gained `confirm_key_id`, and immediate deletion is now
gated on the server's `allow_immediate_deletion`. Scheduled deletions pass
`None`; the four specs that destroy a key outright echo the key id back
and opt the harness config in, which is what the gate asks of a real
caller.
`LocalBackupExportRequest` gained `sanitized_config`. These specs cover
the key-material path, so they seal no configuration and pass `None`.
`KmsCacheStats` became a named struct with real hit, miss, and eviction
counters. `cache_stats_returns_an_entry_count_and_no_hit_or_miss_data`
existed to pin the old placeholder behavior — that the second tuple
element was always zero — which main has since fixed, so it is now
`cache_stats_reports_hits_and_misses_separately` and asserts the counters
actually move.
Starting the service provisions the reserved probe key, so it shows up in
listings and backup bundles. Exact-set assertions filter it through a new
`without_probe_key` helper rather than naming it, keeping those specs
about the keys they seeded.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(kms): bind the AAD to the stored context bytes
Review caught that canonicalizing the AAD on decrypt breaks objects sealed
before canonicalization existed, and it was right. The AAD is the
*serialization* of the encryption context, and `x-rustfs-encryption-context`
stores that exact byte sequence: `encrypt_object` fed one `HashMap` to the
AEAD and then moved the same map into the metadata the header is written
from, so the stored string is byte-identical to the AAD the object was
sealed under. Those objects are therefore recoverable — but only while
nothing round-trips the value through a `HashMap` and re-serializes it.
Recomputing sorted AAD on decrypt would have turned a readable object into
a permanently unreadable one. The previous behavior was worse than the
first analysis credited: it did not merely fail intermittently, it made
the failure deterministic.
`EncryptionMetadata` now carries `context_aad`, the bytes the object was
actually sealed with. Encryption records what it fed the AEAD, the header
projection stores those bytes verbatim (and preserves a legacy ordering
across a re-projection rather than rewriting it into sorted form), and
`headers_to_metadata` carries the stored string through untouched. Both
decrypt paths, SSE-KMS and SSE-C, prefer it and fall back to canonical
serialization only when no stored serialization exists. Canonicalization
still applies to everything newly sealed, so the original ordering bug
cannot recur.
Two tests pin this: a legacy record whose sealed bytes are non-canonical
must survive a full header round trip unchanged, and a context header
rewritten to an equivalent-but-reordered serialization must fail
authentication rather than silently re-deriving a working AAD. Both were
mutation-checked against the reinstated bug on each side.
Also from review: the lifecycle churn test asserted only that every
request was accounted for, which holds whether the state gate exists or
not, so both branches are now pinned deterministically after the churn
(asserting `refused > 0` on the concurrent phase would only trade the hole
for a scheduling flake). And the Local and Vault KV2 envelopes compare
`encryption_context` without authenticating it — `DekCrypto` seals only
the plaintext — which is now documented at both sites; closing it needs a
versioned envelope, since existing ciphertext was sealed without AAD.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: ccccpj <ccccpj@outlook.com>
Co-authored-by: 唐小鸭 <tangtang1251@qq.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Preserve metadata replication operations in the durable MRF and route tagging, retention, and legal-hold updates through the existing full-object replication transport. Keep ACL propagation outside the contract because the current object model has no durable object ACL state.
Refs #1616
Add RAII guards for replication runtime backlog tickets so active worker and queue counters unwind on every terminal path.
Expose node-local MRF pending, dropped, missed, and flush-failure metrics through the bucket replication Prometheus collector while keeping the existing current backlog and durable MRF gauges additive.
Update durable MRF summary maintenance to aggregate incrementally during the persister loop, avoiding repeated full-entry scans on each successful flush.
Co-authored-by: heihutu <heihutu@gmail.com>
Co-authored-by: zhi22915 <qiuzgang@gmail.com>
Only replication-target writes took the bucket transaction lock. Every other
config write (policy, tagging, lifecycle, versioning, ...) went straight to
the process-local metadata-system guard, which serializes nothing across
nodes.
Each config write is a read-modify-write of one whole BucketMetadata blob:
load the blob, replace one field, save the blob back. The namespace locks
inside read_config/save_config are taken and released separately, so they do
not span that cycle. Two nodes updating different config files of the same
bucket therefore both load the same blob, each set their own field, and the
later save drops the other's -- with both clients already told 2xx. This is
not last-writer-wins on one document; an orthogonal config silently vanishes.
Route update(), delete() and update_config_with() through
acquire_config_write_guards(), which takes the cluster-wide transaction lock
first and the metadata-system write guard second. That order is load-bearing:
taking the process-local guard first would park every local reader and writer
of every bucket behind a lock whose holder may be another node, turning
remote contention into a local stall.
The lock is per bucket rather than per config file, since a per-file key
would let exactly the offending pair run concurrently. Rename the helper to
acquire_bucket_metadata_transaction_lock to match, but deliberately keep the
lock resource string as "bucket-targets/{bucket}/transaction.lock": the key
is what nodes agree on, so renaming it would leave a mixed-version cluster
with two disjoint keys and stop old and new nodes from excluding each other
on the very writes that are serialized today.
update_config_with() already narrowed its staleness window to a single load
and save, but its exclusion was explicitly process-local; it is now
cluster-wide, so its doc comment no longer disclaims cross-node races.
Also make update_and_parse load through self.api instead of the ambient store
handle, so the read and the write of one read-modify-write cannot resolve to
different instances.
The new tests drive two BucketMetadataSys instances over one ECStore -- the
in-process stand-in for two nodes, since they share no RwLock and can only be
serialized by the namespace lock. Verified the lost-update test has teeth by
removing the lock and confirming it fails on round 0, with the tagging config
clobbered to empty by the concurrent policy write.
The LoadBucketMetadata peer-notification handler loaded bucket metadata
with the fabricating loader (ConfigNotFound -> BucketMetadata::new) and
unconditionally cached the result. On a transient read-quorum dip during
a reload notification, a peer cached an authoritative "no Object Lock"
default for a lock-enabled bucket, disabling the batch-delete retention
gate (object_lock_delete_check_required) on that node until the next
refresh, and wiping its bucket-target/durability sync state.
Production changes:
- New BucketMetadataSys::reload_from_store (metadata_sys::
reload_bucket_metadata): the peer reload path uses the presence-aware
loader and installs only metadata actually read from persisted
storage. A load miss returns an error (surfaced to the notifying peer
as success=false) and leaves the cache untouched; deletion still
propagates only through the dedicated DeleteBucketMetadata
notification.
- The reload runs under the outer metadata-sys write guard, load
included, mirroring update(): every other cache installer holds that
lock, so a reload snapshot can never land after - and roll back - a
newer concurrent install (the stale-load lost-update from the
review), and the install-plus-registry-sync sequence stays atomic
against concurrent removes and reloads (previously only the set call
was write-guarded, with the load outside any lock).
- The peer-visible miss error is a fixed string: the notifying peer
substring-matches error text against network-failure needles
(is_network_like_error), so interpolating a bucket name (e.g. a legal
bucket literally named "unavailable") could mark a healthy peer
offline.
- get_config's lazy insert routes through set(), picking up the
negative-cache invalidation.
An earlier draft instead guarded set() with a per-config updated_at
freshness comparison. Adversarial validation rejected it (three roles
independently): update_config stamps with the handling node's wall
clock, so within the skew the cluster already tolerates (+/-300s RPC
auth window) a config rewritten with an earlier stamp - e.g. revoking a
public-read policy through a second node, or any same-field rewrite
after an NTP step-back - would be skipped by every peer forever,
silently pinning the revoked permissive config with no re-convergence
path (the 15-minute refresh also routed through the guard). Race
staleness is second-scale while skew is minute-scale, so no tolerance
bound can separate them; the write-guard serialization closes the same
race without clocks and preserves the refresh loop's unconditional
converge-to-disk property, which is the cluster's self-healing
mechanism.
Startup audit (BucketMetadataSys::init): concurrent_load's
insert-if-vacant still installs a fabricated default when a transient
miss hits at boot - indistinguishable from a legacy bucket without a
metadata file at this layer - bounded by the next successful persisted
load. Making the object-lock gate fail closed on such entries is filed
as a follow-up, alongside the bare "unavailable" needle in
is_network_like_error and the Swift cache-only metadata writes.
Tests:
- bucket::metadata_sys::tests::
peer_reload_never_caches_fabricated_defaults_as_authoritative:
miss installs nothing / miss keeps the existing entry intact
(asserting the dedicated non-persisted error) / persisted reload
converges the cache over a stale entry.
- node_service::tests::
test_load_bucket_metadata_failure_skips_scanner_maintenance:
a failed reload reports failure and does not advance scanner
maintenance activity (previously recorded even on a miss).
- The handler success path stays uncovered at the RPC layer (needs an
isolated global object layer, like the pre-existing ignored test);
the composition is pinned at the sys level instead.
Verification:
- cargo fmt --check and cargo clippy --lib --tests clean on
rustfs-ecstore and rustfs.
- Targeted suites green; full cargo test -p rustfs-ecstore --lib:
3198/3200 with two parallelism-sensitive lock-test flakes from the
known baseline (pass in isolation; a different pair flakes per run).
- Adversarial validation (high-risk tier, all seven roles as
independent parallel reviewers) run per AGENTS.md; all findings
fixed or rebutted with evidence, three out-of-scope findings filed
as follow-up tasks.
Co-authored-by: houseme <housemecn@gmail.com>
Swift container and account metadata handlers cloned the cached
BucketMetadata, set the tagging fields, and called set_bucket_metadata,
which only updates the in-memory cache map. Nothing reached
.metadata.bin, so every Swift metadata POST was lost on restart and
silently overwritten by the next disk-truth reload (a peer
LoadBucketMetadata notification or the 15-minute refresh loop) — while
the client had already been told 2xx.
Route these writes through a new metadata_sys::update_config_with: a
read-modify-write that loads the on-disk metadata and persists the
result under the same write guard metadata_sys::update uses, so the
rewrite merges against disk truth instead of a possibly stale cache and
cannot clobber a concurrent update to another config file. Peers are
notified afterwards, matching the S3 config handlers.
Persisting these writes required hardening the paths that now produce
durable state:
- Account metadata writes validate account ownership. This metadata
holds the account's TempURL signing key, so an unauthenticated write
for someone else's account would have become a durable, cluster-wide
takeover of that account's pre-signed URLs. Reads stay open because
TempURL signature validation runs before credentials exist.
- disable_versioning verifies the container exists. Without it the
metadata loader's "no metadata on disk" default would be persisted,
creating an orphan metadata file and caching a fabricated default as
authoritative.
- Container and account metadata are size- and count-limited, reusing
the Swift limits object metadata already enforces; these tags land in
the bucket metadata file that every later config write rewrites whole.
- A rewrite refuses to run when the persisted tagging config is
unreadable, instead of merging onto an empty set and wiping the
container ACL and versioning tags. It reports 409 naming the remedy.
- Storage errors are logged in full and reported generically, since they
now carry real disk and quorum detail.
The tagging arm of BucketMetadata::update_config also clears the parsed
config, as the lifecycle arm does: parse_all_configs skips empty XML
rather than clearing, so a cleared config kept serving the old tags.
Tagging is serialized with the S3 XML serializer the loader can parse
back, not quick_xml, whose output was never round-trippable.
cxymds