ci: pin s3tests Python tools

.github/workflows/e2e-s3tests.yml

@@ -178,9 +178,14 @@ jobs:
 
       - name: Install Python tools
         run: |
-          python3 -m pip install --user --upgrade pip awscurl tox
+          python3 -m pip install --user --upgrade pip "awscurl==0.44" "tox==4.60.0"
           echo "$HOME/.local/bin" >> "$GITHUB_PATH"
 
+      - name: Verify Python tools
+        run: |
+          test "$(python3 -c 'import importlib.metadata as m; print(m.version("awscurl"))')" = "0.44"
+          test "$(python3 -c 'import importlib.metadata as m; print(m.version("tox"))')" = "4.60.0"
+
       - name: Enable buildx
         uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f # v3
 

.github/workflows/nightly-gnu.yml

@@ -39,10 +39,11 @@ jobs:
     env:
       FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
     steps:
-      - name: Checkout repository
+      - name: Checkout main branch
         uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
         with:
           persist-credentials: false
+          ref: main
 
       - name: Setup Rust environment
         uses: ./.github/actions/setup
@@ -88,10 +89,11 @@ jobs:
       # either casing.
       NO_PROXY: 127.0.0.1,localhost
     steps:
-      - name: Checkout repository
+      - name: Checkout main branch
         uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
         with:
           persist-credentials: false
+          ref: main
 
       - name: Setup Rust environment
         uses: ./.github/actions/setup
@@ -176,10 +178,11 @@ jobs:
       FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
       NO_PROXY: 127.0.0.1,localhost
     steps:
-      - name: Checkout repository
+      - name: Checkout main branch
         uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
         with:
           persist-credentials: false
+          ref: main
 
       - name: Setup Rust environment
         uses: ./.github/actions/setup

CLAUDE.md

@@ -30,8 +30,7 @@ make build-docker BUILD_OS=ubuntu22.04
 - 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)
+- CI gates: `.github/workflows/ci.yml` (source of truth; never copy its steps into docs)
 - 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):

CONTRIBUTING.md

@@ -70,8 +70,6 @@ make pre-pr
 
 > For the full test-layer taxonomy (unit / ecstore black-box / e2e / s3s-e2e / S3 compatibility / chaos / fuzz / bench), each layer's entry command, the naming conventions the migration gate depends on, and the serial/nextest rules, see [docs/testing/README.md](docs/testing/README.md).
 
-> For the event, timeout, required-status, and local reproduction matrix, see [docs/testing/ci-gates.md](docs/testing/ci-gates.md).
-
 ### 🔒 Automated Pre-commit Hooks
 #### What `make pre-commit` and `make pre-pr` actually run
 

Cargo.lock

@@ -1858,9 +1858,9 @@ dependencies = [
 
 [[package]]
 name = "cc"
-version = "1.4.4"
+version = "1.4.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0ad534f4357a5264cce5019c989cf66a4f0dc4e0d1b1d15f8aacec0ff7360273"
+checksum = "509591b7bcd67f4ef775afad7662703b4935daaa6ec0e5605cfb1090b32a2b6d"
 dependencies = [
  "find-msvc-tools",
  "jobserver",
@@ -2522,6 +2522,12 @@ dependencies = [
  "subtle",
 ]
 
+[[package]]
+name = "cty"
+version = "0.2.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b365fabc795046672053e29c954733ec3b05e4be654ab130fe8f1f94d7051f35"
+
 [[package]]
 name = "curve25519-dalek"
 version = "4.1.3"
@@ -5982,6 +5988,15 @@ version = "0.2.16"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b6d2cec3eae94f9f509c767b45932f1ada8350c4bdb85af2fcab4a3c14807981"
 
+[[package]]
+name = "libmimalloc-sys"
+version = "0.1.49"
+source = "git+https://github.com/xonatius/mimalloc_rust.git?rev=6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11#6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11"
+dependencies = [
+ "cc",
+ "cty",
+]
+
 [[package]]
 name = "libredox"
 version = "0.1.20"
@@ -6382,6 +6397,14 @@ dependencies = [
  "synstructure 0.13.2",
 ]
 
+[[package]]
+name = "mimalloc"
+version = "0.1.52"
+source = "git+https://github.com/xonatius/mimalloc_rust.git?rev=6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11#6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11"
+dependencies = [
+ "libmimalloc-sys",
+]
+
 [[package]]
 name = "mime"
 version = "0.3.17"
@@ -9139,11 +9162,13 @@ dependencies = [
  "insta",
  "jiff",
  "libc",
+ "libmimalloc-sys",
  "libsystemd",
  "matchit 0.9.2",
  "md-5 0.11.0",
  "metrics",
  "metrics-util",
+ "mimalloc",
  "mime_guess",
  "opentelemetry",
  "opentelemetry_sdk",
@@ -9179,8 +9204,6 @@ dependencies = [
  "rustfs-lock",
  "rustfs-log-analyzer",
  "rustfs-madmin",
- "rustfs-mimalloc",
- "rustfs-mimalloc-sys",
  "rustfs-notify",
  "rustfs-object-capacity",
  "rustfs-object-data-cache",
@@ -9852,24 +9875,6 @@ dependencies = [
  "tokio",
 ]
 
-[[package]]
-name = "rustfs-mimalloc"
-version = "0.5.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a406f4aa07084301d485beec873af6dccc8e3f8762da244743df92038b1db1a6"
-dependencies = [
- "rustfs-mimalloc-sys",
-]
-
-[[package]]
-name = "rustfs-mimalloc-sys"
-version = "0.5.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c3051b819175f58445d4c369a72f0ab88149f3885ba8bea2aff3be01f53fe7cd"
-dependencies = [
- "cc",
-]
-
 [[package]]
 name = "rustfs-notify"
 version = "1.0.0-rc.3"
@@ -12683,7 +12688,6 @@ dependencies = [
  "js-sys",
  "rand 0.10.2",
  "serde_core",
- "sha1_smol",
  "wasm-bindgen",
 ]
 

Cargo.toml

@@ -350,8 +350,8 @@ russh-sftp = "2.4.0"
 dav-server = "0.11.0"
 
 # Performance Analysis and Memory Profiling
-rustfs-mimalloc = { version = "0.5.0" }
-rustfs-mimalloc-sys = { version = "0.5.0" }
+mimalloc = { version = "0.1.52", git = "https://github.com/xonatius/mimalloc_rust.git", rev = "6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11" }
+libmimalloc-sys = { version = "0.1.49", git = "https://github.com/xonatius/mimalloc_rust.git", rev = "6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11", features = ["extended"] }
 hotpath = { version = "0.23.3", default-features = false }
 # Snapshot testing for output format regression detection
 insta = { version = "1.48" }

crates/ecstore/src/core/pools.rs

@@ -13,8 +13,6 @@
 // limitations under the License.
 
 use crate::bucket::replication::replication_state_from_filemeta;
-#[cfg(test)]
-use crate::bucket::utils::is_meta_bucketname;
 use crate::bucket::versioning_sys::BucketVersioningSys;
 use crate::bucket::{
     lifecycle::{
@@ -1349,53 +1347,6 @@ fn should_cleanup_decommission_source_entry(decommissioned: usize, total_version
     decommissioned.saturating_add(expired) == total_versions
 }
 
-const DECOMMISSION_FREE_VERSION_MIGRATED_REASON: &str = "tier_free_version_migrated";
-const DECOMMISSION_FREE_VERSION_CONSUMED_REASON: &str = "tier_free_version_already_consumed";
-const DECOMMISSION_FREE_VERSION_RETAINED_REASON: &str = "tier_free_version_migration_failed";
-const DECOMMISSION_FREE_VERSION_SWEEP_REASON: &str = "tier_free_version_unresolved_after_decommission";
-const DECOMMISSION_FREE_VERSION_DISPOSITION_REASON: &str = "tier_free_version_disposition_recorded";
-
-#[derive(Debug, Default, Clone, Copy, PartialEq, Eq)]
-struct DecommissionFreeVersionDisposition {
-    migrated: usize,
-    consumed: usize,
-    retained: usize,
-}
-
-impl DecommissionFreeVersionDisposition {
-    fn record_migrated(&mut self) {
-        self.migrated += 1;
-    }
-
-    fn record_consumed(&mut self) {
-        self.consumed += 1;
-    }
-
-    fn record_retained(&mut self) {
-        self.retained += 1;
-    }
-
-    fn total(self) -> usize {
-        self.migrated.saturating_add(self.consumed).saturating_add(self.retained)
-    }
-}
-
-enum DecommissionFreeVersionAttempt {
-    Migrated,
-    Consumed,
-    CapacityFailure(Error),
-    Retry(Error),
-}
-
-fn classify_decommission_free_version_attempt(result: Result<()>) -> DecommissionFreeVersionAttempt {
-    match result {
-        Ok(()) => DecommissionFreeVersionAttempt::Migrated,
-        Err(err) if is_decommission_copy_cleanup_safe_error(&err) => DecommissionFreeVersionAttempt::Consumed,
-        Err(err) if is_decommission_target_capacity_error(&err) => DecommissionFreeVersionAttempt::CapacityFailure(err),
-        Err(err) => DecommissionFreeVersionAttempt::Retry(err),
-    }
-}
-
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 #[allow(
     dead_code,
@@ -2075,7 +2026,7 @@ impl PoolMeta {
         self.load_no_lock(pool).await
     }
 
-    pub(crate) async fn load_no_lock<S>(&mut self, pool: Arc<S>) -> Result<()>
+    async fn load_no_lock<S>(&mut self, pool: Arc<S>) -> Result<()>
     where
         S: EcstoreObjectIO,
     {
@@ -2664,12 +2615,8 @@ fn determine_decommission_final_state(items_failed: usize, was_cancelled: bool)
     }
 }
 
-fn decommission_remaining_version_count(versions: &[rustfs_filemeta::FileInfo], expired: usize) -> usize {
-    versions
-        .iter()
-        .filter(|version| !version.tier_free_version())
-        .count()
-        .saturating_sub(expired)
+fn decommission_remaining_version_count(total_versions: usize, expired: usize) -> usize {
+    total_versions.saturating_sub(expired)
 }
 
 fn should_skip_decommission_delete_marker(
@@ -2732,7 +2679,6 @@ fn decommission_remote_tiered_opts(
         user_defined: version.metadata.clone(),
         src_pool_idx,
         data_movement: true,
-        incl_free_versions: version.tier_free_version(),
         include_part_checksums: true,
         http_preconditions: Some(crate::data_movement::data_movement_target_precondition()),
         expected_bucket_incarnation_id,
@@ -3876,7 +3822,6 @@ impl ECStore {
 
         let mut decommissioned: usize = 0;
         let mut expired: usize = 0;
-        let mut free_version_disposition = DecommissionFreeVersionDisposition::default();
         let mut cleanup_preflight_allowed_missing = Vec::new();
 
         for version in fivs.versions.iter() {
@@ -3885,115 +3830,6 @@ impl ECStore {
             }
             decommission_cancel_signal_result(rx.is_cancelled())?;
 
-            if version.tier_free_version() {
-                let version_id = version.version_id.map(|v| v.to_string());
-                let mut migration_error = None;
-                let mut migrated = false;
-                let mut consumed = false;
-                let mut capacity_failure = false;
-                for _ in 0..3 {
-                    match classify_decommission_free_version_attempt(
-                        run_decommission_side_effect(&rx, &operation_gate, || async {
-                            self.decommission_tiered_object(
-                                bucket.as_str(),
-                                &version.name,
-                                version,
-                                &decommission_remote_tiered_opts(
-                                    version,
-                                    version_id.clone(),
-                                    idx,
-                                    expected_bucket_incarnation_id,
-                                ),
-                            )
-                            .await
-                        })
-                        .await,
-                    ) {
-                        DecommissionFreeVersionAttempt::Migrated => {
-                            migrated = true;
-                            migration_error = None;
-                            break;
-                        }
-                        DecommissionFreeVersionAttempt::Consumed => {
-                            consumed = true;
-                            migration_error = None;
-                            break;
-                        }
-                        DecommissionFreeVersionAttempt::CapacityFailure(err) => {
-                            capacity_failure = true;
-                            migration_error = Some(err);
-                            break;
-                        }
-                        DecommissionFreeVersionAttempt::Retry(err) => migration_error = Some(err),
-                    }
-                }
-
-                {
-                    let mut pool_meta = self.pool_meta.write().await;
-                    ensure_decommission_generation(&pool_meta, idx, generation)?;
-                    if let Err(err) = count_decommission_item(&mut pool_meta, idx, 0, !migrated && !consumed) {
-                        return Err(with_decommission_entry_context(
-                            "count_decommission_item",
-                            bucket.as_str(),
-                            entry.name.as_str(),
-                            err,
-                        ));
-                    }
-                }
-
-                if migrated || consumed {
-                    decommissioned += 1;
-                    cleanup_preflight_allowed_missing.push(data_movement::source_cleanup_version_identity(version));
-                }
-                if migrated {
-                    free_version_disposition.record_migrated();
-                } else if consumed {
-                    free_version_disposition.record_consumed();
-                } else {
-                    free_version_disposition.record_retained();
-                }
-
-                debug!(
-                    event = EVENT_DECOMMISSION_ENTRY,
-                    component = LOG_COMPONENT_ECSTORE,
-                    subsystem = LOG_SUBSYSTEM_POOLS,
-                    pool_index = idx,
-                    bucket = %bucket,
-                    object = %version.name,
-                    version_id = ?version_id,
-                    result = ?migration_error,
-                    reason = if migrated {
-                        DECOMMISSION_FREE_VERSION_MIGRATED_REASON
-                    } else if consumed {
-                        DECOMMISSION_FREE_VERSION_CONSUMED_REASON
-                    } else {
-                        DECOMMISSION_FREE_VERSION_RETAINED_REASON
-                    },
-                    state = if migrated {
-                        "free_version_migrated"
-                    } else if consumed {
-                        "free_version_consumed"
-                    } else {
-                        "free_version_retained"
-                    },
-                    "Decommission free-version disposition recorded"
-                );
-
-                if capacity_failure {
-                    return Err(with_decommission_entry_context(
-                        "decommission_tier_free_version",
-                        bucket.as_str(),
-                        version.name.as_str(),
-                        migration_error.expect("capacity failure must retain its error"),
-                    ));
-                }
-
-                if !migrated && !consumed {
-                    break;
-                }
-                continue;
-            }
-
             if run_decommission_side_effect(&rx, &operation_gate, || async {
                 should_skip_lifecycle_for_data_movement(
                     self.clone(),
@@ -4014,7 +3850,7 @@ impl ECStore {
                 continue;
             }
 
-            let remaining_versions = decommission_remaining_version_count(&fivs.versions, expired);
+            let remaining_versions = decommission_remaining_version_count(fivs.versions.len(), expired);
             if should_skip_decommission_delete_marker(version, remaining_versions, replication_config.is_some()) {
                 //
                 decommissioned += 1;
@@ -4293,24 +4129,6 @@ impl ECStore {
             }
         }
 
-        if free_version_disposition.total() > 0 {
-            debug!(
-                event = EVENT_DECOMMISSION_ENTRY,
-                component = LOG_COMPONENT_ECSTORE,
-                subsystem = LOG_SUBSYSTEM_POOLS,
-                pool_index = idx,
-                bucket = %bucket,
-                object = %entry.name,
-                free_versions_migrated = free_version_disposition.migrated,
-                free_versions_consumed = free_version_disposition.consumed,
-                free_versions_retained = free_version_disposition.retained,
-                free_versions_total = free_version_disposition.total(),
-                reason = DECOMMISSION_FREE_VERSION_DISPOSITION_REASON,
-                state = "free_version_disposition",
-                "Decommission free-version disposition summary"
-            );
-        }
-
         if should_cleanup_decommission_source_entry(decommissioned, fivs.versions.len(), expired) {
             if bucket_incarnation_fence.as_ref().is_some_and(|guard| guard.is_lock_lost()) {
                 return Err(Error::other("decommission bucket incarnation fence was lost before source cleanup"));
@@ -4480,39 +4298,6 @@ impl ECStore {
         .await
     }
 
-    #[cfg(test)]
-    pub(crate) async fn decommission_entry_for_test_with_bucket_incarnation(
-        self: &Arc<Self>,
-        idx: usize,
-        entry: MetaCacheEntry,
-        bucket: String,
-        set: Arc<SetDisks>,
-    ) -> Result<()> {
-        let expected_bucket_incarnation_id = if is_meta_bucketname(&bucket) {
-            None
-        } else {
-            Some(self.bucket_incarnation_id_from_disk(&bucket).await?)
-        };
-        self.decommission_entry(
-            CancellationToken::new(),
-            idx,
-            OffsetDateTime::now_utc(),
-            entry,
-            bucket,
-            set,
-            None,
-            None,
-            None,
-            expected_bucket_incarnation_id,
-        )
-        .await
-    }
-
-    #[cfg(test)]
-    pub(crate) async fn check_after_decommission_for_test(self: &Arc<Self>, idx: usize) -> Result<()> {
-        self.check_after_decommission(idx).await
-    }
-
     #[tracing::instrument(skip(self, rx))]
     async fn decommission_pool(
         self: &Arc<Self>,
@@ -5377,7 +5162,6 @@ impl ECStore {
                 let lifecycle_config_cb = lifecycle_config.clone();
                 let object_lock_config_cb = object_lock_config.clone();
                 let store = Arc::clone(self);
-                let set_cb = Arc::clone(set);
                 let callback_rx_cb = callback_rx.clone();
 
                 let callback: ListCallback = Arc::new(move |entry: MetaCacheEntry| {
@@ -5387,7 +5171,6 @@ impl ECStore {
                     let lifecycle_config = lifecycle_config_cb.clone();
                     let object_lock_config = object_lock_config_cb.clone();
                     let store = Arc::clone(&store);
-                    let set = Arc::clone(&set_cb);
                     let callback_rx = callback_rx_cb.clone();
                     Box::pin(async move {
                         if callback_rx.is_cancelled() {
@@ -5402,14 +5185,11 @@ impl ECStore {
                             return;
                         }
 
-                        let fivs = match load_decommission_entry_exact_versions(
-                            &set,
+                        let fivs = match load_decommission_entry_versions(
                             &entry,
                             &bucket_name,
                             "check_after_decommission.file_info_versions",
-                        )
-                        .await
-                        {
+                        ) {
                             Ok(fivs) => fivs,
                             Err(err) => {
                                 let mut first_err = entry_error.lock().await;
@@ -5422,23 +5202,7 @@ impl ECStore {
                         };
 
                         let mut remaining = 0;
-                        for version in fivs.versions.iter().chain(fivs.free_versions.iter()) {
-                            if version.tier_free_version() {
-                                remaining += 1;
-                                debug!(
-                                    event = EVENT_DECOMMISSION_ENTRY,
-                                    component = LOG_COMPONENT_ECSTORE,
-                                    subsystem = LOG_SUBSYSTEM_POOLS,
-                                    pool_index = idx,
-                                    bucket = %bucket_name,
-                                    object = %entry.name,
-                                    version_id = ?version.version_id,
-                                    reason = DECOMMISSION_FREE_VERSION_SWEEP_REASON,
-                                    state = "free_version_retained",
-                                    "Decommission final sweep retained a free version"
-                                );
-                                continue;
-                            }
+                        for version in &fivs.versions {
                             if version.deleted {
                                 continue;
                             }
@@ -5553,6 +5317,13 @@ mod tests {
         assert_eq!(determine_decommission_final_state(0, true), DecommissionFinalState::Failed);
     }
 
+    #[test]
+    fn decommission_remaining_version_count_excludes_only_expired_versions() {
+        assert_eq!(decommission_remaining_version_count(1, 0), 1);
+        assert_eq!(decommission_remaining_version_count(2, 1), 1);
+        assert_eq!(decommission_remaining_version_count(1, 1), 0);
+    }
+
     #[test]
     fn lifecycle_action_removes_data_movement_version_rejects_delete_marker_action() {
         assert!(!lifecycle_action_removes_data_movement_version(IlmAction::DeleteAction));
@@ -5609,21 +5380,6 @@ mod tests {
         )));
     }
 
-    #[test]
-    fn decommission_free_version_attempt_treats_missing_source_as_consumed() {
-        let attempt =
-            classify_decommission_free_version_attempt(Err(Error::ObjectNotFound("bucket".to_string(), "object".to_string())));
-
-        assert!(matches!(attempt, DecommissionFreeVersionAttempt::Consumed));
-    }
-
-    #[test]
-    fn decommission_free_version_attempt_preserves_capacity_failure() {
-        let attempt = classify_decommission_free_version_attempt(Err(Error::DiskFull));
-
-        assert!(matches!(attempt, DecommissionFreeVersionAttempt::CapacityFailure(Error::DiskFull)));
-    }
-
     #[test]
     fn decommission_delete_marker_copy_error_rejects_data_movement_overwrite() {
         let err = Error::DataMovementOverwriteErr("bucket".to_string(), "object".to_string(), "version".to_string());
@@ -5805,12 +5561,6 @@ mod tests {
         assert!(opts.include_part_checksums);
         assert!(opts.http_preconditions.is_some());
         assert_eq!(opts.expected_bucket_incarnation_id, Some(incarnation));
-        assert!(!opts.incl_free_versions);
-
-        let mut free_version = version;
-        free_version.set_tier_free_version();
-        let free_opts = decommission_remote_tiered_opts(&free_version, Some("free-version-id".to_string()), 9, Some(incarnation));
-        assert!(free_opts.incl_free_versions);
     }
 
     #[test]

crates/ecstore/src/core/sets.rs

@@ -988,11 +988,14 @@ impl crate::storage_api_contracts::multipart::MultipartOperations for Sets {
     }
 }
 
-impl Sets {
-    pub(crate) async fn heal_format_with_fence<F>(&self, dry_run: bool, fence_lost: F) -> Result<(HealResultItem, Option<Error>)>
-    where
-        F: Fn() -> bool + Send + Sync,
-    {
+#[async_trait::async_trait]
+impl crate::storage_api_contracts::heal::HealOperations for Sets {
+    type Error = Error;
+    type HealResultItem = HealResultItem;
+    type HealOptions = HealOpts;
+
+    #[tracing::instrument(skip(self))]
+    async fn heal_format(&self, dry_run: bool) -> Result<(HealResultItem, Option<Error>)> {
         let (disks, init_errs) = init_storage_disks_with_errors(
             &self.endpoints.endpoints,
             &DiskOption {
@@ -1065,9 +1068,6 @@ impl Sets {
             // Save new formats `format.json` on unformatted disks.
             for (index, (fm, disk)) in tmp_new_formats.iter_mut().zip(disks.iter()).enumerate() {
                 if fm.is_some() && disk.is_some() {
-                    if fence_lost() {
-                        return Ok((res, Some(StorageError::SlowDown)));
-                    }
                     if let Err(err) = save_format_file(disk, fm).await {
                         if let Some(disk) = disk.as_ref() {
                             let _ = disk.close().await;
@@ -1101,18 +1101,6 @@ impl Sets {
         }
         Ok((res, None))
     }
-}
-
-#[async_trait::async_trait]
-impl crate::storage_api_contracts::heal::HealOperations for Sets {
-    type Error = Error;
-    type HealResultItem = HealResultItem;
-    type HealOptions = HealOpts;
-
-    #[tracing::instrument(skip(self))]
-    async fn heal_format(&self, dry_run: bool) -> Result<(HealResultItem, Option<Error>)> {
-        self.heal_format_with_fence(dry_run, || false).await
-    }
     #[tracing::instrument(skip(self))]
     async fn heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem> {
         let mut result = HealResultItem {

crates/ecstore/src/data_movement/mod.rs

@@ -1950,19 +1950,6 @@ mod tests {
         assert!(source_cleanup_versions_match_with_allowed_missing(&expected, &current, &allowed_missing));
     }
 
-    #[test]
-    fn test_decommission_cleanup_preflight_accepts_migrated_free_version_consumed_from_source() {
-        let migrated = cleanup_test_file_info("object.txt", Uuid::from_u128(1), "migrated");
-        let mut free_version = cleanup_test_file_info("object.txt", Uuid::from_u128(2), "tier-cleanup");
-        free_version.deleted = true;
-        free_version.set_tier_free_version();
-        let expected = cleanup_test_versions(vec![migrated.clone(), free_version.clone()]);
-        let current = cleanup_test_versions(vec![migrated]);
-        let allowed_missing = vec![source_cleanup_version_identity(&free_version)];
-
-        assert!(source_cleanup_versions_match_with_allowed_missing(&expected, &current, &allowed_missing));
-    }
-
     #[test]
     fn test_decommission_cleanup_preflight_rejects_unexpected_missing_version() {
         let migrated = cleanup_test_file_info("object.txt", Uuid::from_u128(1), "migrated");

crates/ecstore/src/io_support/bitrot.rs

@@ -784,24 +784,6 @@ pub(crate) fn create_deferred_bitrot_reader_with_stripe_handle(
 ///
 /// # Returns
 /// A Result containing the BitrotWriterWrapper or an error
-/// Size hint handed to `DiskAPI::create_file` for a bitrot-wrapped shard.
-///
-/// A known length is grown by one checksum per shard so the on-disk file size
-/// matches what the bitrot writer emits. A negative length is the
-/// unknown-size sentinel (`HashReader::SIZE_PRESERVE_LAYER`, used by SSE and
-/// compression) and must be preserved: `RemoteDisk::create_file` forwards it
-/// in the `put_file_stream` query, and the receiver only treats `size > 0` as
-/// a fixed body length when locating the authenticated trailer. Clamping it
-/// to `0` would claim an empty body and misframe the stream. `0` stays `0`
-/// because a genuinely empty object still means an empty body.
-fn bitrot_create_file_size(length: i64, shard_size: usize, checksum_algo: &HashAlgorithm) -> i64 {
-    if length <= 0 {
-        return length;
-    }
-    let length = length as usize;
-    (length.div_ceil(shard_size) * checksum_algo.size() + length) as i64
-}
-
 pub async fn create_bitrot_writer(
     is_inline_buffer: bool,
     disk: Option<&DiskStore>,
@@ -814,7 +796,12 @@ pub async fn create_bitrot_writer(
     let writer = if is_inline_buffer {
         CustomWriter::new_inline_buffer()
     } else if let Some(disk) = disk {
-        let length = bitrot_create_file_size(length, shard_size, &checksum_algo);
+        let length = if length > 0 {
+            let length = length as usize;
+            (length.div_ceil(shard_size) * checksum_algo.size() + length) as i64
+        } else {
+            0
+        };
 
         let file = disk.create_file("", volume, path, length).await?;
         #[cfg(feature = "hotpath")]
@@ -833,25 +820,6 @@ mod tests {
     use rustfs_rio::ChunkReader;
     use std::collections::VecDeque;
 
-    #[test]
-    fn bitrot_create_file_size_grows_known_length_by_checksums() {
-        // 10 bytes over 4-byte shards = 3 shards, each followed by a 32-byte hash.
-        assert_eq!(bitrot_create_file_size(10, 4, &HashAlgorithm::HighwayHash256), 10 + 3 * 32);
-        assert_eq!(bitrot_create_file_size(10, 4, &HashAlgorithm::None), 10);
-    }
-
-    #[test]
-    fn bitrot_create_file_size_keeps_empty_and_unknown_distinct() {
-        assert_eq!(bitrot_create_file_size(0, 4, &HashAlgorithm::HighwayHash256), 0);
-        // SSE/compression streams advertise SIZE_PRESERVE_LAYER (-1); the remote
-        // put_file_stream receiver relies on a non-positive size to parse the auth
-        // trailer from the stream tail, so the sentinel must survive untouched.
-        assert_eq!(
-            bitrot_create_file_size(rustfs_rio::HashReader::SIZE_PRESERVE_LAYER, 4, &HashAlgorithm::HighwayHash256),
-            rustfs_rio::HashReader::SIZE_PRESERVE_LAYER
-        );
-    }
-
     struct TestChunkReader {
         chunks: VecDeque<Bytes>,
     }

crates/ecstore/src/set_disk/mod.rs

@@ -227,70 +227,6 @@ pub(super) fn restore_commit_operation_id_from_metadata(metadata: &HashMap<Strin
     restore_operation_id_from_metadata(metadata)
 }
 
-async fn inspect_decommission_tier_free_version_target(
-    disk: &DiskStore,
-    bucket: &str,
-    object: &str,
-    source: &FileInfo,
-) -> Result<bool> {
-    let raw = match disk.read_xl(bucket, object, false).await {
-        Ok(raw) => raw,
-        Err(DiskError::FileNotFound | DiskError::FileVersionNotFound | DiskError::VolumeNotFound) => return Ok(false),
-        Err(err) => return Err(err.into()),
-    };
-    let meta = FileMeta::load(&raw.buf)?;
-    let source_version_id = source.version_id.filter(|version_id| !version_id.is_nil());
-    let mut matching_count = 0;
-    let mut all_matching_versions_equivalent = true;
-    for existing in meta
-        .versions
-        .iter()
-        .filter(|version| version.header.version_id.filter(|version_id| !version_id.is_nil()) == source_version_id)
-    {
-        matching_count += 1;
-        let existing = existing.into_fileinfo(bucket, object, true)?;
-        existing.validate_for_metadata_read()?;
-        if !existing.tier_free_version() || !crate::store::tiered_data_movement_source_matches(source, &existing)? {
-            all_matching_versions_equivalent = false;
-        }
-    }
-    if matching_count == 0 {
-        return Ok(false);
-    }
-    if matching_count == 1 && all_matching_versions_equivalent {
-        return Ok(true);
-    }
-
-    Err(StorageError::DataMovementOverwriteErr(
-        bucket.to_owned(),
-        object.to_owned(),
-        source_version_id.map(|version_id| version_id.to_string()).unwrap_or_default(),
-    )
-    .into())
-}
-
-fn ensure_decommission_tier_free_version_commit_fence(bucket: &str, object: &str, opts: &ObjectOptions) -> Result<()> {
-    if opts
-        .namespace_lock_fence
-        .as_ref()
-        .is_some_and(NamespaceLockFence::is_lock_lost)
-        || opts
-            .bucket_lifecycle_lock_fence
-            .as_ref()
-            .is_some_and(NamespaceLockFence::is_lock_lost)
-    {
-        return Err(StorageError::NamespaceLockQuorumUnavailable {
-            mode: "decommission_tier_free_version_commit",
-            bucket: bucket.to_string(),
-            object: object.to_string(),
-            required: 1,
-            achieved: 0,
-        });
-    }
-
-    Ok(())
-}
-
 impl SetDisks {
     pub(super) async fn require_current_restore_operation_id(
         &self,
@@ -4729,98 +4665,6 @@ fn resolve_delete_version_state(opts: &ObjectOptions, goi: &ObjectInfo, version_
 }
 
 impl SetDisks {
-    /// Publish an internal tier free-version record without changing its
-    /// delete-marker shape or remote-tier identity. The caller holds the
-    /// source and target object locks; a write quorum is required before the
-    /// source cleanup may remove the original record.
-    #[tracing::instrument(skip(self, fi, opts))]
-    pub(crate) async fn decommission_tier_free_version(
-        &self,
-        bucket: &str,
-        object: &str,
-        fi: &FileInfo,
-        opts: &ObjectOptions,
-    ) -> Result<()> {
-        if !fi.deleted || !fi.tier_free_version() {
-            return Err(Error::other("decommission tier free-version write requires a free version record"));
-        }
-        ensure_decommission_tier_free_version_commit_fence(bucket, object, opts)?;
-
-        self.validate_decommission_tier_free_version_target(bucket, object, fi)
-            .await?;
-        ensure_decommission_tier_free_version_commit_fence(bucket, object, opts)?;
-
-        let disks = self.disks.read().await.clone();
-        let write_quorum = self.default_write_quorum();
-        let futures = disks.into_iter().map(|disk| {
-            let file_info = fi.clone();
-            async move {
-                if let Some(disk) = disk {
-                    disk.write_metadata("", bucket, object, file_info).await
-                } else {
-                    Err(DiskError::DiskNotFound)
-                }
-            }
-        });
-
-        let mut errs = Vec::new();
-        for result in join_all(futures).await {
-            match result {
-                Ok(_) => errs.push(None),
-                Err(err) => errs.push(Some(err)),
-            }
-        }
-
-        ensure_decommission_tier_free_version_commit_fence(bucket, object, opts)?;
-
-        resolve_tiered_decommission_write_quorum_result(&errs, write_quorum, bucket, object)
-    }
-
-    pub(crate) async fn validate_decommission_tier_free_version_target(
-        &self,
-        bucket: &str,
-        object: &str,
-        fi: &FileInfo,
-    ) -> Result<()> {
-        // The caller holds the source and target object locks. Inspect every
-        // target disk before an idempotent return or metadata fan-out so a
-        // sub-quorum conflict cannot be hidden by a successful quorum.
-        let disks = self.disks.read().await.clone();
-        let preflight = disks
-            .iter()
-            .flatten()
-            .map(|disk| inspect_decommission_tier_free_version_target(disk, bucket, object, fi));
-        for result in join_all(preflight).await {
-            result?;
-        }
-        Ok(())
-    }
-
-    pub(crate) async fn has_decommission_tier_free_version_write_quorum(
-        &self,
-        bucket: &str,
-        object: &str,
-        fi: &FileInfo,
-        opts: &ObjectOptions,
-    ) -> Result<bool> {
-        ensure_decommission_tier_free_version_commit_fence(bucket, object, opts)?;
-        let disks = self.disks.read().await.clone();
-        let preflight = disks.iter().map(|disk| async {
-            match disk {
-                Some(disk) => inspect_decommission_tier_free_version_target(disk, bucket, object, fi).await,
-                None => Ok(false),
-            }
-        });
-        let mut equivalent = 0;
-        for result in join_all(preflight).await {
-            if result? {
-                equivalent += 1;
-            }
-        }
-        ensure_decommission_tier_free_version_commit_fence(bucket, object, opts)?;
-        Ok(equivalent >= self.default_write_quorum())
-    }
-
     #[tracing::instrument(skip(self, fi, opts))]
     pub(crate) async fn decommission_tiered_object(
         &self,
@@ -10043,147 +9887,6 @@ mod tests {
         assert_ne!(updated.erasure.distribution, original.erasure.distribution);
     }
 
-    #[tokio::test]
-    async fn decommission_tier_free_version_preserves_remote_identity() {
-        let set_disks = make_local_bucket_test_set_disks().await;
-        let bucket = "free-version-decommission";
-        let object = "object.txt";
-        set_disks
-            .make_bucket(bucket, &MakeBucketOptions::default())
-            .await
-            .expect("target bucket should exist before free-version migration");
-        let version_id = Uuid::new_v4();
-        let mut free_version = FileInfo {
-            name: object.to_string(),
-            volume: bucket.to_string(),
-            version_id: Some(version_id),
-            mod_time: Some(time::OffsetDateTime::now_utc()),
-            deleted: true,
-            transition_tier: "WARM-TIER".to_string(),
-            transitioned_objname: "remote/object".to_string(),
-            ..Default::default()
-        };
-        free_version.set_tier_free_version();
-        // Decoded free versions always carry the on-disk free-version
-        // suffix alongside the in-memory tier marker; mirror that here so
-        // the record satisfies delete-marker metadata validation.
-        rustfs_utils::http::metadata_compat::insert_str(
-            &mut free_version.metadata,
-            rustfs_utils::http::metadata_compat::SUFFIX_FREE_VERSION,
-            String::new(),
-        );
-
-        set_disks
-            .decommission_tier_free_version(bucket, object, &free_version, &ObjectOptions::default())
-            .await
-            .expect("free-version metadata should reach the target quorum");
-        set_disks
-            .decommission_tier_free_version(bucket, object, &free_version, &ObjectOptions::default())
-            .await
-            .expect("replaying the same free-version metadata should be idempotent");
-
-        let versions = set_disks
-            .load_file_info_versions_exact(bucket, object)
-            .await
-            .expect("migrated free-version metadata should decode")
-            .expect("migrated free-version metadata should exist");
-        let migrated = versions
-            .versions
-            .iter()
-            .find(|version| version.version_id == Some(version_id))
-            .expect("free version should be present on the target");
-
-        assert_eq!(
-            versions
-                .versions
-                .iter()
-                .filter(|version| version.version_id == Some(version_id))
-                .count(),
-            1
-        );
-        assert!(migrated.tier_free_version());
-        assert_eq!(migrated.transition_tier, "WARM-TIER");
-        assert_eq!(migrated.transitioned_objname, "remote/object");
-    }
-
-    #[tokio::test]
-    async fn decommission_tier_free_version_resume_requires_write_quorum() {
-        let set_disks = make_local_bucket_test_set_disks_with_drive_count(4).await;
-        let bucket = "free-version-decommission-resume";
-        let object = "object.txt";
-        set_disks
-            .make_bucket(bucket, &MakeBucketOptions::default())
-            .await
-            .expect("target bucket should exist before free-version migration");
-        let mut free_version = FileInfo {
-            name: object.to_string(),
-            volume: bucket.to_string(),
-            version_id: Some(Uuid::new_v4()),
-            mod_time: Some(time::OffsetDateTime::now_utc()),
-            deleted: true,
-            transition_tier: "WARM-TIER".to_string(),
-            transitioned_objname: "remote/object".to_string(),
-            ..Default::default()
-        };
-        free_version.set_tier_free_version();
-        // Decoded free versions always carry the on-disk free-version
-        // suffix alongside the in-memory tier marker; mirror that here so
-        // the record satisfies delete-marker metadata validation.
-        rustfs_utils::http::metadata_compat::insert_str(
-            &mut free_version.metadata,
-            rustfs_utils::http::metadata_compat::SUFFIX_FREE_VERSION,
-            String::new(),
-        );
-        let opts = ObjectOptions::default();
-
-        let disks = set_disks.get_disks_internal().await;
-        for disk in disks.iter().take(2).flatten() {
-            disk.write_metadata("", bucket, object, free_version.clone())
-                .await
-                .expect("partial first attempt should leave equivalent metadata");
-        }
-        assert!(
-            !set_disks
-                .has_decommission_tier_free_version_write_quorum(bucket, object, &free_version, &opts)
-                .await
-                .expect("partial target metadata should remain valid"),
-            "write-quorum-minus-one must not be accepted as an idempotent migration"
-        );
-
-        disks[2]
-            .as_ref()
-            .expect("third target disk should be online")
-            .write_metadata("", bucket, object, free_version.clone())
-            .await
-            .expect("third equivalent target write should complete quorum");
-        assert!(
-            set_disks
-                .has_decommission_tier_free_version_write_quorum(bucket, object, &free_version, &opts)
-                .await
-                .expect("write-quorum target metadata should remain valid")
-        );
-    }
-
-    #[test]
-    fn decommission_tier_free_version_commit_rejects_lost_fence() {
-        let opts = ObjectOptions {
-            namespace_lock_fence: Some(NamespaceLockFence::lost_for_test()),
-            ..Default::default()
-        };
-
-        let err = ensure_decommission_tier_free_version_commit_fence("bucket", "object", &opts)
-            .expect_err("lost target lock must fail the free-version commit");
-        assert!(matches!(
-            err,
-            Error::NamespaceLockQuorumUnavailable {
-                mode: "decommission_tier_free_version_commit",
-                required: 1,
-                achieved: 0,
-                ..
-            }
-        ));
-    }
-
     #[test]
     fn test_resolve_tiered_decommission_write_quorum_result_allows_successful_quorum() {
         let errs = vec![None, None, Some(DiskError::DiskNotFound), None];

crates/ecstore/src/set_disk/ops/object.rs

@@ -2124,13 +2124,26 @@ impl SetDisks {
 
             let put_object_size = known_put_object_storage_size(data.size());
             let shard_file_size_raw = erasure.shard_file_size(put_object_size);
-            let is_inline_buffer = storage_class_config.should_inline(shard_file_size_raw, erasure.data_shards, opts.versioned);
+            let is_inline_buffer =
+                storage_class_config.should_inline(shard_file_size_raw, erasure.data_shards, opts.versioned);
 
             let collect_stage_timing = rustfs_io_metrics::put_stage_metrics_enabled() || issue3031_diag_enabled();
             let shard_file_size = shard_file_size_raw;
             let shard_size = erasure.shard_size();
             let write_path = classify_put_write_path(is_inline_buffer, put_object_size, fi.erasure.block_size);
             let direct_inline_commit = matches!(write_path, SmallWritePath::Inline);
+            {
+                use std::io::Write;
+                let msg = format!(
+                    "INLINE_DEBUG: bucket={} obj={} size={} shard_fs={} ds={} bs={} inline={} direct={} path={} iblock={} ver={}\n",
+                    bucket, object, put_object_size, shard_file_size_raw, erasure.data_shards, fi.erasure.block_size,
+                    is_inline_buffer, direct_inline_commit, write_path.metric_label(), storage_class_config.inline_block(), opts.versioned
+                );
+                if let Ok(mut f) = std::fs::OpenOptions::new().create(true).append(true).open("/tmp/rustfs_inline_debug.log") {
+                    let _ = f.write_all(msg.as_bytes());
+                }
+                let _ = std::io::stderr().write_all(msg.as_bytes());
+            }
             rustfs_io_metrics::record_put_object_path(write_path.metric_label());
             let writer_setup_stage_start = collect_stage_timing.then(Instant::now);
             let (mut writers, errors) = if direct_inline_commit {

crates/ecstore/src/store/heal.rs

@@ -13,12 +13,7 @@
 // limitations under the License.
 
 use super::*;
-use crate::core::pools::POOL_META_NAME;
-use crate::services::rebalance::{REBAL_META_NAME, RebalStatus};
-use crate::set_disk::get_lock_acquire_timeout;
 use crate::storage_api_contracts::heal::HealOperations as _;
-use crate::storage_api_contracts::namespace::NamespaceLocking as _;
-use rustfs_lock::NamespaceLockGuard;
 use tracing::trace;
 
 const LOG_COMPONENT_ECSTORE: &str = "ecstore";
@@ -35,119 +30,7 @@ fn invalid_heal_pool_index(pool_idx: usize, pool_count: usize) -> Error {
     )
 }
 
-#[derive(Debug, Clone, Copy)]
-enum HealFormatPoolSkip {
-    Completed,
-    Retryable,
-}
-
-fn classify_heal_format_pool(
-    pool_idx: usize,
-    pool_cmd_line: &str,
-    pool_meta: &PoolMeta,
-    rebalance_meta: Option<&RebalanceMeta>,
-) -> Option<HealFormatPoolSkip> {
-    let Some(pool) = pool_meta.pools.get(pool_idx) else {
-        return Some(HealFormatPoolSkip::Retryable);
-    };
-
-    if pool.id != pool_idx || pool_cmd_line.is_empty() || pool.cmd_line.is_empty() || pool.cmd_line != pool_cmd_line {
-        return Some(HealFormatPoolSkip::Retryable);
-    }
-
-    if let Some(decommission) = pool.decommission.as_ref() {
-        if decommission.complete {
-            return Some(HealFormatPoolSkip::Completed);
-        }
-        if decommission.failed || decommission.canceled || decommission.queued || pool_meta.is_suspended(pool_idx) {
-            return Some(HealFormatPoolSkip::Retryable);
-        }
-    }
-
-    if let Some(meta) = rebalance_meta {
-        let Some(pool_stats) = meta.pool_stats.get(pool_idx) else {
-            return Some(HealFormatPoolSkip::Retryable);
-        };
-        if pool_stats.info.stopping || (pool_stats.participating && pool_stats.info.status == RebalStatus::Started) {
-            return Some(HealFormatPoolSkip::Retryable);
-        }
-    }
-
-    None
-}
-
-fn heal_format_pool_skip_error(skip: HealFormatPoolSkip) -> Error {
-    match skip {
-        HealFormatPoolSkip::Completed => StorageError::NoHealRequired,
-        HealFormatPoolSkip::Retryable => StorageError::SlowDown,
-    }
-}
-
-fn heal_format_fence_lost_error() -> Error {
-    StorageError::SlowDown
-}
-
 impl ECStore {
-    async fn acquire_heal_format_fence(
-        &self,
-    ) -> Result<(NamespaceLockGuard, NamespaceLockGuard, PoolMeta, Option<RebalanceMeta>)> {
-        let metadata_pool = self
-            .pools
-            .first()
-            .cloned()
-            .ok_or_else(|| Error::other("heal format requires at least one storage pool"))?;
-
-        // Metadata fence order is part of the decommission/rebalance protocol:
-        // pool.bin must always be acquired before rebalance.bin.
-        let pool_lock = metadata_pool.new_ns_lock(RUSTFS_META_BUCKET, POOL_META_NAME).await?;
-        let pool_guard = pool_lock.get_write_lock(get_lock_acquire_timeout()).await?;
-        let rebalance_lock = metadata_pool.new_ns_lock(RUSTFS_META_BUCKET, REBAL_META_NAME).await?;
-        let rebalance_guard = rebalance_lock.get_write_lock(get_lock_acquire_timeout()).await?;
-
-        if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
-            return Err(heal_format_fence_lost_error());
-        }
-
-        let mut pool_meta = PoolMeta::default();
-        pool_meta.load_no_lock(metadata_pool.clone()).await?;
-        if pool_meta.pools.len() != self.pools.len()
-            || pool_meta.pools.iter().enumerate().any(|(pool_idx, pool)| {
-                pool.id != pool_idx || pool.cmd_line.is_empty() || pool.cmd_line != self.pools[pool_idx].endpoints.cmd_line
-            })
-        {
-            return Err(heal_format_fence_lost_error());
-        }
-
-        let mut rebalance_meta = RebalanceMeta::new();
-        let rebalance_meta = match rebalance_meta
-            .load_with_opts(
-                metadata_pool,
-                ObjectOptions {
-                    no_lock: true,
-                    ..Default::default()
-                },
-            )
-            .await
-        {
-            Ok(()) => Some(rebalance_meta),
-            Err(Error::ConfigNotFound) => None,
-            Err(err) => return Err(err),
-        };
-
-        if rebalance_meta
-            .as_ref()
-            .is_some_and(|meta| meta.pool_stats.len() != self.pools.len())
-        {
-            return Err(heal_format_fence_lost_error());
-        }
-
-        if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
-            return Err(heal_format_fence_lost_error());
-        }
-
-        Ok((pool_guard, rebalance_guard, pool_meta, rebalance_meta))
-    }
-
     fn get_pools_for_heal_object(&self, opts: &HealOpts) -> Result<Vec<Arc<Sets>>> {
         match opts.pool {
             Some(pool_idx) => Ok(vec![
@@ -169,26 +52,9 @@ impl ECStore {
         };
 
         let mut count_no_heal = 0;
-        let mut count_completed = 0;
         let mut first_error = None;
-        for (pool_idx, pool) in self.pools.iter().enumerate() {
-            let (pool_guard, rebalance_guard, pool_meta, rebalance_meta) = self.acquire_heal_format_fence().await?;
-            if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
-                first_error.get_or_insert(heal_format_fence_lost_error());
-                break;
-            }
-            if let Some(skip) = classify_heal_format_pool(pool_idx, &pool.endpoints.cmd_line, &pool_meta, rebalance_meta.as_ref())
-            {
-                if matches!(skip, HealFormatPoolSkip::Completed) {
-                    count_completed += 1;
-                } else {
-                    first_error.get_or_insert(heal_format_pool_skip_error(skip));
-                }
-                continue;
-            }
-
-            let fence_lost = || pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost();
-            let (mut result, err) = pool.heal_format_with_fence(dry_run, fence_lost).await?;
+        for pool in self.pools.iter() {
+            let (mut result, err) = pool.heal_format(dry_run).await?;
             if let Some(err) = err {
                 match err {
                     StorageError::NoHealRequired => {
@@ -203,18 +69,11 @@ impl ECStore {
             r.set_count += result.set_count;
             r.before.drives.append(&mut result.before.drives);
             r.after.drives.append(&mut result.after.drives);
-
-            // A lease can be lost after the final write; fail closed before
-            // reporting the pool as successfully healed.
-            if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
-                first_error.get_or_insert(heal_format_fence_lost_error());
-                break;
-            }
         }
         if let Some(err) = first_error {
             return Ok((r, Some(err)));
         }
-        if count_no_heal + count_completed == self.pools.len() {
+        if count_no_heal == self.pools.len() {
             info!(
                 event = EVENT_HEAL_FORMAT_COMPLETED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -443,7 +302,6 @@ mod tests {
     use crate::disk::{DeleteOptions, DiskOption, format::FormatV3, new_disk};
     use crate::layout::endpoints::{EndpointServerPools, Endpoints, PoolEndpoints};
     use crate::runtime::instance::InstanceContext;
-    use crate::services::rebalance::{RebalanceInfo, RebalanceStats};
     use crate::storage_api_contracts::bucket::{BucketOperations, MakeBucketOptions};
     use crate::storage_api_contracts::object::{ObjectIO as _, ObjectOperations};
     use crate::store::init_format::{load_format_erasure, save_format_file};
@@ -495,164 +353,6 @@ mod tests {
         }
     }
 
-    fn pool_meta_with_decommission(info: PoolDecommissionInfo) -> PoolMeta {
-        PoolMeta {
-            pools: vec![PoolStatus {
-                id: 0,
-                cmd_line: "pool-0".to_string(),
-                last_update: OffsetDateTime::UNIX_EPOCH,
-                decommission: Some(info),
-            }],
-            ..Default::default()
-        }
-    }
-
-    #[test]
-    fn heal_format_pool_state_barriers_are_classified() {
-        let active = pool_meta_with_decommission(PoolDecommissionInfo {
-            start_time: Some(OffsetDateTime::UNIX_EPOCH),
-            ..Default::default()
-        });
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-0", &active, None),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        for info in [
-            PoolDecommissionInfo {
-                failed: true,
-                ..Default::default()
-            },
-            PoolDecommissionInfo {
-                canceled: true,
-                ..Default::default()
-            },
-        ] {
-            assert!(matches!(
-                classify_heal_format_pool(0, "pool-0", &pool_meta_with_decommission(info), None),
-                Some(HealFormatPoolSkip::Retryable)
-            ));
-        }
-
-        let completed = pool_meta_with_decommission(PoolDecommissionInfo {
-            complete: true,
-            ..Default::default()
-        });
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-0", &completed, None),
-            Some(HealFormatPoolSkip::Completed)
-        ));
-    }
-
-    #[test]
-    fn heal_format_pool_rebalance_barriers_and_identity_are_fail_closed() {
-        let identity_meta = pool_meta_with_decommission(PoolDecommissionInfo::default());
-        let rebalance = RebalanceMeta {
-            pool_stats: vec![RebalanceStats {
-                participating: true,
-                info: RebalanceInfo {
-                    status: RebalStatus::Started,
-                    ..Default::default()
-                },
-                ..Default::default()
-            }],
-            ..Default::default()
-        };
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&rebalance)),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        let stopping = RebalanceMeta {
-            pool_stats: vec![RebalanceStats {
-                info: RebalanceInfo {
-                    stopping: true,
-                    ..Default::default()
-                },
-                ..Default::default()
-            }],
-            ..Default::default()
-        };
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&stopping)),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        let identity = pool_meta_with_decommission(PoolDecommissionInfo::default());
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-new", &identity, None),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        let identity_without_decommission = PoolMeta {
-            pools: vec![PoolStatus {
-                id: 0,
-                cmd_line: "pool-0".to_string(),
-                last_update: OffsetDateTime::UNIX_EPOCH,
-                decommission: None,
-            }],
-            ..Default::default()
-        };
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-new", &identity_without_decommission, None),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        assert!(matches!(
-            classify_heal_format_pool(0, "", &identity_meta, None),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-0", &PoolMeta::default(), None),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-
-        let stopped = RebalanceMeta {
-            stopped_at: Some(OffsetDateTime::UNIX_EPOCH),
-            pool_stats: vec![RebalanceStats {
-                participating: true,
-                info: RebalanceInfo {
-                    status: RebalStatus::Stopped,
-                    ..Default::default()
-                },
-                ..Default::default()
-            }],
-            ..Default::default()
-        };
-        assert!(classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&stopped)).is_none());
-
-        let stopping_after_stop = RebalanceMeta {
-            stopped_at: Some(OffsetDateTime::UNIX_EPOCH),
-            pool_stats: vec![RebalanceStats {
-                participating: true,
-                info: RebalanceInfo {
-                    status: RebalStatus::Started,
-                    stopping: true,
-                    ..Default::default()
-                },
-                ..Default::default()
-            }],
-            ..Default::default()
-        };
-        assert!(matches!(
-            classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&stopping_after_stop)),
-            Some(HealFormatPoolSkip::Retryable)
-        ));
-    }
-
-    #[test]
-    fn skipped_heal_format_pool_is_never_reported_as_success() {
-        assert!(matches!(
-            heal_format_pool_skip_error(HealFormatPoolSkip::Retryable),
-            StorageError::SlowDown
-        ));
-        assert!(matches!(
-            heal_format_pool_skip_error(HealFormatPoolSkip::Completed),
-            StorageError::NoHealRequired
-        ));
-    }
-
     async fn multi_pool_heal_store() -> (tempfile::TempDir, Arc<ECStore>, CancellationToken) {
         let temp_dir = tempfile::tempdir().expect("multi-pool heal test directory should be created");
         let mut pool_endpoints = Vec::new();
@@ -1189,18 +889,6 @@ mod tests {
             bucket_fence_registry: std::sync::Arc::default(),
         };
 
-        let err = store
-            .handle_heal_format(false)
-            .await
-            .expect_err("missing pool metadata must fail closed before format writes");
-        assert!(matches!(err, StorageError::SlowDown));
-
-        let pool_meta = PoolMeta::new(&store.pools, &PoolMeta::default());
-        pool_meta
-            .save(store.pools.clone())
-            .await
-            .expect("pool metadata should be persisted before format heal");
-
         let (result, err) = store
             .handle_heal_format(false)
             .await
@@ -1214,22 +902,5 @@ mod tests {
             .await
             .expect("the later pool should be healed despite the first pool error");
         assert_eq!(healed.erasure.this, recoverable_format.erasure.sets[0][2]);
-
-        let mut completed_meta = PoolMeta::new(&store.pools, &PoolMeta::default());
-        for status in &mut completed_meta.pools {
-            status.decommission = Some(PoolDecommissionInfo {
-                complete: true,
-                ..Default::default()
-            });
-        }
-        completed_meta
-            .save(store.pools.clone())
-            .await
-            .expect("completed pool metadata should be persisted");
-        let (_, err) = store
-            .handle_heal_format(false)
-            .await
-            .expect("completed pools should be reported as a no-op");
-        assert!(matches!(err, Some(StorageError::NoHealRequired)));
     }
 }

crates/ecstore/src/store/init.rs

@@ -553,8 +553,6 @@ mod tests {
         should_retry_format_load, should_retry_local_decommission_resume, wait_for_local_decommission_resume_delay,
     };
     #[cfg(feature = "test-util")]
-    use crate::disk::DiskAPI;
-    #[cfg(feature = "test-util")]
     use crate::{
         bucket::lifecycle::{
             lifecycle::{TRANSITION_PENDING, TransitionOptions},
@@ -1309,77 +1307,6 @@ mod tests {
         });
     }
 
-    #[cfg(feature = "test-util")]
-    async fn seed_transitioned_free_version(
-        ctx: &Arc<crate::runtime::instance::InstanceContext>,
-        store: &Arc<crate::store::ECStore>,
-        bucket: &str,
-        object: &str,
-    ) -> (uuid::Uuid, uuid::Uuid) {
-        let tier_name = format!("DECOMFREE{}", uuid::Uuid::new_v4().simple());
-        register_mock_tier(&ctx.tier_config_mgr(), &tier_name).await;
-
-        let mut reader = PutObjReader::from_vec(b"transitioned source bytes".to_vec());
-        let source = store.pools[0]
-            .put_object(
-                bucket,
-                object,
-                &mut reader,
-                &ObjectOptions {
-                    versioned: true,
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("write transitioned decommission source");
-        let source_version = source.version_id.expect("transitioned source must be versioned");
-        store.pools[0]
-            .transition_object(
-                bucket,
-                object,
-                &ObjectOptions {
-                    versioned: true,
-                    version_id: Some(source_version.to_string()),
-                    transition: TransitionOptions {
-                        status: TRANSITION_PENDING.to_string(),
-                        tier: tier_name,
-                        etag: source.etag.clone().expect("transitioned source must have an ETag"),
-                        ..Default::default()
-                    },
-                    mod_time: source.mod_time,
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("transition source before decommission");
-        store.pools[0]
-            .delete_object(
-                bucket,
-                object,
-                ObjectOptions {
-                    versioned: true,
-                    version_id: Some(source_version.to_string()),
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("delete transitioned source version");
-
-        let versions = store.pools[0]
-            .get_disks_by_key(object)
-            .load_file_info_versions_exact(bucket, object)
-            .await
-            .expect("source versions should decode after transition delete")
-            .expect("source free version should remain after transition delete");
-        let free_version = versions
-            .versions
-            .iter()
-            .find(|version| version.tier_free_version())
-            .and_then(|version| version.version_id)
-            .expect("transition delete should create a free version");
-        (source_version, free_version)
-    }
-
     async fn write_decommission_test_multipart_source(
         store: &Arc<crate::store::ECStore>,
         pool_idx: usize,
@@ -3919,383 +3846,6 @@ mod tests {
         shutdown.cancel();
     }
 
-    #[cfg(feature = "test-util")]
-    #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
-    #[serial_test::serial(storage_class_env)]
-    async fn decommission_entry_skips_cleanup_only_marker_when_free_version_is_present() {
-        let temp_dir = tempfile::tempdir().expect("create free-version decommission store dir");
-        let (ctx, store, shutdown) =
-            without_storage_class_env(build_isolated_test_store(temp_dir.path(), "decommission-free-marker", &[4, 4])).await;
-        crate::bucket::metadata_sys::init_bucket_metadata_sys(store.clone(), Vec::new()).await;
-        let bucket = format!("decom-free-marker-{}", uuid::Uuid::new_v4());
-        let object = "free-marker-object";
-        store
-            .make_bucket(&bucket, &MakeBucketOptions::default())
-            .await
-            .expect("create free-version decommission bucket");
-        let (_, free_version) = seed_transitioned_free_version(&ctx, &store, &bucket, object).await;
-        let source_free = store.pools[0]
-            .get_disks_by_key(object)
-            .load_file_info_versions_exact(&bucket, object)
-            .await
-            .expect("source free-version metadata should decode")
-            .and_then(|versions| {
-                versions
-                    .versions
-                    .into_iter()
-                    .find(|version| version.version_id == Some(free_version) && version.tier_free_version())
-            })
-            .expect("source free-version identity should be present before decommission");
-        let mut target_reader = PutObjReader::from_vec(b"target ordinary bytes".to_vec());
-        let target_w = store.pools[1]
-            .put_object(
-                &bucket,
-                object,
-                &mut target_reader,
-                &ObjectOptions {
-                    versioned: true,
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("write unrelated target version");
-        let target_w_version = target_w.version_id.expect("target version should have an id");
-        let marker = store.pools[0]
-            .delete_object(
-                &bucket,
-                object,
-                ObjectOptions {
-                    versioned: true,
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("write cleanup-only delete marker");
-        assert!(marker.delete_marker);
-
-        mark_test_pool_decommissioning(&store, 0).await;
-        let source_set = store.pools[0].get_disks_by_key(object);
-        store
-            .decommission_entry_for_test_with_bucket_incarnation(
-                0,
-                MetaCacheEntry {
-                    name: object.to_string(),
-                    ..Default::default()
-                },
-                bucket.clone(),
-                source_set.clone(),
-            )
-            .await
-            .expect("real decommission entry should migrate the free version");
-
-        let target_versions = store.pools[1]
-            .get_disks_by_key(object)
-            .load_file_info_versions_exact(&bucket, object)
-            .await
-            .expect("target versions should decode")
-            .expect("target free version should be present");
-        assert!(
-            target_versions
-                .versions
-                .iter()
-                .any(|version| { version.version_id == Some(free_version) && version.tier_free_version() })
-        );
-        let migrated_free = target_versions
-            .versions
-            .iter()
-            .find(|version| version.version_id == Some(free_version) && version.tier_free_version())
-            .expect("migrated free-version identity should remain readable from target disks");
-        assert!(
-            crate::store::tiered_data_movement_source_matches(&source_free, migrated_free)
-                .expect("migrated free-version identity should decode")
-        );
-        let retained_w = target_versions
-            .versions
-            .iter()
-            .find(|version| version.version_id == Some(target_w_version))
-            .expect("unrelated target version should remain");
-        assert!(!retained_w.deleted && !retained_w.tier_free_version());
-        assert_eq!(retained_w.size, target_w.size);
-        assert_eq!(retained_w.get_etag(), target_w.etag);
-        assert!(
-            target_versions
-                .versions
-                .iter()
-                .all(|version| { version.tier_free_version() || !version.deleted })
-        );
-        assert!(
-            source_set
-                .load_file_info_versions_exact(&bucket, object)
-                .await
-                .expect("source versions should be readable after cleanup")
-                .is_none(),
-            "successful free-version migration should permit source cleanup"
-        );
-        let (heal_versions, _, _) = store
-            .heal_walk_versions_page(1, 0, &bucket, "", None, 2, 16, true)
-            .await
-            .expect("heal walk should decode the migrated free version");
-        let free_version_string = free_version.to_string();
-        let healed_free = heal_versions
-            .iter()
-            .find(|version| version.version_id.as_deref() == Some(free_version_string.as_str()))
-            .expect("heal walk should surface the migrated free version");
-        let healed_info = healed_free
-            .lifecycle_object_info
-            .as_ref()
-            .expect("heal walk should retain lifecycle identity for the migrated free version");
-        assert!(healed_info.transitioned_object.free_version);
-        assert_eq!(healed_info.transitioned_object.tier, source_free.transition_tier);
-        assert_eq!(healed_info.transitioned_object.name, source_free.transitioned_objname);
-        shutdown.cancel();
-    }
-
-    #[cfg(feature = "test-util")]
-    #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
-    #[serial_test::serial(storage_class_env)]
-    async fn decommission_entry_allows_free_version_consumed_before_source_lock() {
-        let temp_dir = tempfile::tempdir().expect("create consumed free-version store dir");
-        let (ctx, store, shutdown) =
-            without_storage_class_env(build_isolated_test_store(temp_dir.path(), "decommission-free-consumed", &[4, 4])).await;
-        crate::bucket::metadata_sys::init_bucket_metadata_sys(store.clone(), Vec::new()).await;
-        let bucket = format!("decom-free-consumed-{}", uuid::Uuid::new_v4());
-        let object = "free-consumed-object";
-        store
-            .make_bucket(&bucket, &MakeBucketOptions::default())
-            .await
-            .expect("create consumed free-version bucket");
-        let (_, free_version) = seed_transitioned_free_version(&ctx, &store, &bucket, object).await;
-
-        mark_test_pool_decommissioning(&store, 0).await;
-        let source_set = store.pools[0].get_disks_by_key(object);
-        let barrier = crate::store::object::DecommissionFreeVersionSourceRaceBarrier::install(&bucket, object);
-        let decommission = tokio::spawn({
-            let store = store.clone();
-            let bucket = bucket.clone();
-            let source_set = source_set.clone();
-            async move {
-                store
-                    .decommission_entry_for_test_with_bucket_incarnation(
-                        0,
-                        MetaCacheEntry {
-                            name: object.to_string(),
-                            ..Default::default()
-                        },
-                        bucket,
-                        source_set,
-                    )
-                    .await
-            }
-        });
-
-        barrier.wait_until_paused().await;
-        store.pools[0]
-            .delete_object(
-                &bucket,
-                object,
-                ObjectOptions {
-                    versioned: true,
-                    version_id: Some(free_version.to_string()),
-                    incl_free_versions: true,
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("lifecycle should consume the source free version before decommission locks it");
-        assert!(
-            source_set
-                .load_file_info_versions_exact(&bucket, object)
-                .await
-                .expect("consumed source metadata should remain readable")
-                .is_none(),
-            "the lifecycle delete should remove the source free version"
-        );
-
-        barrier.release();
-        decommission
-            .await
-            .expect("decommission task should join")
-            .expect("a concurrently consumed free version should not fail source cleanup");
-        assert!(
-            store.pools[1]
-                .get_disks_by_key(object)
-                .load_file_info_versions_exact(&bucket, object)
-                .await
-                .expect("target metadata should remain readable")
-                .is_none(),
-            "an already consumed free version should not be recreated on the target"
-        );
-        shutdown.cancel();
-    }
-
-    #[cfg(feature = "test-util")]
-    #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
-    #[serial_test::serial(storage_class_env)]
-    async fn decommission_entry_rejects_subquorum_free_version_conflict_and_retains_source() {
-        let temp_dir = tempfile::tempdir().expect("create sub-quorum free-version store dir");
-        let (ctx, store, shutdown) =
-            without_storage_class_env(build_isolated_test_store(temp_dir.path(), "decommission-free-conflict", &[4, 4])).await;
-        crate::bucket::metadata_sys::init_bucket_metadata_sys(store.clone(), Vec::new()).await;
-        let bucket = format!("decom-free-conflict-{}", uuid::Uuid::new_v4());
-        let object = "free-conflict-object";
-        store
-            .make_bucket(&bucket, &MakeBucketOptions::default())
-            .await
-            .expect("create sub-quorum conflict bucket");
-        let (_, free_version) = seed_transitioned_free_version(&ctx, &store, &bucket, object).await;
-        let source_free = store.pools[0]
-            .get_disks_by_key(object)
-            .load_file_info_versions_exact(&bucket, object)
-            .await
-            .expect("source free version should decode before crash replay setup")
-            .and_then(|versions| {
-                versions
-                    .versions
-                    .into_iter()
-                    .find(|version| version.version_id == Some(free_version))
-            })
-            .expect("source free version should be available for crash replay setup");
-
-        let mut target_reader = PutObjReader::from_vec(b"target ordinary bytes".to_vec());
-        let target = store.pools[1]
-            .put_object(
-                &bucket,
-                object,
-                &mut target_reader,
-                &ObjectOptions {
-                    versioned: true,
-                    ..Default::default()
-                },
-            )
-            .await
-            .expect("seed ordinary target version");
-        let target_version = target.version_id.expect("target version must have an ID");
-        let target_disks = store.pools[1].get_disks_by_key(object).disks.read().await.clone();
-        for disk in target_disks.iter().skip(1) {
-            disk.as_ref()
-                .expect("target crash replay quorum disk should be online")
-                .write_metadata("", &bucket, object, source_free.clone())
-                .await
-                .expect("seed an equivalent free version on the target quorum");
-        }
-        let conflict_path = temp_dir
-            .path()
-            .join(format!("pool1/set0/disk0/{bucket}/{object}/{STORAGE_FORMAT_FILE}"));
-        let encoded = tokio::fs::read(&conflict_path)
-            .await
-            .expect("target metadata should be readable");
-        let mut metadata = FileMeta::load(&encoded).expect("target metadata should decode");
-        let target_index = metadata
-            .versions
-            .iter()
-            .position(|version| version.header.version_id == Some(target_version))
-            .expect("target version should be present on the conflict disk");
-        let mut target_meta = metadata.versions[target_index]
-            .parse_version_meta()
-            .expect("target version metadata should decode");
-        target_meta
-            .object
-            .as_mut()
-            .expect("target conflict must remain an ordinary object")
-            .version_id = Some(free_version);
-        metadata.versions[target_index] = target_meta.try_into().expect("conflict metadata should encode");
-        let expected_conflict_meta = metadata.versions[target_index].meta.clone();
-        let expected_conflict = metadata.versions[target_index]
-            .into_fileinfo(&bucket, object, true)
-            .expect("conflict metadata should decode as an ordinary object");
-        let duplicate_free: rustfs_filemeta::FileMetaShallowVersion = rustfs_filemeta::FileMetaVersion::from(source_free.clone())
-            .try_into()
-            .expect("duplicate free metadata should encode");
-        metadata.versions.insert(target_index, duplicate_free);
-        tokio::fs::write(&conflict_path, metadata.marshal_msg().expect("conflict metadata should encode"))
-            .await
-            .expect("write sub-quorum conflict metadata");
-
-        mark_test_pool_decommissioning(&store, 0).await;
-        let source_set = store.pools[0].get_disks_by_key(object);
-        store
-            .decommission_entry_for_test_with_bucket_incarnation(
-                0,
-                MetaCacheEntry {
-                    name: object.to_string(),
-                    ..Default::default()
-                },
-                bucket.clone(),
-                source_set.clone(),
-            )
-            .await
-            .expect("conflicted decommission entry should retain the source and retry later");
-
-        let source_versions = source_set
-            .load_file_info_versions_exact(&bucket, object)
-            .await
-            .expect("retained source versions should decode")
-            .expect("source free version should be retained after conflict");
-        assert!(
-            source_versions
-                .versions
-                .iter()
-                .any(|version| { version.version_id == Some(free_version) && version.tier_free_version() })
-        );
-        let post_encoded = tokio::fs::read(&conflict_path)
-            .await
-            .expect("conflict metadata should remain readable");
-        let post_metadata = FileMeta::load(&post_encoded).expect("post-conflict metadata should decode");
-        let same_id = post_metadata
-            .versions
-            .iter()
-            .filter(|version| version.header.version_id == Some(free_version))
-            .collect::<Vec<_>>();
-        assert_eq!(same_id.len(), 2, "conflict metadata should retain both same-ID records");
-        assert_eq!(same_id.iter().filter(|version| version.header.free_version()).count(), 1);
-        assert_eq!(same_id.iter().filter(|version| !version.header.free_version()).count(), 1);
-        let post_conflict = same_id
-            .into_iter()
-            .find(|version| !version.header.free_version())
-            .expect("ordinary conflict version must remain addressable by the source ID");
-        let post_conflict_info = post_conflict
-            .into_fileinfo(&bucket, object, true)
-            .expect("post-conflict ordinary metadata should decode");
-        assert!(!post_conflict_info.deleted && !post_conflict_info.tier_free_version());
-        assert_eq!(post_conflict.meta, expected_conflict_meta);
-        assert_eq!(post_conflict_info.size, expected_conflict.size);
-        assert_eq!(post_conflict_info.data_dir, expected_conflict.data_dir);
-        assert_eq!(post_conflict_info.metadata, expected_conflict.metadata);
-        assert_eq!(post_conflict_info.get_etag(), expected_conflict.get_etag());
-
-        for disk_index in 0..4 {
-            let target_path = temp_dir
-                .path()
-                .join(format!("pool1/set0/disk{disk_index}/{bucket}/{object}/{STORAGE_FORMAT_FILE}"));
-            let target_encoded = tokio::fs::read(&target_path)
-                .await
-                .expect("target metadata should remain readable");
-            let target_meta = FileMeta::load(&target_encoded).expect("target metadata should decode");
-            let same_id = target_meta
-                .versions
-                .iter()
-                .filter(|version| version.header.version_id == Some(free_version))
-                .collect::<Vec<_>>();
-            if disk_index == 0 {
-                assert_eq!(same_id.len(), 2);
-                assert!(same_id[0].header.free_version());
-                assert!(!same_id[1].header.free_version());
-            } else {
-                assert_eq!(same_id.len(), 1);
-                assert!(same_id[0].header.free_version());
-            }
-        }
-        let sweep_err = store
-            .check_after_decommission_for_test(0)
-            .await
-            .expect_err("final sweep must report the retained free version");
-        assert!(
-            sweep_err.to_string().contains("version(s) were found"),
-            "unexpected final sweep error: {sweep_err}"
-        );
-        shutdown.cancel();
-    }
-
     #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
     #[serial_test::serial(storage_class_env)]
     async fn versioned_batch_delete_marker_skips_decommission_source() {

crates/ecstore/src/store/mod.rs

@@ -151,7 +151,7 @@ pub(crate) mod init_format;
 pub(crate) mod list_objects;
 mod multipart;
 mod object;
-pub(crate) use object::{ObjectLockDiagGuard, SourceCleanupMutationFence, tiered_data_movement_source_matches};
+pub(crate) use object::{ObjectLockDiagGuard, SourceCleanupMutationFence};
 pub use object::{
     PrepareSelectObjectSnapshotError, PreparedGetObjectReader, SelectObjectSnapshot, SelectObjectSnapshotReadError,
     SnapshotConsistencyError,

crates/ecstore/src/store/object.rs

@@ -490,82 +490,6 @@ fn decommission_mutation_fence_for_test(
         .map(|hook| hook.fence.clone())
 }
 
-#[cfg(test)]
-struct DecommissionFreeVersionSourceRaceState {
-    bucket: String,
-    object: String,
-    arrived: tokio::sync::Notify,
-    release: tokio::sync::Notify,
-}
-
-#[cfg(test)]
-pub(crate) struct DecommissionFreeVersionSourceRaceBarrier {
-    state: Arc<DecommissionFreeVersionSourceRaceState>,
-}
-
-#[cfg(test)]
-static DECOMMISSION_FREE_VERSION_SOURCE_RACE_BARRIER: std::sync::OnceLock<
-    std::sync::Mutex<Option<Arc<DecommissionFreeVersionSourceRaceState>>>,
-> = std::sync::OnceLock::new();
-
-#[cfg(test)]
-impl DecommissionFreeVersionSourceRaceBarrier {
-    pub(crate) fn install(bucket: &str, object: &str) -> Self {
-        let state = Arc::new(DecommissionFreeVersionSourceRaceState {
-            bucket: bucket.to_string(),
-            object: object.to_string(),
-            arrived: tokio::sync::Notify::new(),
-            release: tokio::sync::Notify::new(),
-        });
-        let mut slot = DECOMMISSION_FREE_VERSION_SOURCE_RACE_BARRIER
-            .get_or_init(|| std::sync::Mutex::new(None))
-            .lock()
-            .expect("decommission free-version source race barrier should not poison");
-        assert!(slot.is_none(), "decommission free-version source race barrier must be unique");
-        *slot = Some(Arc::clone(&state));
-        Self { state }
-    }
-
-    pub(crate) async fn wait_until_paused(&self) {
-        tokio::time::timeout(Duration::from_secs(30), self.state.arrived.notified())
-            .await
-            .expect("decommission should pause before acquiring the free-version source lock");
-    }
-
-    pub(crate) fn release(&self) {
-        self.state.release.notify_one();
-    }
-}
-
-#[cfg(test)]
-impl Drop for DecommissionFreeVersionSourceRaceBarrier {
-    fn drop(&mut self) {
-        self.state.release.notify_one();
-        let mut slot = DECOMMISSION_FREE_VERSION_SOURCE_RACE_BARRIER
-            .get_or_init(|| std::sync::Mutex::new(None))
-            .lock()
-            .expect("decommission free-version source race barrier should not poison");
-        if slot.as_ref().is_some_and(|state| Arc::ptr_eq(state, &self.state)) {
-            *slot = None;
-        }
-    }
-}
-
-#[cfg(test)]
-async fn pause_decommission_free_version_before_source_lock(bucket: &str, object: &str) {
-    let state = DECOMMISSION_FREE_VERSION_SOURCE_RACE_BARRIER
-        .get_or_init(|| std::sync::Mutex::new(None))
-        .lock()
-        .expect("decommission free-version source race barrier should not poison")
-        .as_ref()
-        .filter(|state| state.bucket == bucket && state.object == object)
-        .cloned();
-    if let Some(state) = state {
-        state.arrived.notify_one();
-        state.release.notified().await;
-    }
-}
-
 pub(crate) struct SourceCleanupMutationFence {
     guard: ObjectLockDiagGuard,
     source_lock_covered: bool,
@@ -1570,15 +1494,13 @@ fn is_equivalent_data_movement_tiered_object(source: &rustfs_filemeta::FileInfo,
         && source_actual_size == target_actual_size
 }
 
-pub(crate) fn tiered_data_movement_source_matches(
+fn tiered_data_movement_source_matches(
     expected: &rustfs_filemeta::FileInfo,
     current: &rustfs_filemeta::FileInfo,
 ) -> Result<bool> {
     let expected_backend = crate::services::tier::tier::tier_destination_id_from_metadata(&expected.metadata)?;
     let current_backend = crate::services::tier::tier::tier_destination_id_from_metadata(&current.metadata)?;
     Ok(expected.version_id == current.version_id
-        && expected.deleted == current.deleted
-        && expected.tier_free_version() == current.tier_free_version()
         && expected.data_dir == current.data_dir
         && expected.mod_time == current.mod_time
         && expected.size == current.size
@@ -1592,15 +1514,6 @@ pub(crate) fn tiered_data_movement_source_matches(
         && expected_backend == current_backend)
 }
 
-fn decommission_free_version_overwrite_error(bucket: &str, object: &str, version_id: Option<Uuid>) -> Error {
-    StorageError::DataMovementOverwriteErr(
-        bucket.to_owned(),
-        object.to_owned(),
-        version_id.map(|id| id.to_string()).unwrap_or_default(),
-    )
-    .into()
-}
-
 fn should_check_data_movement_resume_target(src_pool_idx: usize, target_pool_idx: usize) -> bool {
     target_pool_idx != src_pool_idx
 }
@@ -2308,23 +2221,6 @@ impl ECStore {
         )
     }
 
-    async fn has_equivalent_data_movement_tier_free_version(
-        &self,
-        bucket: &str,
-        object: &str,
-        source: &rustfs_filemeta::FileInfo,
-        opts: &ObjectOptions,
-        target_pool_idx: usize,
-    ) -> Result<bool> {
-        let pool = self
-            .pools
-            .get(target_pool_idx)
-            .ok_or_else(|| Error::other(format!("invalid tiered data movement target pool {target_pool_idx}")))?;
-        pool.get_disks_by_key(object)
-            .has_decommission_tier_free_version_write_quorum(bucket, object, source, opts)
-            .await
-    }
-
     fn resolve_decommission_target_pool_idx_result(result: Result<usize>, bucket: &str, object: &str) -> Result<usize> {
         result.map_err(|err| Error::other(format!("failed to select decommission target pool for {bucket}/{object}: {err}")))
     }
@@ -2344,10 +2240,6 @@ impl ECStore {
         check_put_object_args(bucket, object)?;
 
         let mut opts = opts.clone();
-        let is_free_version = fi.tier_free_version();
-        if is_free_version {
-            opts.incl_free_versions = true;
-        }
         let bucket_incarnation_fence = if is_meta_bucketname(bucket) {
             None
         } else {
@@ -2385,10 +2277,6 @@ impl ECStore {
                 &object,
             )?
         };
-        #[cfg(test)]
-        if is_free_version {
-            pause_decommission_free_version_before_source_lock(bucket, logical_object).await;
-        }
         let _object_guards = self
             .acquire_data_movement_object_write_locks(bucket, &object, opts.src_pool_idx, idx, &mut opts)
             .await?;
@@ -2406,7 +2294,7 @@ impl ECStore {
                 versions
                     .versions
                     .iter()
-                    .find(|current| current.version_id == fi.version_id && current.tier_free_version() == is_free_version)
+                    .find(|current| current.version_id == fi.version_id && !current.tier_free_version())
             })
             .ok_or_else(|| to_object_err(StorageError::FileNotFound, vec![bucket, object.as_str()]))?;
         if !tiered_data_movement_source_matches(fi, current_source)? {
@@ -2421,40 +2309,24 @@ impl ECStore {
                 .get_available_pool_idx_excluding(bucket, &object, fi.size, opts.src_pool_idx)
                 .await;
             let target_pool_idx = resolve_data_movement_resume_target_pool(idx, resume_target_pool_idx, opts.src_pool_idx);
-            if is_free_version && target_pool_idx == opts.src_pool_idx {
-                return Err(Error::DiskFull);
-            }
-            let equivalent = if is_free_version {
-                self.has_equivalent_data_movement_tier_free_version(bucket, &object, &fi, &opts, target_pool_idx)
-                    .await?
-            } else {
-                self.has_equivalent_data_movement_tiered_object(bucket, &object, &fi, &opts, target_pool_idx)
-                    .await?
-            };
-            if equivalent {
-                return Ok(());
-            }
-
-            return Err(decommission_free_version_overwrite_error(bucket, &object, fi.version_id));
-        }
-
-        let result = if is_free_version {
             if self
-                .has_equivalent_data_movement_tier_free_version(bucket, &object, &fi, &opts, idx)
+                .has_equivalent_data_movement_tiered_object(bucket, &object, &fi, &opts, target_pool_idx)
                 .await?
             {
                 return Ok(());
             }
-            self.pools[idx]
-                .get_disks_by_key(&object)
-                .decommission_tier_free_version(bucket, &object, &fi, &opts)
-                .await
-        } else {
-            self.pools[idx]
-                .get_disks_by_key(&object)
-                .decommission_tiered_object(bucket, &object, &fi, &opts)
-                .await
-        };
+
+            return Err(StorageError::DataMovementOverwriteErr(
+                bucket.to_owned(),
+                object.to_owned(),
+                opts.version_id.clone().unwrap_or_default(),
+            ));
+        }
+
+        let result = self.pools[idx]
+            .get_disks_by_key(&object)
+            .decommission_tiered_object(bucket, &object, &fi, &opts)
+            .await;
         if matches!(result, Err(Error::PreconditionFailed)) {
             if self
                 .has_equivalent_data_movement_tiered_object(bucket, &object, &fi, &opts, idx)
@@ -3322,7 +3194,7 @@ impl ECStore {
 
         // Default return value
         let mut del_objects = vec![DeletedObject::default(); objects.len()];
-        let accounting = vec![None; objects.len()];
+        let mut accounting = vec![None; objects.len()];
 
         let mut del_errs = Vec::with_capacity(objects.len());
         for _ in 0..objects.len() {

crates/ecstore/src/store/rebalance/support.rs

@@ -271,7 +271,7 @@ pub(super) fn resolve_latest_object_info_candidates(
             .filter(|candidate| latest_candidate_mod_time(candidate) == Some(latest_mod_time))
             .collect::<Vec<_>>();
 
-        latest_candidates.sort_by_key(|candidate| std::cmp::Reverse(candidate.idx));
+        latest_candidates.sort_by(|left, right| right.idx.cmp(&left.idx));
 
         let Some(winner) = latest_candidates.first() else {
             return Err(Error::ErasureReadQuorum);

crates/filemeta/src/filemeta.rs

@@ -90,21 +90,6 @@ fn legacy_data_key_for_version(version_id: Option<Uuid>) -> Option<String> {
 pub const TRANSITION_COMPLETE: &str = "complete";
 pub const TRANSITION_PENDING: &str = "pending";
 
-/// xl.meta key marking a tier free-version record.
-///
-/// A free version is a delete-marker-shaped cleanup hint appended by
-/// [`MetaObject::delete_version`] when a version whose remote transition
-/// completed is removed from xl.meta; it carries the remote tier identity for
-/// an idempotent remote delete and is never a user-visible version
-/// (`num_versions` excludes it). While the record exists it is consumed by the
-/// lifecycle free-version recovery scan and the usage scanner, which re-enqueue
-/// the pending remote delete, and by heal metadata walks. On S3 and lifecycle
-/// delete paths the same obligation is also carried by a committed tier-journal
-/// entry; deletes without such an entry (for example a removed version whose
-/// transition state decodes as unknown) rely on this record alone until the
-/// worker removes it after a successful remote delete. Decommission preserves
-/// the record and its remote identity on the target pool before source cleanup
-/// — see docs/architecture/decommission-compatibility.md.
 pub const FREE_VERSION: &str = "free-version";
 
 pub const TRANSITION_STATUS: &str = "transition-status";
@@ -462,10 +447,6 @@ impl FileMeta {
         };
 
         if let Some(fidx) = existing_idx {
-            let existing = self.versions[fidx].parse_version_meta()?;
-            if existing.free_version() != version.free_version() {
-                return Err(Error::other("cannot replace a free version with a non-free version"));
-            }
             return self.set_idx(fidx, version);
         }
 

crates/filemeta/src/filemeta/version.rs

@@ -2725,15 +2725,6 @@ impl MetaObject {
         self.meta_sys.retain(|k, _| !k.starts_with("X-Amz-Restore"));
     }
 
-    /// Builds the free-version cleanup record appended when a transitioned
-    /// version is removed from xl.meta. The record keeps the remote tier
-    /// identity so the lifecycle worker can issue the idempotent remote delete
-    /// and only then remove the record; until then the recovery scan and the
-    /// usage scanner keep re-enqueueing it. S3 and lifecycle deletes also
-    /// persist a committed tier-journal entry for the same remote delete. The
-    /// decommission path copies this record unchanged before source cleanup,
-    /// including when the transition state is unknown — see
-    /// docs/architecture/decommission-compatibility.md.
     pub fn init_free_version(&self, fi: &FileInfo) -> Result<(FileMetaVersion, bool)> {
         if fi.skip_tier_free_version() {
             return Ok((FileMetaVersion::default(), false));

crates/heal/src/heal/task/heal_erasure_set.rs

@@ -231,10 +231,6 @@ impl HealTask {
                             "Heal erasure set format repair skipped because no format heal was required"
                         );
                     } else {
-                        let error = e;
-                        if error.is_recoverable_heal() {
-                            return Err(error);
-                        }
                         error!(
                             target: "rustfs::heal::task",
                             event = EVENT_HEAL_ERASURE_SET_RESULT,
@@ -243,7 +239,7 @@ impl HealTask {
                             task_id = %self.id,
                             set_disk_id,
                             result = "format_failed",
-                            error = %error,
+                            error = %e,
                             "Heal erasure set failed"
                         );
                         {
@@ -251,7 +247,7 @@ impl HealTask {
                             progress.update_progress(4, 4, 0, 0);
                         }
                         return Err(Error::TaskExecutionFailed {
-                            message: format!("Failed to heal disk format for {set_disk_id}: {error}"),
+                            message: format!("Failed to heal disk format for {set_disk_id}: {e}"),
                         });
                     }
                 } else {
@@ -288,9 +284,6 @@ impl HealTask {
             Err(Error::TaskCancelled) => return Err(Error::TaskCancelled),
             Err(Error::TaskTimeout) => return Err(Error::TaskTimeout),
             Err(e) => {
-                if e.is_recoverable_heal() {
-                    return Err(e);
-                }
                 error!(
                     target: "rustfs::heal::task",
                     event = EVENT_HEAL_ERASURE_SET_RESULT,

crates/heal/src/heal/task/tests.rs

@@ -547,7 +547,6 @@ struct MockStorage {
     heal_object_outcome: Mutex<Option<MockHealObjectOutcome>>,
     heal_object_outcomes: Mutex<HashMap<String, VecDeque<MockHealObjectOutcome>>>,
     format_no_heal_required: Mutex<bool>,
-    format_error: Mutex<Option<Error>>,
     global_format_calls: Mutex<u32>,
     replacement_format_calls: Mutex<Vec<(usize, usize, Vec<String>)>>,
     replacement_targets_ready: Mutex<bool>,
@@ -868,9 +867,6 @@ impl HealStorageAPI for MockStorage {
 
     async fn heal_format(&self, _dry_run: bool) -> Result<(HealResultItem, Option<Error>)> {
         *self.global_format_calls.lock().unwrap() += 1;
-        if let Some(error) = self.format_error.lock().unwrap().take() {
-            return Err(error);
-        }
         let no_heal_required = *self.format_no_heal_required.lock().unwrap();
         if no_heal_required {
             Ok((HealResultItem::default(), Some(Error::Storage(EcstoreError::NoHealRequired))))
@@ -2056,30 +2052,6 @@ async fn test_erasure_set_heal_continues_after_format_no_heal_required() {
     );
 }
 
-#[tokio::test]
-async fn erasure_set_format_slowdown_is_propagated() {
-    let storage = Arc::new(MockStorage {
-        format_error: Mutex::new(Some(Error::Storage(EcstoreError::SlowDown))),
-        ..Default::default()
-    });
-    let request = HealRequest::new(
-        HealType::ErasureSet {
-            buckets: Vec::new(),
-            set_disk_id: "pool_0_set_0".to_string(),
-        },
-        HealOptions::default(),
-        HealPriority::Normal,
-    );
-    let task = HealTask::from_request(request, storage);
-
-    let error = task
-        .execute()
-        .await
-        .expect_err("format SlowDown must remain recoverable for the task manager");
-
-    assert!(matches!(error, Error::Storage(EcstoreError::SlowDown)));
-}
-
 #[tokio::test]
 async fn erasure_set_bucket_prepass_failure_stops_before_object_heal() {
     let temp = TempDir::new().expect("temporary directory should be created");

crates/kms/tests/vault_ha_failover_live.rs

@@ -16,14 +16,14 @@
 //!
 //! `scripts/test/vault_ha_kms_live.sh` owns the official Vault containers and
 //! kills the active node while this test continuously decrypts through a
-//! surviving standby. KV2 and Transit must recover after the bounded circuit
-//! interval, use a bounded number of attempts, and leave the circuit and
-//! in-flight gauges at zero after a new leader is elected.
+//! surviving standby. KV2 and Transit requests must remain successful, use a
+//! bounded number of attempts, and leave the circuit and in-flight gauges at
+//! zero after a new leader is elected.
 
 use std::collections::HashMap;
 use std::path::{Path, PathBuf};
+use std::sync::Arc;
 use std::sync::atomic::{AtomicBool, AtomicU64, Ordering};
-use std::sync::{Arc, Mutex};
 use std::time::Duration;
 
 use metrics_util::MetricKind;
@@ -43,11 +43,6 @@ const OPERATION_ATTEMPTS: &str = "rustfs_kms_backend_operation_attempts";
 const IN_FLIGHT: &str = "rustfs_kms_backend_in_flight";
 const CIRCUIT_OPEN: &str = "rustfs_kms_backend_circuit_open";
 const MAX_ATTEMPTS: u32 = 10;
-const ATTEMPT_TIMEOUT: Duration = Duration::from_secs(2);
-const HEALTHY_PROGRESS_TIMEOUT: Duration = Duration::from_secs(20);
-// The circuit remains open for 30s after five failed attempts.
-const POST_FAILOVER_PROGRESS_TIMEOUT: Duration = Duration::from_secs(35);
-const FAILOVER_ERROR_POLL_INTERVAL: Duration = Duration::from_millis(100);
 
 type MetricEntry = (
     metrics_util::CompositeKey,
@@ -69,7 +64,7 @@ fn config(backend: KmsBackend, backend_config: BackendConfig) -> KmsConfig {
         backend,
         backend_config,
         allow_insecure_dev_defaults: true,
-        timeout: ATTEMPT_TIMEOUT,
+        timeout: Duration::from_secs(2),
         retry_attempts: MAX_ATTEMPTS,
         enable_cache: false,
         ..KmsConfig::default()
@@ -169,31 +164,14 @@ fn retryable_failures(snapshot: &[MetricEntry], operation: &str) -> u64 {
         .sum()
 }
 
-async fn wait_for_count(
-    counter: &AtomicU64,
-    failure: &Mutex<Option<String>>,
-    minimum: u64,
-    description: &str,
-    timeout: Duration,
-) {
-    tokio::time::timeout(timeout, async {
+async fn wait_for_count(counter: &AtomicU64, minimum: u64, description: &str) {
+    tokio::time::timeout(Duration::from_secs(20), async {
         while counter.load(Ordering::SeqCst) < minimum {
-            if let Some(error) = failure.lock().expect("decrypt failure lock poisoned").as_ref() {
-                panic!(
-                    "{description} worker failed after {} successful decrypts: {error}",
-                    counter.load(Ordering::SeqCst)
-                );
-            }
             tokio::time::sleep(Duration::from_millis(25)).await;
         }
     })
     .await
-    .unwrap_or_else(|_| {
-        panic!(
-            "timed out after {timeout:?} waiting for {description}: completed {}, expected {minimum}",
-            counter.load(Ordering::SeqCst)
-        )
-    });
+    .unwrap_or_else(|_| panic!("timed out waiting for {description}"));
 }
 
 async fn wait_for_file(path: &Path, description: &str) {
@@ -211,8 +189,7 @@ async fn decrypt_loop<B: KmsBackendTrait + Send + Sync + 'static>(
     request: DecryptRequest,
     expected: Vec<u8>,
     completed: Arc<AtomicU64>,
-    allow_failover_errors: Arc<AtomicBool>,
-    failure: Arc<Mutex<Option<String>>>,
+    failed: Arc<AtomicBool>,
     stop: CancellationToken,
 ) {
     while !stop.is_cancelled() {
@@ -220,18 +197,8 @@ async fn decrypt_loop<B: KmsBackendTrait + Send + Sync + 'static>(
             Ok(response) if response.plaintext == expected => {
                 completed.fetch_add(1, Ordering::SeqCst);
             }
-            Ok(_) => {
-                *failure.lock().expect("decrypt failure lock poisoned") =
-                    Some("decrypt returned unexpected plaintext".to_string());
-                return;
-            }
-            Err(rustfs_kms::KmsError::BackendError { .. } | rustfs_kms::KmsError::OperationTimedOut { .. })
-                if allow_failover_errors.load(Ordering::SeqCst) =>
-            {
-                tokio::time::sleep(FAILOVER_ERROR_POLL_INTERVAL).await;
-            }
-            Err(error) => {
-                *failure.lock().expect("decrypt failure lock poisoned") = Some(error.to_string());
+            Ok(_) | Err(_) => {
+                failed.store(true, Ordering::SeqCst);
                 return;
             }
         }
@@ -329,9 +296,7 @@ async fn exercise_failover(snapshotter: &Snapshotter) {
     );
 
     let stop = CancellationToken::new();
-    let allow_failover_errors = Arc::new(AtomicBool::new(false));
-    let kv2_failure = Arc::new(Mutex::new(None));
-    let transit_failure = Arc::new(Mutex::new(None));
+    let failed = Arc::new(AtomicBool::new(false));
     let kv2_completed = Arc::new(AtomicU64::new(0));
     let transit_completed = Arc::new(AtomicU64::new(0));
     let kv2_worker = tokio::spawn(decrypt_loop(
@@ -339,8 +304,7 @@ async fn exercise_failover(snapshotter: &Snapshotter) {
         kv2_request,
         kv2_data_key.plaintext_key,
         Arc::clone(&kv2_completed),
-        Arc::clone(&allow_failover_errors),
-        Arc::clone(&kv2_failure),
+        Arc::clone(&failed),
         stop.clone(),
     ));
     let transit_worker = tokio::spawn(decrypt_loop(
@@ -348,21 +312,12 @@ async fn exercise_failover(snapshotter: &Snapshotter) {
         transit_request,
         transit_data_key.plaintext_key,
         Arc::clone(&transit_completed),
-        Arc::clone(&allow_failover_errors),
-        Arc::clone(&transit_failure),
+        Arc::clone(&failed),
         stop.clone(),
     ));
 
-    wait_for_count(&kv2_completed, &kv2_failure, 2, "two healthy KV2 decrypts", HEALTHY_PROGRESS_TIMEOUT).await;
-    wait_for_count(
-        &transit_completed,
-        &transit_failure,
-        2,
-        "two healthy Transit decrypts",
-        HEALTHY_PROGRESS_TIMEOUT,
-    )
-    .await;
-    allow_failover_errors.store(true, Ordering::SeqCst);
+    wait_for_count(&kv2_completed, 2, "two healthy KV2 decrypts").await;
+    wait_for_count(&transit_completed, 2, "two healthy Transit decrypts").await;
     std::fs::write(&marker, b"ready").expect("publish failover readiness marker");
 
     wait_for_file(&elected, "the replacement Vault leader").await;
@@ -371,39 +326,18 @@ async fn exercise_failover(snapshotter: &Snapshotter) {
 
     let kv2_after_election = kv2_completed.load(Ordering::SeqCst) + 2;
     let transit_after_election = transit_completed.load(Ordering::SeqCst) + 2;
-    wait_for_count(
-        &kv2_completed,
-        &kv2_failure,
-        kv2_after_election,
-        "post-failover KV2 decrypts",
-        POST_FAILOVER_PROGRESS_TIMEOUT,
-    )
-    .await;
-    wait_for_count(
-        &transit_completed,
-        &transit_failure,
-        transit_after_election,
-        "post-failover Transit decrypts",
-        POST_FAILOVER_PROGRESS_TIMEOUT,
-    )
-    .await;
+    wait_for_count(&kv2_completed, kv2_after_election, "post-failover KV2 decrypts").await;
+    wait_for_count(&transit_completed, transit_after_election, "post-failover Transit decrypts").await;
 
     stop.cancel();
     kv2_worker.await.expect("KV2 decrypt worker must join");
     transit_worker.await.expect("Transit decrypt worker must join");
-    assert!(
-        kv2_failure.lock().expect("KV2 failure lock poisoned").is_none(),
-        "no KV2 decrypt may fail or return different plaintext"
-    );
-    assert!(
-        transit_failure.lock().expect("Transit failure lock poisoned").is_none(),
-        "no Transit decrypt may fail or return different plaintext"
-    );
+    assert!(!failed.load(Ordering::SeqCst), "no decrypt may fail or return different plaintext");
 }
 
 #[test]
 #[ignore = "requires a real three-node Vault Raft cluster; run scripts/test/vault_ha_kms_live.sh"]
-fn vault_raft_leader_failure_recovers_kv2_and_transit_decrypts() {
+fn vault_raft_leader_failure_preserves_kv2_and_transit_decrypts() {
     let recorder = DebuggingRecorder::new();
     let snapshotter = recorder.snapshotter();
     metrics::with_local_recorder(&recorder, || {
@@ -415,6 +349,11 @@ fn vault_raft_leader_failure_recovers_kv2_and_transit_decrypts() {
     });
     let snapshot = snapshotter.snapshot().into_vec();
 
+    assert_eq!(
+        counter_value(&snapshot, OPERATIONS_TOTAL, &[("outcome", "circuit_open")]),
+        0,
+        "a bounded leader election must not open the circuit"
+    );
     assert_eq!(
         counter_value(&snapshot, OPERATIONS_TOTAL, &[("outcome", "budget_exhausted")]),
         0,

crates/test-utils/src/lib.rs

@@ -245,18 +245,6 @@ impl TestECStoreEnvBuilder {
             .await
             .expect("build test ECStore");
 
-        // The production bootstrap only persists pool.bin from the elected
-        // first cluster node.  Test stores intentionally have no cluster
-        // election, but heal-format still requires that durable fence before
-        // it can write any disk format.  Materialize the validated topology
-        // here so the shared fixture models a ready single-node store.
-        let mut pool_meta = ecstore.pool_meta.read().await.clone();
-        pool_meta.dont_save = false;
-        pool_meta
-            .save(ecstore.pools.clone())
-            .await
-            .expect("persist test pool metadata");
-
         if self.init_bucket_metadata {
             let buckets_list = ecstore
                 .list_bucket(&BucketOptions {

deny.toml

@@ -43,6 +43,9 @@ allow-git = [
     # RustFS fork carrying presigned expiry and constant-time authentication fixes.
     # owner: rustfs-maintainers  review: 2026-10
     "https://github.com/rustfs/s3s.git",
+    # MiMalloc fork pinned for hotpath allocation counting support.
+    # owner: houseme  review: 2026-10
+    "https://github.com/xonatius/mimalloc_rust.git",
 ]
 
 [bans]

docs/architecture/decommission-compatibility.md

@@ -153,99 +153,6 @@ No migration step is required for these decisions because this note documents th
 current RustFS behavior. Changing either decision later requires an operator
 compatibility note and updated characterization tests.
 
-## Tier Free Versions During Decommission
-
-A tier free version is an internal xl.meta record (`rustfs_filemeta::FREE_VERSION`,
-flagged `XL_FLAG_FREE_VERSION`) shaped like a delete marker. It is created by
-`MetaObject::init_free_version` when a version whose remote transition completed is
-deleted locally: the visible version is removed and the record keeps the remote-tier
-identity (tier, object name, version id, state, destination id) needed for an
-idempotent remote delete. Free versions are not user-visible versions; `num_versions`
-and all listing/GET paths exclude them.
-
-### Lifecycle And Consumers
-
-Creation: any local delete that removes a version whose transition status is
-`complete` appends the record via `MetaObject::delete_version` →
-`init_free_version` (skipped only when `skip_tier_free_version` is set, as on
-data-movement copies). The same deletes also persist a durable tier-journal
-entry on every user-facing path: S3 single deletes (`execute_delete_object` →
-`delete_object_with_tier_delete_journal`), S3 batch deletes, lifecycle expiry,
-and lifecycle delete-all all prepare and commit a journal entry around the
-delete. A journal entry is omitted when the removed version's transition state
-decodes as `TransitionVersionState::Unknown`, or on internal journal-less
-delete paths that never touch transitioned user objects.
-
-Consumption while the record exists: the background recovery loop started by
-`init_background_expiry` (spawned by `spawn_tier_free_version_recovery_once`,
-enabled by default) scans disks for pending records and re-enqueues them; the
-usage scanner does the same; the lifecycle worker then deletes the remote tier
-object idempotently and only afterwards removes the local record. Heal walks
-include free-version records in metadata healing. Transition planning,
-replication, restore, GET, listings, and usage aggregation never depend on
-them.
-
-### Decommission Handling
-
-The exact decommission inventory loader (`load_file_info_versions_exact` via
-`get_all_file_info_versions`) keeps free-version records inline in `versions`.
-The migration loop handles them before lifecycle expiry and delete-marker
-shortcuts. It selects a target pool using the free-version-aware lookup, then
-writes the original free record to every target disk with the normal metadata
-write quorum. The free-version marker, local version id, transition identity,
-transition state, and destination id are preserved at the FileInfo/metadata
-boundary.
-
-The source record is physically removed only after the target write quorum has
-committed and the source cleanup preflight still matches the exact inventory.
-If the lifecycle worker has already completed the remote delete and removed the
-source record before decommission acquires the source lock, decommission records
-that identity as already consumed and treats the missing source record as safe.
-If target capacity, metadata validation, lock fencing, or quorum fails, the
-source record remains and the entry records `state = "free_version_retained"`
-with reason `tier_free_version_migration_failed`; the worker retries the
-operation on a later pass. A target record with the same version id is accepted
-only when its free-version identity matches; a conflicting ordinary version or
-different free record is an overwrite error. This makes retries idempotent and
-prevents a free record from replacing a user-visible version.
-
-### Reference-Audit Result
-
-After migration, user-facing GET/list/transition/replication/restore paths still
-exclude the record. Recovery, usage scanning, lifecycle tier cleanup, and heal
-continue to see it when they request free versions, so an unresolved remote
-delete remains actionable on the target pool. The committed tier journal remains
-an independent retry source where one exists; it is not used as a reason to drop
-the xl.meta record. In particular, `Unknown` transition state records are
-migrated unchanged rather than discarded: the lifecycle worker retains them if
-remote identity validation cannot make a delete request.
-
-Each migrated record emits `state = "free_version_migrated"` with reason
-`tier_free_version_migrated`. A record consumed before migration emits
-`state = "free_version_consumed"` with reason
-`tier_free_version_already_consumed`. Each failed record emits the retained state
-and failure reason above. The entry also emits a disposition summary with
-migrated, consumed, retained, and total counts. The final decommission sweep uses
-the exact loader, counts free records still present, and emits one retained
-record/reason for each unresolved free version before failing the sweep. This
-makes successful migration, completed cleanup, and retained cleanup obligations
-visible instead of silently omitting free records.
-
-No new S3-visible version or admin response field is needed: free versions remain
-internal and are never counted as user-visible versions. The structured
-`decommission_entry` events are the operational status surface for the
-free-version disposition; the existing decommission item/failed counters still
-report the enclosing object migration result.
-
-Regression guard:
-
-- `decommission_tier_free_version_preserves_remote_identity`
-- `decommission_tier_free_version_resume_requires_write_quorum`
-- `decommission_tier_free_version_commit_rejects_lost_fence`
-- `test_decommission_cleanup_preflight_accepts_migrated_free_version_consumed_from_source`
-- `decommission_entry_skips_cleanup_only_marker_when_free_version_is_present`
-- `decommission_entry_rejects_subquorum_free_version_conflict_and_retains_source`
-
 ## Regression Guard
 
 The queued multi-pool contract is guarded by:

docs/testing/ci-gates.md

@@ -1,149 +0,0 @@
-# CI gate matrix
-
-This file is the source of truth for which validation runs on each event, its
-configured wall-clock budget, and whether it can block a merge. Test taxonomy,
-naming, and nextest serialization rules remain in [README.md](README.md); e2e
-membership and counts remain in
-[e2e-suite-inventory.md](e2e-suite-inventory.md).
-
-The distinction between **required** and **report-only** is load-bearing:
-a failing job blocks a merge only when its exact check name is present in the
-live `main` ruleset. A workflow name, a `merge_group` trigger, or a red PR check
-does not make a job required by itself.
-
-## Required merge checks
-
-The live `main` ruleset (`6436880`) currently requires exactly these contexts:
-
-| Required context | Producer | Validation |
-|---|---|---|
-| `CLA Check` | `.github/workflows/cla.yml` | Contributor agreement |
-| `Quick Checks` | `.github/workflows/ci.yml` | Formatting and repository guard scripts |
-| `Test and Lint` | `.github/workflows/ci.yml` | Clippy, workspace nextest excluding `e2e_test`, doctests, and migration proofs |
-
-For pull requests limited to the paths excluded by the main CI workflow,
-`.github/workflows/ci-docs-only.yml` reports `Quick Checks` and
-`Test and Lint` under the same names. It runs the real quick checks and the
-planning-document guard; it does not claim that Rust compilation or runtime
-tests ran. Despite the workflow name, these paths also include selected deploy,
-workflow, and lock files.
-
-Verify the live rule rather than trusting this snapshot before changing merge
-policy:
-
-```bash
-gh api repos/rustfs/rustfs/rulesets/6436880 \
-  --jq '.rules[] | select(.type == "required_status_checks") | .parameters'
-```
-
-The ruleset currently has `strict_required_status_checks_policy=false`.
-`Continuous Integration` accepts `merge_group` events and runs `e2e-full` for
-them, but `End-to-End Tests (full merge gate)` is not currently a required
-context. Therefore the repository is prepared to test a merge-queue SHA, but
-the workflow alone does not prove that every merge passed that lane.
-
-## Pull request and merge matrix
-
-Budgets below are job `timeout-minutes`, not typical runtimes. “Report-only”
-means the result is visible and actionable but is not in the live required
-context list.
-
-| Event | Validation | Budget | Merge status | Reproduction |
-|---|---|---:|---|---|
-| PR, non-doc change | `Quick Checks` | 10 min | Required | `make pre-commit` (broader local umbrella) |
-| PR, non-doc change | `Test and Lint` | 90 min | Required | `cargo nextest run --profile ci --all --exclude e2e_test` |
-| PR, non-doc change | `Typos` | 10 min | Report-only | `typos` |
-| PR, non-doc change | `ILM Integration (serial)` | 90 min | Report-only | Use the exact command in `.github/workflows/ci.yml` |
-| PR, non-doc change | rio-v2 / swift / sftp test-and-lint variants | 90 min each | Report-only | `cargo nextest run` with the workflow's feature set |
-| PR, non-doc change | `Build RustFS Debug Binary` | 30 min | Report-only; prerequisite for black-box lanes | `cargo build -p rustfs --bins` |
-| PR, non-doc change | `io_uring Integration (real)` | 30 min | Report-only | `cargo test -p rustfs-ecstore --lib uring_ -- --test-threads=1 --nocapture` |
-| PR, non-doc change | `End-to-End Tests` (`e2e-smoke` plus `s3s-e2e`) | 30 min | Report-only | `cargo nextest run --profile e2e-smoke -p e2e_test`; then `./scripts/e2e-run.sh ./target/debug/rustfs <data-dir>` |
-| PR, non-doc change | `S3 Implemented Tests` | 60 min | Report-only | Build `rustfs`, then run `scripts/s3-tests/run.sh` with `DEPLOY_MODE=binary`, `TEST_MODE=single`, and `MAXFAIL=0` |
-| PR, non-doc change | `S3 Lifecycle Behavior Tests` | 30 min | Report-only | Use the accelerated scanner environment in `.github/workflows/ci.yml` with `scripts/s3-tests/run.sh` |
-| PR touching dependency or workflow inputs | Cargo Deny / Workflow Pin Report / Dependency Review | 20 / 5 / 30 min | Report-only | `cargo deny check`; `scripts/security/check_workflow_pins.sh` |
-| PR touching architecture rules or architecture docs | `Architecture Migration Rules` | 10 min | Report-only | `scripts/check_architecture_migration_rules.sh` |
-| PR touching Nix or workspace manifests | `Nix Build & Check` | 60 min | Report-only | `nix flake check` |
-| PR limited to main-CI-excluded paths | companion `Quick Checks` and `Test and Lint` | 10 min each | Required | `git diff --check`; `make doc-paths-check` when documentation paths changed |
-| `merge_group` | Standard CI plus `e2e-full` | 55 min for `e2e-full` | Standard required contexts only; `e2e-full` report-only | `cargo nextest run --profile e2e-full -p e2e_test` |
-| Push to `main` | Standard CI plus `e2e-full` | 55 min for `e2e-full` | Post-merge detection | Same as `merge_group` |
-| PR touching fuzz inputs or harness paths | Build plus five 60-second fuzz smoke targets | 60 min build; 30 min per target | Report-only | `MAX_TOTAL_TIME=60 ./scripts/fuzz/run.sh` |
-| PR touching selected ecstore disk/format paths | `Rename Safety` on Windows | 60 min | Report-only | Run the four `cargo test -p rustfs-ecstore --lib <filter>` commands in `windows-filesystem.yml` on Windows |
-
-The authoritative e2e filters live in `.config/nextest.toml`; extend a profile
-instead of adding a second ad-hoc selector. Before a profile runs,
-`scripts/check_test_wiring.py` compares its exact membership to the committed
-digest so a silent test drop fails closed.
-
-## Scheduled and manual validation
-
-Scheduled lanes are independent fault domains. They do not block a pull
-request, but their workflow-local gate can fail the run and scheduled failures
-are routed to the shared failure-issue action. The scheduled-validation
-watchdog and freshness workflow separately detect incomplete runs and missing
-schedules.
-
-| Cadence (UTC unless noted) | Workflow / validation | Budget | Verdict and artifacts | Reproduction |
-|---|---|---:|---|---|
-| Daily 02:17 | Fuzz: five nightly corpus targets | 60 min build; 60 min per target | Gate; corpus/crash artifacts, scheduled failure alert | `MAX_TOTAL_TIME=<seconds> ./scripts/fuzz/run.sh` |
-| Daily 03:17 | MinIO interop (EC + SSE read parity) | 40 min | Gate; scheduled failure alert | Dispatch `minio-interop.yml` or follow its pinned Docker fixture steps |
-| Daily 04:29 | Replication / cluster-fault / protocol e2e | 45 / 90 / 90 min | Three independent gates; JUnit, membership, and server logs | `cargo nextest run --profile e2e-repl-nightly -p e2e_test`; `--profile e2e-nightly`; `-j 1 --profile e2e-protocols` |
-| Daily 06:31 | Warp performance A/B | 180 min | Regression budget gate; A/B summaries and server logs | `bash scripts/run_hotpath_warp_abba.sh --help` |
-| Daily 00:07 Asia/Shanghai (16:07 UTC previous day) | Nightly GNU build and Vault lanes | 150 / 90 / 60 min | Build, live Vault, and HA failover gates | Use the commands and pinned Vault images in `nightly-gnu.yml` |
-| Daily 03:23 | Security Audit | 20 / 5 min, plus 30 min on PR dependency review | Cargo Deny and workflow-pin gates; scheduled failure alert | `cargo deny check`; `scripts/security/check_workflow_pins.sh` |
-| Daily 23:47 | Scheduled Validation Freshness | 10 min | Fails when a critical schedule was never created or is stale | Dispatch `scheduled-validation-freshness.yml` |
-| Sunday 00:11 | Full `Continuous Integration` matrix | Per-job budgets above | Weekly variant coverage, including dormant rio-v2 binary/e2e lanes | Dispatch `ci.yml` |
-| Sunday 01:13 | Seven-platform build matrix | 150 min per platform | Build/package integrity; scheduled failure alert | Dispatch `build.yml` with an exact platform set |
-| Sunday 02:19 | Ceph s3-tests full sweep: single and real four-node, four shards each | 180 min per shard | Compatibility gate; report, JUnit, exact node IDs, and server logs | `scripts/s3-tests/run.sh` against an existing single or distributed target |
-| Sunday 06:41 | Mint | 120 min | **Report-only by design**; per-suite PASS/FAIL/NA and raw `log.json` | Reproduce the pinned Docker sequence in `mint.yml` or dispatch it |
-| Sunday 07:43 | Workspace line coverage | 120 min | Report-only trend; lcov and JSON retained 90 days | `make coverage` |
-| Monthly, day 1 06:37 | Runner Hygiene | 15 min | Validates runner ephemerality; scheduled failure alert | Dispatch `runner-hygiene.yml` |
-
-Manual `workflow_dispatch` exists for the scheduled workflows above. Manual
-runs are debugging evidence and intentionally do not open scheduled-failure
-issues. A manual performance run may explicitly allow a known regression; that
-override must not be treated as an ordinary passing baseline.
-
-## Release validation
-
-Release validation is post-merge and tag-driven; it does not substitute for a
-pull-request gate.
-
-| Event | Validation | Budget | Result |
-|---|---|---:|---|
-| Push to `main` or weekly schedule | `Build and Release` platform matrix | 150 min per platform | Build artifacts for all selected targets; no release publication on a main push |
-| Valid release or preview tag | `Build and Release` plus asset checks | 150 min per platform | Draft release, checksummed assets, and publish step |
-| Successful non-preview release-tag build | Docker image build and image scan | 60 min build; 30 min scan | Multi-architecture images plus vulnerability report |
-| Successful release-tag build | DEB/RPM packaging | 30 min per architecture | Packages and checksum files uploaded to the release |
-| Successful non-preview release-tag build | Helm template test and package | 30 min build; 30 min publish | Versioned chart and repository index |
-
-Use an exact preview tag for end-to-end release rehearsal. Manual dispatches
-are backfill/debug paths and do not prove the automatic `workflow_run` chain.
-
-## Evidence requirements
-
-A green check is useful only when it proves the intended behavior ran:
-
-- Record the exact commit SHA and run URL.
-- Separate product failure from runner prerequisites, service readiness, and
-  cancellation. Repair the precondition, then rerun the exact workload.
-- Preserve membership manifests, JUnit, raw compatibility logs, seeds, and
-  server logs where the workflow provides them.
-- For a bug fix or a new fault checker, provide sensitivity evidence: the old
-  behavior or an intentional mutation must fail the new oracle, and the fixed
-  behavior must pass it.
-- Never promote a report-only lane to required from one green run. Require at
-  least 14 days and 30 representative pull requests with at least 99% complete
-  execution, then update the ruleset and this table together.
-
-## Change checklist
-
-Update this file in the same pull request when any of these change:
-
-- workflow triggers, job names, timeouts, or nextest profile ownership;
-- required status contexts or strict/merge-queue policy;
-- scheduled cadence, alert routing, artifact contract, or local reproduction;
-- report-only versus gating semantics.
-
-Do not copy per-module test counts here. Update
-[e2e-suite-inventory.md](e2e-suite-inventory.md) and its enforced membership
-digest instead.

rustfs/Cargo.toml

@@ -322,7 +322,7 @@ thiserror = { workspace = true }
 tracing.workspace = true
 url = { workspace = true }
 urlencoding = { workspace = true }
-uuid = { workspace = true, features = ["v4", "v5", "fast-rng", "macro-diagnostics"] }
+uuid = { workspace = true, features = ["v4", "fast-rng", "macro-diagnostics"] }
 zip = { workspace = true }
 libc = { workspace = true }
 rand = { workspace = true, features = ["serde"] }
@@ -336,16 +336,16 @@ opentelemetry = { workspace = true }
 tracing-opentelemetry = { workspace = true }
 # Data structures
 hashbrown = { workspace = true, features = ["serde", "rayon"] }
-rustfs-mimalloc = { workspace = true }
+mimalloc = { workspace = true }
 
 [target.'cfg(target_os = "linux")'.dependencies]
 libsystemd.workspace = true
 
 [target.'cfg(not(target_os = "windows"))'.dependencies]
-rustfs-mimalloc-sys.workspace = true
+libmimalloc-sys.workspace = true
 
 [dev-dependencies]
-uuid = { workspace = true, features = ["v4", "v5", "fast-rng", "macro-diagnostics"] }
+uuid = { workspace = true, features = ["v4", "fast-rng", "macro-diagnostics"] }
 serial_test = { workspace = true }
 tempfile = { workspace = true }
 aws-config = { workspace = true }

rustfs/src/admin/handlers/site_replication.rs

@@ -41,7 +41,7 @@ use crate::admin::storage_api::config::save_admin_config;
 use crate::admin::storage_api::contract::bucket::{
     BucketOperations, BucketOptions, DeleteBucketOptions, MakeBucketOptions, SRBucketDeleteOp,
 };
-use crate::admin::storage_api::error::{Error as StorageError, is_err_bucket_not_found};
+use crate::admin::storage_api::error::Error as StorageError;
 use crate::admin::storage_api::runtime::ECStore;
 use crate::admin::utils::{encode_compatible_admin_payload, read_compatible_admin_body};
 use crate::auth::constant_time_eq;
@@ -55,7 +55,6 @@ use crate::storage::storage_api::{
 use base64::Engine;
 use base64::engine::general_purpose::STANDARD as BASE64_STANDARD;
 use base64::engine::general_purpose::URL_SAFE_NO_PAD;
-use futures::StreamExt;
 use hmac::{Hmac, Mac};
 use http::header::{CONTENT_TYPE, HOST};
 use http::{HeaderMap, HeaderValue, Uri};
@@ -2097,18 +2096,6 @@ async fn remote_add_preflight_info(site: &PeerSite) -> S3Result<SiteReplicationA
             format!("invalid site replication metainfo from `{}`: {e}", site.endpoint),
         )
     })?;
-    if info.deployment_id.is_empty() {
-        // The peer will be tracked under a locally derived fallback ID
-        // (deployment_id_for_endpoint) instead of its real deployment ID.
-        warn!(
-            event = EVENT_ADMIN_SITE_REPLICATION_STATE,
-            component = LOG_COMPONENT_ADMIN,
-            subsystem = LOG_SUBSYSTEM_SITE_REPLICATION,
-            result = "peer_deployment_id_missing",
-            peer_endpoint = %site.endpoint,
-            "admin site replication state"
-        );
-    }
 
     let idp_body = send_peer_admin_get_request_with_client(
         &client,
@@ -2219,30 +2206,20 @@ fn site_replication_bootstrap_token(uri: &Uri) -> Option<String> {
     query_pairs(uri).get("bootstrapToken").cloned()
 }
 
-/// Query for a peer `make-with-versioning` bucket op. `versioningEnabled`
-/// always travels so the outbound query matches MinIO's site-replication
-/// make-bucket wire contract: MinIO's own create-bucket hook sends
-/// `versioningEnabled=true` on this op. RustFS's inbound handler
-/// force-enables versioning either way.
-fn make_with_versioning_bucket_op_path(bucket: &str, created_at: Option<&str>, lock_enabled: bool) -> String {
+fn bootstrap_bucket_make_op_path(bucket: &SRBucketInfo) -> String {
     let mut query = form_urlencoded::Serializer::new(String::new());
-    query.append_pair("bucket", bucket);
-    query.append_pair("operation", SITE_REPLICATION_BUCKET_OP_MAKE_WITH_VERSIONING);
-    query.append_pair("versioningEnabled", "true");
-    if let Some(created_at) = created_at {
-        query.append_pair("createdAt", created_at);
+    query.append_pair("bucket", &bucket.bucket);
+    query.append_pair("operation", "make-with-versioning");
+    if let Some(created_at) = bucket
+        .created_at
+        .and_then(|value| value.format(&time::format_description::well_known::Rfc3339).ok())
+    {
+        query.append_pair("createdAt", &created_at);
     }
-    if lock_enabled {
+    if bucket.object_lock_config.is_some() {
         query.append_pair("lockEnabled", "true");
     }
-    format!("{SITE_REPLICATION_PEER_BUCKET_OPS_PATH}?{}", query.finish())
-}
-
-fn bootstrap_bucket_make_op_path(bucket: &SRBucketInfo) -> String {
-    let created_at = bucket
-        .created_at
-        .and_then(|value| value.format(&time::format_description::well_known::Rfc3339).ok());
-    make_with_versioning_bucket_op_path(&bucket.bucket, created_at.as_deref(), bucket.object_lock_config.is_some())
+    format!("/rustfs/admin/v3/site-replication/peer/bucket-ops?{}", query.finish())
 }
 
 fn bootstrap_bucket_meta_item(bucket: &SRBucketInfo, item_type: &str, updated_at: Option<OffsetDateTime>) -> SRBucketMeta {
@@ -4269,7 +4246,16 @@ async fn broadcast_site_replication_make_bucket(
         .format(&time::format_description::well_known::Rfc3339)
         .unwrap_or_default();
 
-    let path = make_with_versioning_bucket_op_path(bucket, Some(&created_at), lock_enabled);
+    let path = {
+        let mut query = form_urlencoded::Serializer::new(String::new());
+        query.append_pair("bucket", bucket);
+        query.append_pair("operation", "make-with-versioning");
+        query.append_pair("createdAt", &created_at);
+        if lock_enabled {
+            query.append_pair("lockEnabled", "true");
+        }
+        format!("/rustfs/admin/v3/site-replication/peer/bucket-ops?{}", query.finish())
+    };
     let path = if let Some(token) = bootstrap_token {
         with_site_replication_bootstrap_token(&path, token)
     } else {
@@ -10220,25 +10206,13 @@ impl Operation for SiteReplicationStatusHandler {
     }
 }
 
-/// `POST /v3/site-replication/devnull` — peer link-check upload drain.
-/// MinIO streams multi-megabyte probe bodies here during site netperf link
-/// checks and expects an unbounded discard (its handler copies to io.Discard);
-/// buffering through the 1MB admin body cap turned any larger probe into a
-/// 400 and a false link failure. Stream and discard instead — no size cap.
-async fn drain_site_replication_devnull(mut input: Body) -> S3Result<()> {
-    while let Some(chunk) = input.next().await {
-        chunk.map_err(|e| s3_error!(InvalidRequest, "failed to read devnull stream: {}", e))?;
-    }
-    Ok(())
-}
-
 pub struct SiteReplicationDevNullHandler {}
 
 #[async_trait::async_trait]
 impl Operation for SiteReplicationDevNullHandler {
     async fn call(&self, req: S3Request<Body>, _params: Params<'_, '_>) -> S3Result<S3Response<(StatusCode, Body)>> {
         validate_site_replication_admin_request(&req, AdminAction::SiteReplicationOperationAction).await?;
-        drain_site_replication_devnull(req.input).await?;
+        let _ = read_plain_admin_body(req.input).await?;
         Ok(empty_response(StatusCode::NO_CONTENT))
     }
 }
@@ -10497,19 +10471,6 @@ impl Operation for SRPeerJoinHandler {
     }
 }
 
-/// Outcome of a peer-driven `purge-deleted-bucket` replay. A bucket that is
-/// already gone means the purge raced an earlier replay or a local delete —
-/// that is success — but any other failure must reach the sender like the
-/// sibling delete branches do: swallowing it answered 200 while the bucket
-/// survived on this site.
-fn purge_deleted_bucket_result(result: Result<(), StorageError>) -> S3Result<()> {
-    match result {
-        Ok(()) => Ok(()),
-        Err(err) if is_err_bucket_not_found(&err) => Ok(()),
-        Err(err) => Err(ApiError::from(err).into()),
-    }
-}
-
 pub struct SRPeerBucketOpsHandler {}
 
 #[async_trait::async_trait]
@@ -10609,18 +10570,16 @@ impl Operation for SRPeerBucketOpsHandler {
                     .map_err(ApiError::from)?;
             }
             "purge-deleted-bucket" => {
-                purge_deleted_bucket_result(
-                    store
-                        .delete_bucket(
-                            &bucket,
-                            &DeleteBucketOptions {
-                                force: true,
-                                srdelete_op: SRBucketDeleteOp::Purge,
-                                ..Default::default()
-                            },
-                        )
-                        .await,
-                )?;
+                let _ = store
+                    .delete_bucket(
+                        &bucket,
+                        &DeleteBucketOptions {
+                            force: true,
+                            srdelete_op: SRBucketDeleteOp::Purge,
+                            ..Default::default()
+                        },
+                    )
+                    .await;
             }
             _ => return Err(s3_error!(InvalidRequest, "unsupported site replication bucket operation")),
         }
@@ -13966,54 +13925,6 @@ mod tests {
         assert!(!query_flag(&uri, "missing"));
     }
 
-    /// A5 red-light: a `purge-deleted-bucket` replay must report success when
-    /// the bucket is already gone, and must propagate every other failure —
-    /// the swallowed error answered 200 while the bucket survived.
-    #[test]
-    fn test_purge_deleted_bucket_result_tolerates_only_missing_bucket() {
-        assert!(purge_deleted_bucket_result(Ok(())).is_ok());
-        assert!(purge_deleted_bucket_result(Err(StorageError::BucketNotFound("photos".to_string()))).is_ok());
-        assert!(purge_deleted_bucket_result(Err(StorageError::VolumeNotFound)).is_ok());
-        let err = purge_deleted_bucket_result(Err(StorageError::StorageFull))
-            .expect_err("non-not-found delete failures must propagate");
-        assert_ne!(*err.code(), S3ErrorCode::NoSuchBucket);
-    }
-
-    /// C5 red-light: the site-replication devnull drain must accept bodies
-    /// beyond the 1MB admin body cap — MinIO's link check streams large
-    /// probe bodies and treats a 400 as a broken link.
-    #[tokio::test]
-    async fn test_site_replication_devnull_drains_body_beyond_admin_cap() {
-        let body = Body::from(vec![0u8; MAX_ADMIN_REQUEST_BODY_SIZE + 1]);
-        drain_site_replication_devnull(body)
-            .await
-            .expect("devnull must drain bodies larger than the admin body cap");
-    }
-
-    /// A3 red-light: `versioningEnabled` must travel on every outbound
-    /// make-with-versioning bucket op so the query matches MinIO's
-    /// site-replication make-bucket wire contract (MinIO's own hook sends
-    /// `versioningEnabled=true` on this op).
-    #[test]
-    fn test_make_with_versioning_op_paths_send_versioning_enabled() {
-        let bucket = SRBucketInfo {
-            bucket: "photos".to_string(),
-            created_at: Some(OffsetDateTime::UNIX_EPOCH),
-            object_lock_config: Some(BASE64_STANDARD.encode("<ObjectLockConfiguration/>")),
-            ..Default::default()
-        };
-        let bootstrap = bootstrap_bucket_make_op_path(&bucket);
-        assert!(bootstrap.contains("operation=make-with-versioning"), "{bootstrap}");
-        assert!(bootstrap.contains("versioningEnabled=true"), "{bootstrap}");
-        assert!(bootstrap.contains("createdAt="), "{bootstrap}");
-        assert!(bootstrap.contains("lockEnabled=true"), "{bootstrap}");
-
-        // The broadcast path (create-bucket hook) shares the same builder.
-        let broadcast = make_with_versioning_bucket_op_path("photos", Some("1970-01-01T00:00:00Z"), false);
-        assert!(broadcast.contains("versioningEnabled=true"), "{broadcast}");
-        assert!(!broadcast.contains("lockEnabled"), "{broadcast}");
-    }
-
     #[tokio::test]
     #[serial]
     async fn test_add_bootstrap_scope_only_allows_expected_bucket_setup_until_guard_drops() {

rustfs/src/admin/site_replication_identity.rs

@@ -13,9 +13,9 @@
 // limitations under the License.
 
 use rustfs_madmin::{PeerInfo, SyncStatus};
-use std::collections::BTreeMap;
+use std::collections::{BTreeMap, hash_map::DefaultHasher};
+use std::hash::{Hash, Hasher};
 use url::Url;
-use uuid::Uuid;
 
 fn has_http_scheme(endpoint: &str) -> bool {
     endpoint.get(..7).is_some_and(|prefix| prefix.eq_ignore_ascii_case("http://"))
@@ -66,12 +66,10 @@ pub fn site_identity_key(endpoint: &str) -> String {
         .unwrap_or_else(|| trimmed.to_ascii_lowercase())
 }
 
-/// Fallback deployment ID for a peer that reported none. UUIDv5 over the
-/// canonical endpoint: the ID is persisted in site-replication state and
-/// broadcast to peers, so it must be identical across Rust toolchains
-/// (`DefaultHasher` is not) and across spellings of the same endpoint.
 pub fn deployment_id_for_endpoint(endpoint: &str) -> String {
-    Uuid::new_v5(&Uuid::NAMESPACE_URL, canonical_endpoint(endpoint).as_bytes()).to_string()
+    let mut hasher = DefaultHasher::new();
+    endpoint.hash(&mut hasher);
+    format!("{:016x}", hasher.finish())
 }
 
 pub fn same_identity_endpoint(left: &str, right: &str) -> bool {
@@ -176,23 +174,6 @@ mod tests {
         }
     }
 
-    /// B8 red-light: the fallback deployment ID must be a toolchain-stable
-    /// UUIDv5 over the canonical endpoint — `DefaultHasher` output is not
-    /// guaranteed stable across Rust releases, yet the ID is persisted in
-    /// site-replication state and broadcast to peers.
-    #[test]
-    fn deployment_id_for_endpoint_is_stable_uuid_v5_over_canonical_endpoint() {
-        let endpoint = "https://node-a.example.com:9000";
-        let id = deployment_id_for_endpoint(endpoint);
-        let parsed = uuid::Uuid::parse_str(&id).expect("fallback deployment ID must be a UUID");
-        assert_eq!(parsed.get_version_num(), 5, "fallback deployment ID must be UUIDv5");
-        // Deterministic for the same endpoint and for spelling variants that
-        // share a canonical form; distinct endpoints stay distinct.
-        assert_eq!(id, deployment_id_for_endpoint(endpoint));
-        assert_eq!(id, deployment_id_for_endpoint(" HTTPS://Node-A.Example.Com:9000/ "));
-        assert_ne!(id, deployment_id_for_endpoint("https://node-b.example.com:9000"));
-    }
-
     #[test]
     fn canonical_endpoint_accepts_case_insensitive_scheme() {
         assert_eq!(

rustfs/src/admin/storage_api.rs

@@ -51,7 +51,7 @@ mod ecstore_disk {
 }
 
 mod ecstore_error {
-    pub(crate) use crate::storage::storage_api::ecstore_error::{StorageError, is_err_bucket_not_found};
+    pub(crate) use crate::storage::storage_api::ecstore_error::StorageError;
 }
 
 #[allow(unused_imports)]
@@ -919,7 +919,6 @@ pub(crate) mod contract {
 }
 
 pub(crate) mod error {
-    pub(crate) use super::ecstore_error::is_err_bucket_not_found;
     pub(crate) use super::{Error, StorageError};
 }
 

rustfs/src/allocator_reclaim.rs

@@ -369,8 +369,14 @@ pub fn allocator_reclaim_controller_snapshot(ctx: &CancellationToken) -> Allocat
 }
 
 #[cfg(not(target_os = "windows"))]
+#[allow(unsafe_code)]
 fn collect_allocator_memory(force: bool) -> Result<(), String> {
-    rustfs_mimalloc::MiMalloc::collect(force);
+    // SAFETY: `mi_collect` is provided by the active global allocator backend
+    // on this target family. It is explicitly intended to reclaim retained
+    // pages/segments and does not require additional invariants from the caller.
+    unsafe {
+        libmimalloc_sys::mi_collect(force);
+    }
     Ok(())
 }
 

rustfs/src/main.rs

@@ -26,22 +26,22 @@ struct MiMallocAllocator;
 unsafe impl GlobalAlloc for MiMallocAllocator {
     unsafe fn alloc(&self, layout: Layout) -> *mut u8 {
         // SAFETY: the caller upholds GlobalAlloc's contract for layout.
-        unsafe { rustfs_mimalloc::MiMalloc.alloc(layout) }
+        unsafe { mimalloc::MiMalloc.alloc(layout) }
     }
 
     unsafe fn alloc_zeroed(&self, layout: Layout) -> *mut u8 {
         // SAFETY: the caller upholds GlobalAlloc's contract for layout.
-        unsafe { rustfs_mimalloc::MiMalloc.alloc_zeroed(layout) }
+        unsafe { mimalloc::MiMalloc.alloc_zeroed(layout) }
     }
 
     unsafe fn dealloc(&self, ptr: *mut u8, layout: Layout) {
         // SAFETY: ptr and layout came from this allocator and are forwarded unchanged.
-        unsafe { rustfs_mimalloc::MiMalloc.dealloc(ptr, layout) }
+        unsafe { mimalloc::MiMalloc.dealloc(ptr, layout) }
     }
 
     unsafe fn realloc(&self, ptr: *mut u8, layout: Layout, new_size: usize) -> *mut u8 {
         // SAFETY: ptr and layout came from this allocator and are forwarded unchanged.
-        unsafe { rustfs_mimalloc::MiMalloc.realloc(ptr, layout, new_size) }
+        unsafe { mimalloc::MiMalloc.realloc(ptr, layout, new_size) }
     }
 }
 
@@ -51,7 +51,7 @@ static GLOBAL: hotpath::CountingAllocator<MiMallocAllocator> = hotpath::Counting
 
 #[cfg(not(all(feature = "hotpath", feature = "hotpath-alloc")))]
 #[global_allocator]
-static GLOBAL: rustfs_mimalloc::MiMalloc = rustfs_mimalloc::MiMalloc;
+static GLOBAL: mimalloc::MiMalloc = mimalloc::MiMalloc;
 
 fn main() {
     let _hotpath_guard = hotpath::HotpathGuardBuilder::new("main").build();
@@ -71,9 +71,8 @@ mod tests {
         allocation.extend_from_slice(&[7_u8; 64]);
 
         assert_eq!(allocation.len(), 64);
-        let heap = rustfs_mimalloc::heap::Heap::main();
         // SAFETY: the live Vec pointer is valid to inspect for heap ownership.
-        assert!(unsafe { heap.contains(allocation.as_ptr()) });
+        assert!(unsafe { libmimalloc_sys::mi_is_in_heap_region(allocation.as_ptr().cast()) });
     }
 
     #[test]
@@ -86,13 +85,12 @@ mod tests {
         let layout = Layout::from_size_align(32, 8).expect("valid test allocation layout");
         let grown_layout = Layout::from_size_align(64, 8).expect("valid grown test allocation layout");
         let allocator = super::MiMallocAllocator;
-        let heap = rustfs_mimalloc::heap::Heap::main();
 
         // SAFETY: The pointer is checked for null before use and later released
         // through the same allocator with the corresponding layout.
         let ptr = unsafe { allocator.alloc_zeroed(layout) };
         assert!(!ptr.is_null());
-        assert!(unsafe { heap.contains(ptr) });
+        assert!(unsafe { libmimalloc_sys::mi_is_in_heap_region(ptr.cast()) });
         assert!(unsafe { std::slice::from_raw_parts(ptr, 32).iter().all(|byte| *byte == 0) });
 
         // SAFETY: `ptr` was allocated by `allocator` with `layout`; on failure
@@ -104,7 +102,7 @@ mod tests {
             panic!("mimalloc realloc failed in allocator smoke test");
         }
 
-        assert!(unsafe { heap.contains(grown_ptr) });
+        assert!(unsafe { libmimalloc_sys::mi_is_in_heap_region(grown_ptr.cast()) });
         // SAFETY: `grown_ptr` was reallocated by `allocator` and is released
         // with the matching grown layout.
         unsafe { allocator.dealloc(grown_ptr, grown_layout) };

rustfs/src/memory_observability.rs

@@ -17,7 +17,10 @@ use rustfs_io_metrics::{
     record_cpu_usage, record_memory_usage, record_process_memory_split,
 };
 use serde::Serialize;
+#[cfg(any(test, not(target_os = "windows")))]
 use serde_json::Value;
+#[cfg(not(target_os = "windows"))]
+use std::ffi::CStr;
 use std::path::Path;
 use std::sync::{Arc, Mutex, OnceLock};
 use std::time::Duration;
@@ -228,18 +231,7 @@ fn read_cgroup_memory_snapshot() -> Option<CgroupMemorySnapshot> {
     read_cgroup_v2().or_else(read_cgroup_v1)
 }
 
-fn read_allocator_memory_snapshot() -> Option<AllocatorMemorySnapshot> {
-    let json = rustfs_mimalloc::MiMalloc::stats_json();
-    if json.is_empty() {
-        return None;
-    }
-    let observation = parse_mimalloc_stats_json(&json)?;
-    Some(AllocatorMemorySnapshot {
-        backend: crate::allocator_reclaim::allocator_backend(),
-        observation,
-    })
-}
-
+#[cfg(any(test, not(target_os = "windows")))]
 fn numeric_json_value(value: &Value) -> Option<u64> {
     match value {
         Value::Number(number) => number
@@ -250,6 +242,7 @@ fn numeric_json_value(value: &Value) -> Option<u64> {
     }
 }
 
+#[cfg(any(test, not(target_os = "windows")))]
 fn numeric_json_field(value: &Value, field: &str) -> Option<u64> {
     match value {
         Value::Object(fields) => fields
@@ -261,6 +254,7 @@ fn numeric_json_field(value: &Value, field: &str) -> Option<u64> {
     }
 }
 
+#[cfg(any(test, not(target_os = "windows")))]
 fn mimalloc_stat_field(value: &Value, metric: &str, field: &str) -> Option<u64> {
     match value {
         Value::Object(fields) => {
@@ -277,10 +271,12 @@ fn mimalloc_stat_field(value: &Value, metric: &str, field: &str) -> Option<u64>
     }
 }
 
+#[cfg(any(test, not(target_os = "windows")))]
 fn mimalloc_stat_current(value: &Value, metric: &str) -> Option<u64> {
     mimalloc_stat_field(value, metric, "current")
 }
 
+#[cfg(any(test, not(target_os = "windows")))]
 fn mimalloc_stat_sum(value: &Value, metrics: &[&str], field: &str) -> Option<u64> {
     metrics
         .iter()
@@ -289,6 +285,7 @@ fn mimalloc_stat_sum(value: &Value, metrics: &[&str], field: &str) -> Option<u64
         .filter(|value| *value > 0)
 }
 
+#[cfg(any(test, not(target_os = "windows")))]
 fn parse_mimalloc_stats_json(stats_json: &str) -> Option<AllocatorMemoryObservation> {
     let value = serde_json::from_str::<Value>(stats_json).ok()?;
     let malloc_metrics = ["malloc_normal", "malloc_huge"];
@@ -315,6 +312,33 @@ fn parse_mimalloc_stats_json(stats_json: &str) -> Option<AllocatorMemoryObservat
     }
 }
 
+#[cfg(not(target_os = "windows"))]
+#[allow(unsafe_code)]
+fn read_allocator_memory_snapshot() -> Option<AllocatorMemorySnapshot> {
+    // SAFETY: `mi_stats_get_json` returns a null-terminated JSON buffer owned by
+    // mimalloc when called with a null input buffer. The mimalloc API requires
+    // freeing that buffer with `mi_free`; parsing finishes before the buffer is freed.
+    let observation = unsafe {
+        let stats_ptr = libmimalloc_sys::mi_stats_get_json(0, std::ptr::null_mut());
+        if stats_ptr.is_null() {
+            return None;
+        }
+
+        let observation = CStr::from_ptr(stats_ptr).to_str().ok().and_then(parse_mimalloc_stats_json);
+        libmimalloc_sys::mi_free(stats_ptr.cast());
+        observation?
+    };
+    Some(AllocatorMemorySnapshot {
+        backend: crate::allocator_reclaim::allocator_backend(),
+        observation,
+    })
+}
+
+#[cfg(target_os = "windows")]
+fn read_allocator_memory_snapshot() -> Option<AllocatorMemorySnapshot> {
+    None
+}
+
 fn configured_memory_observability_interval_secs() -> u64 {
     rustfs_utils::get_env_u64(ENV_MEMORY_OBSERVABILITY_INTERVAL_SECS, DEFAULT_MEMORY_OBSERVABILITY_INTERVAL_SECS).max(1)
 }
@@ -542,13 +566,6 @@ mod tests {
         assert_eq!(parse_mimalloc_stats_json(r#"{ "allocator": "unknown" }"#), None);
     }
 
-    #[test]
-    fn read_allocator_memory_snapshot_uses_mimalloc_stats_json() {
-        let snapshot = super::read_allocator_memory_snapshot();
-        #[cfg(not(target_os = "windows"))]
-        assert!(snapshot.is_some(), "allocator snapshot should be available on non-Windows");
-    }
-
     #[test]
     fn memory_observability_snapshot_reports_disabled_when_metrics_are_disabled() {
         let snapshot = build_memory_observability_status_snapshot(false, 15, false);

scripts/test/vault_ha_kms_live.sh

@@ -241,7 +241,7 @@ env \
   RUSTFS_TEST_VAULT_FAILOVER_MARKER="$MARKER" \
   RUSTFS_TEST_VAULT_OLD_LEADER="$OLD_LEADER" \
   cargo test -p rustfs-kms --test vault_ha_failover_live \
-    vault_raft_leader_failure_recovers_kv2_and_transit_decrypts -- \
+    vault_raft_leader_failure_preserves_kv2_and_transit_decrypts -- \
     --ignored --nocapture --test-threads=1 &
 TEST_PID=$!