Merge remote-tracking branch 'origin/main' into overtrue/fix-1912-heal-bucket-fence

# Conflicts: # crates/ecstore/src/store/heal.rs
fix(ecstore): fence bucket heal during decommission
2026-08-23 04:39:04 +00:00 · 2026-08-23 04:40:48 +08:00 · 2026-08-23 04:22:12 +08:00
28 changed files with 1130 additions and 360 deletions
@@ -1,2 +1,2 @@
-sha256-darwin=e78234828a8893b8bb0b27f57135edaff5066123cd592c38256be117c350d9d0
-sha256-linux=8084e3e013286bdea17ee18fbd44f28ade8ad0fe2726fa644835278819a41082
+sha256-darwin=9f767b37ed8b1c82da62ea441462d75487785c8086e56f08fb6f6cd89c6e2e52
+sha256-linux=fbdaf42b220958d4b1e8880e0f8b5a7992d38e21051bb60596dd4538424757d6
@@ -399,7 +399,7 @@ path = "junit.xml"
 #     `e2e-smoke` (20 fast) and `e2e-repl-nightly` (55 slow) lanes and reserves
 #     it for those, so e2e-full does not double-run it.
 #   * #[ignore]d tests — nextest skips them by default (no --run-ignored); the
-#     manual-localhost:9000 reliant tests are ci-13's migration.
+#     manual-localhost:9000 reliant/policy tests are ci-13's migration.
 #
 # Each e2e test spawns its own single-node rustfs server on a random port with
 # an isolated temp dir (crates/e2e_test/src/common.rs), so the set is
@@ -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):
@@ -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

@@ -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"
@@ -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" }
@@ -72,7 +72,7 @@ The reason string on each attribute is the classifier. Current classes:

 - **Needs a pre-started server** — `"requires running RustFS server at
  localhost:9000"` / `"Connects to existing rustfs server"`. These are the
-  `reliant/*` tests; start a server first (e.g.
+  `reliant/*` and `policy/test_runner` tests; start a server first (e.g.
  [`scripts/run_e2e_tests.sh`](../../scripts/run_e2e_tests.sh)) or use
  `--run-ignored`.
 - **Heavy / external tool** — `"Starts a rustfs server; enable when running
@@ -11,20 +11,29 @@ The tests cover the following AWS policy variable scenarios:
 3. **Variable concatenation** - Combining variables with static text like `prefix-${aws:username}-suffix`
 4. **Nested variables** - Complex nested variable patterns like `${${aws:username}-test}`
 5. **Deny scenarios** - Testing deny policies with variables
-6. **STS credentials** - Variable resolution inherited by temporary credentials

 ## Prerequisites

+- RustFS server binary
 - `awscurl` utility for admin API calls
 - AWS SDK for Rust (included in the project)

 ## Running Tests

+### Run All Policy Tests Using Unified Test Runner
+
+```bash
+# Run all policy tests with comprehensive reporting
+# Note: Requires a RustFS server running on localhost:9000
+cargo test -p e2e_test policy::test_runner::test_policy_full_suite -- --nocapture --ignored  --test-threads=1
+
+# Run only critical policy tests
+cargo test -p e2e_test policy::test_runner::test_policy_critical_suite -- --nocapture --ignored --test-threads=1
+```
+
 ### Run All Policy Tests

 ```bash
 # From the project root directory
-cargo test -p e2e_test policy:: -- --nocapture
-```
-
-Each test starts an isolated RustFS server on a dynamically allocated local port and cleans it up afterward.
+cargo test -p e2e_test policy:: -- --nocapture --ignored --test-threads=1
+```
@@ -18,3 +18,5 @@
 //! including single-value, multi-value, and nested variable scenarios.

 mod policy_variables_test;
+mod test_env;
+mod test_runner;
@@ -14,17 +14,14 @@

 //! Tests for AWS IAM policy variables with single-value, multi-value, and nested scenarios

-use crate::common::{
-    RustFSTestEnvironment, awscurl_delete, awscurl_put, build_test_s3_config, build_test_sts_client, init_logging,
-};
-use aws_sdk_s3::Client;
-use aws_sdk_s3::error::ProvideErrorMetadata;
+use crate::common::{awscurl_delete, awscurl_put, init_logging};
+use crate::policy::test_env::PolicyTestEnvironment;
 use aws_sdk_s3::primitives::ByteStream;
 use tracing::info;

 /// Helper function to create a regular user with given credentials
 async fn create_user(
-    env: &RustFSTestEnvironment,
+    env: &PolicyTestEnvironment,
    username: &str,
    password: &str,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
@@ -39,9 +36,20 @@ async fn create_user(
    Ok(())
 }

+/// Helper function to create an STS user with given credentials
+async fn create_sts_user(
+    env: &PolicyTestEnvironment,
+    username: &str,
+    password: &str,
+) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    // For STS, we create a regular user first, then use it to assume roles
+    create_user(env, username, password).await?;
+    Ok(())
+}
+
 /// Helper function to create and attach a policy
 async fn create_and_attach_policy(
-    env: &RustFSTestEnvironment,
+    env: &PolicyTestEnvironment,
    policy_name: &str,
    username: &str,
    policy_document: serde_json::Value,
@@ -62,9 +70,9 @@ async fn create_and_attach_policy(
 }

 /// Helper function to clean up test resources
-async fn cleanup_user_and_policy(env: &RustFSTestEnvironment, username: &str, policy_name: &str) {
+async fn cleanup_user_and_policy(env: &PolicyTestEnvironment, username: &str, policy_name: &str) {
    // Create admin client for cleanup
-    let admin_client = env.create_s3_client();
+    let admin_client = env.create_s3_client(&env.access_key, &env.secret_key);

    // Delete buckets that might have been created by this user
    let bucket_patterns = [
@@ -76,7 +84,7 @@ async fn cleanup_user_and_policy(env: &RustFSTestEnvironment, username: &str, po
        format!("{username}-test"),
        format!("{username}-sts-bucket"),
        format!("{username}-service-bucket"),
-        format!("{username}-private-bucket"),
+        "private-test-bucket".to_string(), // For deny test
    ];

    // Try to delete objects and buckets
@@ -113,18 +121,24 @@ async fn cleanup_user_and_policy(env: &RustFSTestEnvironment, username: &str, po

 /// Test AWS policy variables with single-value scenarios
 #[tokio::test(flavor = "multi_thread")]
+#[ignore = "Starts a rustfs server; enable when running full E2E"]
 pub async fn test_aws_policy_variables_single_value() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    test_aws_policy_variables_single_value_impl().await
+}
+
+/// Implementation function for single-value policy variables test
+pub async fn test_aws_policy_variables_single_value_impl() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    init_logging();
    info!("Starting AWS policy variables single-value test");

-    let mut env = RustFSTestEnvironment::new().await?;
-    env.start_rustfs_server(vec![]).await?;
+    let env = PolicyTestEnvironment::with_address("127.0.0.1:9000").await?;

    test_aws_policy_variables_single_value_impl_with_env(&env).await
 }

-async fn test_aws_policy_variables_single_value_impl_with_env(
-    env: &RustFSTestEnvironment,
+/// Implementation function for single-value policy variables test with shared environment
+pub async fn test_aws_policy_variables_single_value_impl_with_env(
+    env: &PolicyTestEnvironment,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    // Create test user
    let test_user = "testuser1";
@@ -184,7 +198,9 @@ async fn test_aws_policy_variables_single_value_impl_with_env(
    awscurl_put(&attach_policy_url, "", &env.access_key, &env.secret_key).await?;

    // Create S3 client for test user
-    let test_client = env.create_s3_client_with_credentials(test_user, test_password);
+    let test_client = env.create_s3_client(test_user, test_password);
+
+    tokio::time::sleep(std::time::Duration::from_millis(500)).await;

    // Test 1: User should be able to list buckets (allowed by policy)
    info!("Test 1: User listing buckets");
@@ -241,13 +257,11 @@ async fn test_aws_policy_variables_single_value_impl_with_env(
    // Test 6: User should NOT be able to create bucket NOT matching username pattern
    info!("Test 6: User attempting to create bucket NOT matching pattern");
    let other_bucket_name = "other-user-bucket";
-    let denied = test_client
-        .create_bucket()
-        .bucket(other_bucket_name)
-        .send()
-        .await
-        .expect_err("a bucket outside the username pattern must be denied");
-    assert_eq!(denied.as_service_error().and_then(ProvideErrorMetadata::code), Some("AccessDenied"));
+    let create_other_result = test_client.create_bucket().bucket(other_bucket_name).send().await;
+    if create_other_result.is_ok() {
+        cleanup().await;
+        return Err("User should NOT be able to create bucket NOT matching username pattern".into());
+    }

    // Cleanup
    info!("Cleaning up test resources");
@@ -259,18 +273,24 @@ async fn test_aws_policy_variables_single_value_impl_with_env(

 /// Test AWS policy variables with multi-value scenarios
 #[tokio::test(flavor = "multi_thread")]
+#[ignore = "Starts a rustfs server; enable when running full E2E"]
 pub async fn test_aws_policy_variables_multi_value() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    test_aws_policy_variables_multi_value_impl().await
+}
+
+/// Implementation function for multi-value policy variables test
+pub async fn test_aws_policy_variables_multi_value_impl() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    init_logging();
    info!("Starting AWS policy variables multi-value test");

-    let mut env = RustFSTestEnvironment::new().await?;
-    env.start_rustfs_server(vec![]).await?;
+    let env = PolicyTestEnvironment::with_address("127.0.0.1:9000").await?;

    test_aws_policy_variables_multi_value_impl_with_env(&env).await
 }

-async fn test_aws_policy_variables_multi_value_impl_with_env(
-    env: &RustFSTestEnvironment,
+/// Implementation function for multi-value policy variables test with shared environment
+pub async fn test_aws_policy_variables_multi_value_impl_with_env(
+    env: &PolicyTestEnvironment,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    // Create test user
    let test_user = "testuser2";
@@ -318,7 +338,7 @@ async fn test_aws_policy_variables_multi_value_impl_with_env(
    create_and_attach_policy(env, policy_name, test_user, policy_document).await?;

    // Create S3 client for test user
-    let test_client = env.create_s3_client_with_credentials(test_user, test_password);
+    let test_client = env.create_s3_client(test_user, test_password);

    // Test 1: User should be able to create buckets matching any of the multi-value patterns
    info!("Test 1: User creating first bucket matching multi-value pattern");
@@ -348,13 +368,11 @@ async fn test_aws_policy_variables_multi_value_impl_with_env(
    // Test 4: User should NOT be able to create bucket NOT matching any multi-value pattern
    info!("Test 4: User attempting to create bucket NOT matching any pattern");
    let other_bucket_name = format!("{test_user}-other-bucket");
-    let denied = test_client
-        .create_bucket()
-        .bucket(&other_bucket_name)
-        .send()
-        .await
-        .expect_err("a bucket outside all allowed patterns must be denied");
-    assert_eq!(denied.as_service_error().and_then(ProvideErrorMetadata::code), Some("AccessDenied"));
+    let create_other_result = test_client.create_bucket().bucket(&other_bucket_name).send().await;
+    if create_other_result.is_ok() {
+        cleanup().await;
+        return Err("User should NOT be able to create bucket NOT matching any multi-value pattern".into());
+    }

    // Test 5: User should be able to list objects in their allowed buckets
    info!("Test 5: User listing objects in allowed buckets");
@@ -380,18 +398,24 @@ async fn test_aws_policy_variables_multi_value_impl_with_env(

 /// Test AWS policy variables with variable concatenation
 #[tokio::test(flavor = "multi_thread")]
+#[ignore = "Starts a rustfs server; enable when running full E2E"]
 pub async fn test_aws_policy_variables_concatenation() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    test_aws_policy_variables_concatenation_impl().await
+}
+
+/// Implementation function for concatenation policy variables test
+pub async fn test_aws_policy_variables_concatenation_impl() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    init_logging();
    info!("Starting AWS policy variables concatenation test");

-    let mut env = RustFSTestEnvironment::new().await?;
-    env.start_rustfs_server(vec![]).await?;
+    let env = PolicyTestEnvironment::with_address("127.0.0.1:9000").await?;

    test_aws_policy_variables_concatenation_impl_with_env(&env).await
 }

-async fn test_aws_policy_variables_concatenation_impl_with_env(
-    env: &RustFSTestEnvironment,
+/// Implementation function for concatenation policy variables test with shared environment
+pub async fn test_aws_policy_variables_concatenation_impl_with_env(
+    env: &PolicyTestEnvironment,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    // Create test user
    let test_user = "testuser3";
@@ -431,7 +455,10 @@ async fn test_aws_policy_variables_concatenation_impl_with_env(
    create_and_attach_policy(env, policy_name, test_user, policy_document).await?;

    // Create S3 client for test user
-    let test_client = env.create_s3_client_with_credentials(test_user, test_password);
+    let test_client = env.create_s3_client(test_user, test_password);
+
+    // Add a small delay to allow policy to propagate
+    tokio::time::sleep(std::time::Duration::from_millis(500)).await;

    // Test: User should be able to create bucket matching concatenated pattern
    info!("Test: User creating bucket matching concatenated pattern");
@@ -460,30 +487,41 @@ async fn test_aws_policy_variables_concatenation_impl_with_env(

 /// Test AWS policy variables with nested scenarios
 #[tokio::test(flavor = "multi_thread")]
+#[ignore = "Starts a rustfs server; enable when running full E2E"]
 pub async fn test_aws_policy_variables_nested() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    test_aws_policy_variables_nested_impl().await
+}
+
+/// Implementation function for nested policy variables test
+pub async fn test_aws_policy_variables_nested_impl() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    init_logging();
    info!("Starting AWS policy variables nested test");

-    let mut env = RustFSTestEnvironment::new().await?;
-    env.start_rustfs_server(vec![]).await?;
+    let env = PolicyTestEnvironment::with_address("127.0.0.1:9000").await?;

    test_aws_policy_variables_nested_impl_with_env(&env).await
 }

 /// Test AWS policy variables with STS temporary credentials
 #[tokio::test(flavor = "multi_thread")]
+#[ignore = "Starts a rustfs server; enable when running full E2E"]
 pub async fn test_aws_policy_variables_sts() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    test_aws_policy_variables_sts_impl().await
+}
+
+/// Implementation function for STS policy variables test
+pub async fn test_aws_policy_variables_sts_impl() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    init_logging();
    info!("Starting AWS policy variables STS test");

-    let mut env = RustFSTestEnvironment::new().await?;
-    env.start_rustfs_server(vec![]).await?;
+    let env = PolicyTestEnvironment::with_address("127.0.0.1:9000").await?;

    test_aws_policy_variables_sts_impl_with_env(&env).await
 }

-async fn test_aws_policy_variables_nested_impl_with_env(
-    env: &RustFSTestEnvironment,
+/// Implementation function for nested policy variables test with shared environment
+pub async fn test_aws_policy_variables_nested_impl_with_env(
+    env: &PolicyTestEnvironment,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    // Create test user
    let test_user = "testuser4";
@@ -523,7 +561,10 @@ async fn test_aws_policy_variables_nested_impl_with_env(
    create_and_attach_policy(env, policy_name, test_user, policy_document).await?;

    // Create S3 client for test user
-    let test_client = env.create_s3_client_with_credentials(test_user, test_password);
+    let test_client = env.create_s3_client(test_user, test_password);
+
+    // Add a small delay to allow policy to propagate
+    tokio::time::sleep(std::time::Duration::from_millis(500)).await;

    // Test nested variable resolution
    info!("Test: Nested variable resolution");
@@ -540,14 +581,14 @@ async fn test_aws_policy_variables_nested_impl_with_env(
        return Err(format!("User should be able to create bucket with nested variable: {e}").into());
    }

-    // Verify a valid bucket name outside the resolved resource is denied.
-    let denied = test_client
-        .create_bucket()
-        .bucket("other-user-test")
-        .send()
-        .await
-        .expect_err("a bucket outside the resolved nested variable must be denied");
-    assert_eq!(denied.as_service_error().and_then(ProvideErrorMetadata::code), Some("AccessDenied"));
+    // Verify bucket creation fails with unresolved variable
+    let unresolved_bucket = format!("${{}}-test {test_user}");
+    let create_unresolved = test_client.create_bucket().bucket(&unresolved_bucket).send().await;
+
+    if create_unresolved.is_ok() {
+        cleanup().await;
+        return Err("User should NOT be able to create bucket with unresolved variable".into());
+    }

    // Cleanup
    info!("Cleaning up test resources");
@@ -557,8 +598,9 @@ async fn test_aws_policy_variables_nested_impl_with_env(
    Ok(())
 }

-async fn test_aws_policy_variables_sts_impl_with_env(
-    env: &RustFSTestEnvironment,
+/// Implementation function for STS policy variables test with shared environment
+pub async fn test_aws_policy_variables_sts_impl_with_env(
+    env: &PolicyTestEnvironment,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    // Create test user for STS
    let test_user = "testuser-sts";
@@ -570,7 +612,8 @@ async fn test_aws_policy_variables_sts_impl_with_env(
        cleanup_user_and_policy(env, test_user, policy_name).await;
    };

-    create_user(env, test_user, test_password).await?;
+    // Create STS user
+    create_sts_user(env, test_user, test_password).await?;

    // Create policy with STS-compatible variables
    let policy_document = serde_json::json!({
@@ -581,11 +624,6 @@ async fn test_aws_policy_variables_sts_impl_with_env(
                "Action": ["s3:ListAllMyBuckets"],
                "Resource": ["arn:aws:s3:::*"]
            },
-            {
-                "Effect": "Allow",
-                "Action": ["sts:AssumeRole"],
-                "Resource": ["arn:aws:s3:::*"]
-            },
            {
                "Effect": "Allow",
                "Action": ["s3:CreateBucket"],
@@ -593,12 +631,7 @@ async fn test_aws_policy_variables_sts_impl_with_env(
            },
            {
                "Effect": "Allow",
-                "Action": ["s3:ListBucket"],
-                "Resource": [format!("arn:aws:s3:::{}-sts-bucket", "${aws:username}")]
-            },
-            {
-                "Effect": "Allow",
-                "Action": ["s3:PutObject", "s3:GetObject"],
+                "Action": ["s3:ListBucket", "s3:PutObject", "s3:GetObject"],
                "Resource": [format!("arn:aws:s3:::{}-sts-bucket/*", "${aws:username}")]
            }
        ]
@@ -606,22 +639,11 @@ async fn test_aws_policy_variables_sts_impl_with_env(

    create_and_attach_policy(env, policy_name, test_user, policy_document).await?;

-    let assumed = build_test_sts_client(&env.url, test_user, test_password, None, "policy-variable-sts")
-        .assume_role()
-        .role_arn("arn:aws:iam::123456789012:role/policy-variable")
-        .role_session_name("policy-variable-e2e")
-        .send()
-        .await?;
-    let credentials = assumed
-        .credentials()
-        .ok_or("AssumeRole response should contain temporary credentials")?;
-    let test_client = Client::from_conf(build_test_s3_config(
-        &env.url,
-        credentials.access_key_id(),
-        credentials.secret_access_key(),
-        Some(credentials.session_token()),
-        "policy-variable-sts-session",
-    ));
+    // Create S3 client for test user
+    let test_client = env.create_s3_client(test_user, test_password);
+
+    // Add a small delay to allow policy to propagate
+    tokio::time::sleep(std::time::Duration::from_millis(500)).await;

    // Test: User should be able to create bucket matching STS pattern
    info!("Test: User creating bucket matching STS pattern");
@@ -677,18 +699,24 @@ async fn test_aws_policy_variables_sts_impl_with_env(

 /// Test AWS policy variables with deny scenarios
 #[tokio::test(flavor = "multi_thread")]
+#[ignore = "Starts a rustfs server; enable when running full E2E"]
 pub async fn test_aws_policy_variables_deny() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    test_aws_policy_variables_deny_impl().await
+}
+
+/// Implementation function for deny policy variables test
+pub async fn test_aws_policy_variables_deny_impl() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    init_logging();
    info!("Starting AWS policy variables deny test");

-    let mut env = RustFSTestEnvironment::new().await?;
-    env.start_rustfs_server(vec![]).await?;
+    let env = PolicyTestEnvironment::with_address("127.0.0.1:9000").await?;

    test_aws_policy_variables_deny_impl_with_env(&env).await
 }

-async fn test_aws_policy_variables_deny_impl_with_env(
-    env: &RustFSTestEnvironment,
+/// Implementation function for deny policy variables test with shared environment
+pub async fn test_aws_policy_variables_deny_impl_with_env(
+    env: &PolicyTestEnvironment,
 ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
    // Create test user
    let test_user = "testuser5";
@@ -731,7 +759,10 @@ async fn test_aws_policy_variables_deny_impl_with_env(
    create_and_attach_policy(env, policy_name, test_user, policy_document).await?;

    // Create S3 client for test user
-    let test_client = env.create_s3_client_with_credentials(test_user, test_password);
+    let test_client = env.create_s3_client(test_user, test_password);
+
+    // Add a small delay to allow policy to propagate
+    tokio::time::sleep(std::time::Duration::from_millis(500)).await;

    // Test 1: User should be able to create bucket matching username pattern
    info!("Test 1: User creating bucket matching username pattern");
@@ -744,14 +775,12 @@ async fn test_aws_policy_variables_deny_impl_with_env(

    // Test 2: User should NOT be able to create bucket with "private" in the name (deny rule)
    info!("Test 2: User attempting to create bucket with 'private' in name (should be denied)");
-    let private_bucket_name = format!("{test_user}-private-bucket");
-    let denied = test_client
-        .create_bucket()
-        .bucket(&private_bucket_name)
-        .send()
-        .await
-        .expect_err("the explicit deny must reject a matching bucket name");
-    assert_eq!(denied.as_service_error().and_then(ProvideErrorMetadata::code), Some("AccessDenied"));
+    let private_bucket_name = "private-test-bucket";
+    let create_private_result = test_client.create_bucket().bucket(private_bucket_name).send().await;
+    if create_private_result.is_ok() {
+        cleanup().await;
+        return Err("User should NOT be able to create bucket with 'private' in name due to deny rule".into());
+    }

    // Cleanup
    info!("Cleaning up test resources");
@@ -0,0 +1,100 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+//! Custom test environment for policy variables tests
+//!
+//! This module provides a custom test environment that doesn't automatically
+//! stop servers when destroyed, addressing the server stopping issue.
+
+use aws_sdk_s3::Client;
+use aws_sdk_s3::config::{Config, Credentials, Region};
+use std::net::TcpStream;
+use std::time::Duration;
+use tokio::time::sleep;
+use tracing::{info, warn};
+
+// Default credentials
+const DEFAULT_ACCESS_KEY: &str = "rustfsadmin";
+const DEFAULT_SECRET_KEY: &str = "rustfsadmin";
+
+/// Custom test environment that doesn't automatically stop servers
+pub struct PolicyTestEnvironment {
+    pub temp_dir: String,
+    pub address: String,
+    pub url: String,
+    pub access_key: String,
+    pub secret_key: String,
+}
+
+impl PolicyTestEnvironment {
+    /// Create a new test environment with specific address
+    /// This environment won't stop any server when dropped
+    pub async fn with_address(address: &str) -> Result<Self, Box<dyn std::error::Error + Send + Sync>> {
+        let temp_dir = format!("/tmp/rustfs_policy_test_{}", uuid::Uuid::new_v4());
+        tokio::fs::create_dir_all(&temp_dir).await?;
+
+        let url = format!("http://{address}");
+
+        Ok(Self {
+            temp_dir,
+            address: address.to_string(),
+            url,
+            access_key: DEFAULT_ACCESS_KEY.to_string(),
+            secret_key: DEFAULT_SECRET_KEY.to_string(),
+        })
+    }
+
+    /// Create an AWS S3 client configured for this RustFS instance
+    pub fn create_s3_client(&self, access_key: &str, secret_key: &str) -> Client {
+        let credentials = Credentials::new(access_key, secret_key, None, None, "policy-test");
+        let config = Config::builder()
+            .credentials_provider(credentials)
+            .region(Region::new("us-east-1"))
+            .endpoint_url(&self.url)
+            .force_path_style(true)
+            .behavior_version_latest()
+            .build();
+        Client::from_conf(config)
+    }
+
+    /// Wait for RustFS server to be ready by checking TCP connectivity
+    pub async fn wait_for_server_ready(&self) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+        info!("Waiting for RustFS server to be ready on {}", self.address);
+
+        for i in 0..30 {
+            if TcpStream::connect(&self.address).is_ok() {
+                info!("✅ RustFS server is ready after {} attempts", i + 1);
+                return Ok(());
+            }
+
+            if i == 29 {
+                return Err("RustFS server failed to become ready within 30 seconds".into());
+            }
+
+            sleep(Duration::from_secs(1)).await;
+        }
+
+        Ok(())
+    }
+}
+
+// Implement Drop trait that doesn't stop servers
+impl Drop for PolicyTestEnvironment {
+    fn drop(&mut self) {
+        // Clean up temp directory only, don't stop any server
+        if let Err(e) = std::fs::remove_dir_all(&self.temp_dir) {
+            warn!("Failed to clean up temp directory {}: {}", self.temp_dir, e);
+        }
+    }
+}
@@ -0,0 +1,230 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use crate::common::init_logging;
+use crate::policy::test_env::PolicyTestEnvironment;
+use std::time::Instant;
+use tokio::time::{Duration, sleep};
+use tracing::{error, info};
+
+/// Test case definition
+#[derive(Debug, Clone)]
+pub struct TestDefinition {
+    pub name: String,
+    pub is_critical: bool,
+}
+
+impl TestDefinition {
+    pub fn new(name: impl Into<String>, is_critical: bool) -> Self {
+        Self {
+            name: name.into(),
+            is_critical,
+        }
+    }
+}
+
+/// Test result
+#[derive(Debug, Clone)]
+pub struct TestResult {
+    pub test_name: String,
+    pub success: bool,
+    pub error_message: Option<String>,
+}
+
+impl TestResult {
+    pub fn success(test_name: String) -> Self {
+        Self {
+            test_name,
+            success: true,
+            error_message: None,
+        }
+    }
+
+    pub fn failure(test_name: String, error: String) -> Self {
+        Self {
+            test_name,
+            success: false,
+            error_message: Some(error),
+        }
+    }
+}
+
+/// Test suite configuration
+#[derive(Debug, Clone, Default)]
+pub struct TestSuiteConfig {
+    pub include_critical_only: bool,
+}
+
+/// Policy test suite
+pub struct PolicyTestSuite {
+    tests: Vec<TestDefinition>,
+    config: TestSuiteConfig,
+}
+
+impl PolicyTestSuite {
+    /// Create default test suite
+    pub fn new() -> Self {
+        let tests = vec![
+            TestDefinition::new("test_aws_policy_variables_single_value", true),
+            TestDefinition::new("test_aws_policy_variables_multi_value", true),
+            TestDefinition::new("test_aws_policy_variables_concatenation", true),
+            TestDefinition::new("test_aws_policy_variables_nested", true),
+            TestDefinition::new("test_aws_policy_variables_deny", true),
+            TestDefinition::new("test_aws_policy_variables_sts", true),
+        ];
+
+        Self {
+            tests,
+            config: TestSuiteConfig::default(),
+        }
+    }
+
+    /// Configure test suite
+    pub fn with_config(mut self, config: TestSuiteConfig) -> Self {
+        self.config = config;
+        self
+    }
+
+    /// Run test suite
+    pub async fn run_test_suite(&self) -> Vec<TestResult> {
+        init_logging();
+        info!("Starting Policy Variables test suite");
+
+        let start_time = Instant::now();
+        let mut results = Vec::new();
+
+        // Create test environment
+        let env = match PolicyTestEnvironment::with_address("127.0.0.1:9000").await {
+            Ok(env) => env,
+            Err(e) => {
+                error!("Failed to create test environment: {}", e);
+                return vec![TestResult::failure("env_creation".into(), e.to_string())];
+            }
+        };
+
+        // Wait for server to be ready
+        if env.wait_for_server_ready().await.is_err() {
+            error!("Server is not ready");
+            return vec![TestResult::failure("server_check".into(), "Server not ready".into())];
+        }
+
+        // Filter tests
+        let tests_to_run: Vec<&TestDefinition> = self
+            .tests
+            .iter()
+            .filter(|test| !self.config.include_critical_only || test.is_critical)
+            .collect();
+
+        info!("Scheduled {} tests", tests_to_run.len());
+
+        // Run tests
+        for (i, test_def) in tests_to_run.iter().enumerate() {
+            info!("Running test {}/{}: {}", i + 1, tests_to_run.len(), test_def.name);
+            let test_start = Instant::now();
+
+            let result = self.run_single_test(test_def, &env).await;
+            let test_duration = test_start.elapsed();
+
+            match result {
+                Ok(_) => {
+                    info!("Test passed: {} ({:.2}s)", test_def.name, test_duration.as_secs_f64());
+                    results.push(TestResult::success(test_def.name.clone()));
+                }
+                Err(e) => {
+                    error!("Test failed: {} ({:.2}s): {}", test_def.name, test_duration.as_secs_f64(), e);
+                    results.push(TestResult::failure(test_def.name.clone(), e.to_string()));
+                }
+            }
+
+            // Delay between tests to avoid resource conflicts
+            if i < tests_to_run.len() - 1 {
+                sleep(Duration::from_secs(2)).await;
+            }
+        }
+
+        // Print summary
+        self.print_summary(&results, start_time.elapsed());
+
+        results
+    }
+
+    /// Run a single test
+    async fn run_single_test(
+        &self,
+        test_def: &TestDefinition,
+        env: &PolicyTestEnvironment,
+    ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+        match test_def.name.as_str() {
+            "test_aws_policy_variables_single_value" => {
+                super::policy_variables_test::test_aws_policy_variables_single_value_impl_with_env(env).await
+            }
+            "test_aws_policy_variables_multi_value" => {
+                super::policy_variables_test::test_aws_policy_variables_multi_value_impl_with_env(env).await
+            }
+            "test_aws_policy_variables_concatenation" => {
+                super::policy_variables_test::test_aws_policy_variables_concatenation_impl_with_env(env).await
+            }
+            "test_aws_policy_variables_nested" => {
+                super::policy_variables_test::test_aws_policy_variables_nested_impl_with_env(env).await
+            }
+            "test_aws_policy_variables_deny" => {
+                super::policy_variables_test::test_aws_policy_variables_deny_impl_with_env(env).await
+            }
+            "test_aws_policy_variables_sts" => {
+                super::policy_variables_test::test_aws_policy_variables_sts_impl_with_env(env).await
+            }
+            _ => Err(format!("Test {} not implemented", test_def.name).into()),
+        }
+    }
+
+    /// Print test summary
+    fn print_summary(&self, results: &[TestResult], total_duration: Duration) {
+        info!("=== Test Suite Summary ===");
+        info!("Total duration: {:.2}s", total_duration.as_secs_f64());
+        info!("Total tests: {}", results.len());
+
+        let passed = results.iter().filter(|r| r.success).count();
+        let failed = results.len() - passed;
+        let success_rate = (passed as f64 / results.len() as f64) * 100.0;
+
+        info!("Passed: {} | Failed: {}", passed, failed);
+        info!("Success rate: {:.1}%", success_rate);
+
+        if failed > 0 {
+            error!("Failed tests:");
+            for result in results.iter().filter(|r| !r.success) {
+                error!("  - {}: {}", result.test_name, result.error_message.as_ref().unwrap());
+            }
+        }
+    }
+}
+
+/// Test suite
+#[tokio::test]
+#[ignore = "Connects to existing rustfs server"]
+async fn test_policy_critical_suite() -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    let config = TestSuiteConfig {
+        include_critical_only: true,
+    };
+    let suite = PolicyTestSuite::new().with_config(config);
+    let results = suite.run_test_suite().await;
+
+    let failed = results.iter().filter(|r| !r.success).count();
+    if failed > 0 {
+        return Err(format!("Critical tests failed: {failed} failures").into());
+    }
+
+    info!("All critical tests passed");
+    Ok(())
+}
@@ -454,13 +454,14 @@ pub mod rpc {
        AuthenticatedChannel, KMS_SIGNAL_SUBSYSTEM, LocalPeerS3Client, PEER_RESTDRY_RUN, PEER_RESTSIGNAL, PEER_RESTSUB_SYS,
        PeerRestClient, PeerS3Client, S3PeerSys, SERVICE_SIGNAL_REFRESH_CONFIG, SERVICE_SIGNAL_RELOAD_DYNAMIC,
        ScannerBucketListing, ScannerPeerActivity, TONIC_RPC_PREFIX, TonicInterceptor, build_put_file_auth_trailer,
-        check_and_record_signed_rpc_nonce, gen_signature_headers, gen_tonic_replay_scope_headers, gen_tonic_signature_headers,
-        gen_tonic_signature_interceptor, node_service_time_out_client, node_service_time_out_client_no_auth,
-        normalize_tonic_rpc_audience, set_tonic_canonical_body_digest, sign_ns_scanner_capability, sign_put_file_capability,
-        sign_tonic_rpc_response_proof, tonic_boot_epoch_challenge, tonic_boot_epoch_response_headers,
-        tonic_rpc_auth_failure_reason, verify_put_file_auth_trailer, verify_put_file_capability, verify_rpc_signature,
-        verify_tonic_boot_epoch_response, verify_tonic_canonical_body_digest, verify_tonic_mutation_body_digest,
-        verify_tonic_rpc_response_proof, verify_tonic_rpc_signature, verify_tonic_rpc_signature_with_bootstrap,
+        check_and_record_signed_rpc_nonce, decode_heal_bucket_rpc_options, encode_heal_bucket_rpc_options, gen_signature_headers,
+        gen_tonic_replay_scope_headers, gen_tonic_signature_headers, gen_tonic_signature_interceptor,
+        node_service_time_out_client, node_service_time_out_client_no_auth, normalize_tonic_rpc_audience,
+        set_tonic_canonical_body_digest, sign_ns_scanner_capability, sign_put_file_capability, sign_tonic_rpc_response_proof,
+        tonic_boot_epoch_challenge, tonic_boot_epoch_response_headers, tonic_rpc_auth_failure_reason,
+        verify_put_file_auth_trailer, verify_put_file_capability, verify_rpc_signature, verify_tonic_boot_epoch_response,
+        verify_tonic_canonical_body_digest, verify_tonic_mutation_body_digest, verify_tonic_rpc_response_proof,
+        verify_tonic_rpc_signature, verify_tonic_rpc_signature_with_bootstrap,
    };
 }

@@ -50,6 +50,9 @@ pub use peer_rest_client::{
    SERVICE_SIGNAL_RELOAD_DYNAMIC, ScannerPeerActivity,
 };
 pub(crate) use peer_s3_client::heal_bucket_local_on_disks;
-pub use peer_s3_client::{LocalPeerS3Client, PeerS3Client, S3PeerSys, ScannerBucketListing, ScannerSetBucketListing};
+pub use peer_s3_client::{
+    LocalPeerS3Client, PeerS3Client, S3PeerSys, ScannerBucketListing, ScannerSetBucketListing, decode_heal_bucket_rpc_options,
+    encode_heal_bucket_rpc_options,
+};
 pub use remote_disk::RemoteDisk;
 pub use remote_locker::RemoteClient;
@@ -18,6 +18,7 @@ use crate::cluster::rpc::client::{
    node_service_time_out_client,
 };
 use crate::cluster::rpc::set_tonic_mutation_body_digest;
+use crate::core::pools::PoolMeta;
 use crate::disk::error::DiskError;
 use crate::disk::error::{Error, Result};
 use crate::disk::error_reduce::{BUCKET_OP_IGNORED_ERRS, is_all_buckets_not_found, reduce_write_quorum_errs};
@@ -46,7 +47,12 @@ use std::sync::{
    Mutex as StdMutex,
    atomic::{AtomicBool, Ordering},
 };
-use std::{collections::HashMap, fmt::Debug, sync::Arc, time::Duration};
+use std::{
+    collections::{BTreeSet, HashMap},
+    fmt::Debug,
+    sync::Arc,
+    time::Duration,
+};
 #[cfg(test)]
 use tokio::sync::Notify;
 use tokio::{net::TcpStream, sync::RwLock, time};
@@ -99,6 +105,9 @@ impl DeleteBucketEmptyScanBarrier {
 #[cfg(test)]
 static DELETE_BUCKET_EMPTY_SCAN_BARRIER: StdMutex<Option<Arc<DeleteBucketEmptyScanBarrier>>> = StdMutex::new(None);

+#[cfg(test)]
+static HEAL_BUCKET_PRE_MUTATION_BARRIER: StdMutex<Option<Arc<DeleteBucketEmptyScanBarrier>>> = StdMutex::new(None);
+
 #[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
 enum HealBucketOperation {
    Make,
@@ -171,6 +180,15 @@ pub(crate) fn install_delete_bucket_empty_scan_barrier() -> Arc<DeleteBucketEmpt
    barrier
 }

+#[cfg(test)]
+fn install_heal_bucket_pre_mutation_barrier() -> Arc<DeleteBucketEmptyScanBarrier> {
+    let barrier = Arc::new(DeleteBucketEmptyScanBarrier::default());
+    *HEAL_BUCKET_PRE_MUTATION_BARRIER
+        .lock()
+        .expect("heal bucket mutation barrier lock should not be poisoned") = Some(barrier.clone());
+    barrier
+}
+
 #[cfg(test)]
 async fn pause_after_delete_bucket_empty_scan() {
    let barrier = DELETE_BUCKET_EMPTY_SCAN_BARRIER
@@ -182,6 +200,20 @@ async fn pause_after_delete_bucket_empty_scan() {
    }
 }

+#[cfg(test)]
+async fn pause_before_heal_bucket_volume_mutation() {
+    let barrier = HEAL_BUCKET_PRE_MUTATION_BARRIER
+        .lock()
+        .expect("heal bucket mutation barrier lock should not be poisoned")
+        .take();
+    if let Some(barrier) = barrier {
+        barrier.pause().await;
+    }
+}
+
+#[cfg(not(test))]
+async fn pause_before_heal_bucket_volume_mutation() {}
+
 #[derive(Clone, Debug)]
 pub struct ScannerBucketListing {
    pub buckets: Vec<BucketInfo>,
@@ -253,9 +285,41 @@ fn resolve_heal_bucket_mode(opts: &mut HealOpts, pool_errs: &[Option<Error>]) ->
    Ok(())
 }

+#[derive(serde::Serialize, serde::Deserialize)]
+#[serde(deny_unknown_fields)]
+struct HealBucketRpcEnvelope {
+    options: HealOpts,
+    #[serde(rename = "fencedPools", default)]
+    fenced_pools: Vec<usize>,
+}
+
+pub fn encode_heal_bucket_rpc_options(opts: HealOpts, fenced_pools: &[usize]) -> Result<String> {
+    serde_json::to_string(&HealBucketRpcEnvelope {
+        options: opts,
+        fenced_pools: fenced_pools.to_vec(),
+    })
+    .map_err(Into::into)
+}
+
+pub fn decode_heal_bucket_rpc_options(payload: &str) -> Result<(HealOpts, Vec<usize>)> {
+    match serde_json::from_str::<HealBucketRpcEnvelope>(payload) {
+        Ok(envelope) => Ok((envelope.options, envelope.fenced_pools)),
+        Err(envelope_err) => serde_json::from_str::<HealOpts>(payload)
+            .map(|options| (options, Vec::new()))
+            .map_err(|legacy_err| {
+                Error::other(format!(
+                    "decode heal bucket RPC options failed: envelope={envelope_err}; legacy={legacy_err}"
+                ))
+            }),
+    }
+}
+
 #[async_trait]
 pub trait PeerS3Client: Debug + Sync + Send + 'static {
    async fn heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem>;
+    async fn heal_bucket_with_fence(&self, bucket: &str, opts: &HealOpts, _fenced_pools: &[usize]) -> Result<HealResultItem> {
+        self.heal_bucket(bucket, opts).await
+    }
    async fn make_bucket(&self, bucket: &str, opts: &MakeBucketOptions) -> Result<()>;
    async fn list_bucket(&self, opts: &BucketOptions) -> Result<Vec<BucketInfo>>;
    async fn delete_bucket(&self, bucket: &str, opts: &DeleteBucketOptions) -> Result<()>;
@@ -309,6 +373,10 @@ impl S3PeerSys {

 impl S3PeerSys {
    pub async fn heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem> {
+        self.heal_bucket_with_fence(bucket, opts, &[]).await
+    }
+
+    pub async fn heal_bucket_with_fence(&self, bucket: &str, opts: &HealOpts, fenced_pools: &[usize]) -> Result<HealResultItem> {
        let mut opts = *opts;
        let mut futures = Vec::with_capacity(self.clients.len());
        for client in self.clients.iter() {
@@ -331,7 +399,7 @@ impl S3PeerSys {
            let opts_clone = opts;
            let heal_bucket_results_clone = heal_bucket_results.clone();
            futures.push(async move {
-                match client.heal_bucket(bucket, &opts_clone).await {
+                match client.heal_bucket_with_fence(bucket, &opts_clone, fenced_pools).await {
                    Ok(res) => {
                        heal_bucket_results_clone.write().await[idx] = res;
                        None
@@ -635,8 +703,18 @@ impl PeerS3Client for LocalPeerS3Client {
    }

    async fn heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem> {
+        self.heal_bucket_with_fence(bucket, opts, &[]).await
+    }
+
+    async fn heal_bucket_with_fence(&self, bucket: &str, opts: &HealOpts, fenced_pools: &[usize]) -> Result<HealResultItem> {
        let disks = self.local_disks_for_pools().await.into_iter().map(Some).collect();
-        heal_bucket_local_on_disks(bucket, opts, disks).await
+        let store = runtime_sources::object_store_handle().filter(|store| Arc::ptr_eq(&store.ctx, &self.instance_ctx));
+        #[cfg(not(test))]
+        if store.is_none() {
+            return Err(Error::other("bucket heal refused: pool metadata is unavailable for this instance"));
+        }
+        heal_bucket_local_on_disks_with_pool_meta(bucket, opts, disks, store.as_ref().map(|store| &store.pool_meta), fenced_pools)
+            .await
    }

    async fn list_bucket(&self, _opts: &BucketOptions) -> Result<Vec<BucketInfo>> {
@@ -1079,9 +1157,13 @@ impl PeerS3Client for RemotePeerS3Client {
    }

    async fn heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem> {
+        self.heal_bucket_with_fence(bucket, opts, &[]).await
+    }
+
+    async fn heal_bucket_with_fence(&self, bucket: &str, opts: &HealOpts, fenced_pools: &[usize]) -> Result<HealResultItem> {
        self.execute_with_timeout(
            || async {
-                let options: String = serde_json::to_string(opts)?;
+                let options = encode_heal_bucket_rpc_options(*opts, fenced_pools)?;
                let mut client = self.get_client().await?;
                let mut request = Request::new(HealBucketRequest {
                    bucket: bucket.to_string(),
@@ -1229,6 +1311,117 @@ pub(crate) async fn heal_bucket_local_on_disks(
    opts: &HealOpts,
    disks: Vec<Option<DiskStore>>,
 ) -> Result<HealResultItem> {
+    if let Some(store) = runtime_sources::object_store_handle() {
+        return heal_bucket_local_on_disks_with_pool_meta(bucket, opts, disks, Some(&store.pool_meta), &[]).await;
+    }
+
+    #[cfg(test)]
+    return heal_bucket_local_on_disks_with_pool_meta(bucket, opts, disks, None, &[]).await;
+
+    #[cfg(not(test))]
+    Err(Error::other("bucket heal refused: pool metadata is unavailable"))
+}
+
+fn disk_pool_index(disk: &DiskStore) -> Result<usize> {
+    usize::try_from(disk.endpoint().pool_idx)
+        .map_err(|_| Error::other(format!("invalid bucket-heal pool index {}", disk.endpoint().pool_idx)))
+}
+
+fn fenced_decommission_drive_state() -> DriveState {
+    DriveState::Unknown("skipped-decommission-suspended".to_string())
+}
+
+fn heal_bucket_fence_detail(fenced_pools: &BTreeSet<usize>) -> Option<String> {
+    if fenced_pools.is_empty() {
+        return None;
+    }
+    let pools = fenced_pools.iter().map(usize::to_string).collect::<Vec<_>>().join(", ");
+    Some(format!("skipped: bucket-volume heal fenced on decommission-suspended pool(s): {pools}"))
+}
+
+async fn snapshot_heal_bucket_fence(
+    disks: &[Option<DiskStore>],
+    pool_meta: Option<&RwLock<PoolMeta>>,
+    dispatch_fenced_pools: &[usize],
+) -> Result<(Vec<bool>, BTreeSet<usize>)> {
+    let mut fenced_disks = vec![false; disks.len()];
+    let mut fenced_pools = dispatch_fenced_pools.iter().copied().collect::<BTreeSet<_>>();
+    let pool_meta = match pool_meta {
+        Some(pool_meta) => Some(pool_meta.read().await),
+        None => None,
+    };
+    if let Some(pool_meta) = pool_meta.as_ref()
+        && let Some(pool_idx) = fenced_pools.iter().find(|pool_idx| **pool_idx >= pool_meta.pools.len())
+    {
+        return Err(Error::other(format!(
+            "bucket-heal dispatch fence pool index {pool_idx} is absent from {} pool metadata entries",
+            pool_meta.pools.len()
+        )));
+    }
+
+    for (disk_index, disk) in disks.iter().enumerate() {
+        let Some(disk) = disk else {
+            continue;
+        };
+        let pool_idx = disk_pool_index(disk)?;
+        if let Some(pool_meta) = pool_meta.as_ref() {
+            if pool_idx >= pool_meta.pools.len() {
+                return Err(Error::other(format!(
+                    "bucket-heal pool index {pool_idx} is absent from {} pool metadata entries",
+                    pool_meta.pools.len()
+                )));
+            }
+            if pool_meta.is_suspended(pool_idx) {
+                fenced_pools.insert(pool_idx);
+            }
+        }
+        if fenced_pools.contains(&pool_idx) {
+            fenced_disks[disk_index] = true;
+        }
+    }
+    Ok((fenced_disks, fenced_pools))
+}
+
+async fn run_heal_bucket_volume_mutation<F, Fut>(
+    disk: &DiskStore,
+    pool_meta: Option<&RwLock<PoolMeta>>,
+    operation: F,
+) -> Result<Option<usize>>
+where
+    F: FnOnce() -> Fut,
+    Fut: std::future::Future<Output = Result<()>>,
+{
+    let Some(pool_meta) = pool_meta else {
+        operation().await?;
+        return Ok(None);
+    };
+    let pool_idx = disk_pool_index(disk)?;
+    let pool_meta = pool_meta.read().await;
+    if pool_idx >= pool_meta.pools.len() {
+        return Err(Error::other(format!(
+            "bucket-heal pool index {pool_idx} is absent from {} pool metadata entries",
+            pool_meta.pools.len()
+        )));
+    }
+    if pool_meta.is_suspended(pool_idx) {
+        return Ok(Some(pool_idx));
+    }
+
+    // Keep the metadata read guard through the disk mutation so a decommission
+    // transition cannot pass between this state check and the destructive action.
+    operation().await?;
+    Ok(None)
+}
+
+async fn heal_bucket_local_on_disks_with_pool_meta(
+    bucket: &str,
+    opts: &HealOpts,
+    disks: Vec<Option<DiskStore>>,
+    pool_meta: Option<&RwLock<PoolMeta>>,
+    dispatch_fenced_pools: &[usize],
+) -> Result<HealResultItem> {
+    let (fenced_disks, mut fenced_pool_idxs) = snapshot_heal_bucket_fence(&disks, pool_meta, dispatch_fenced_pools).await?;
+    let fenced_disks = Arc::new(fenced_disks);
    let before_state = Arc::new(RwLock::new(vec![String::new(); disks.len()]));
    let after_state = Arc::new(RwLock::new(vec![String::new(); disks.len()]));

@@ -1238,7 +1431,14 @@ pub(crate) async fn heal_bucket_local_on_disks(
        let bucket = bucket.to_string();
        let bs_clone = before_state.clone();
        let as_clone = after_state.clone();
+        let fenced_disks = fenced_disks.clone();
        futures.push(async move {
+            if fenced_disks[index] {
+                let skipped = fenced_decommission_drive_state().to_string();
+                bs_clone.write().await[index] = skipped.clone();
+                as_clone.write().await[index] = skipped;
+                return None;
+            }
            let disk = match disk {
                Some(disk) => disk,
                None => {
@@ -1301,9 +1501,14 @@ pub(crate) async fn heal_bucket_local_on_disks(
                state: state.to_string(),
            });
        }
+        if let Some(detail) = heal_bucket_fence_detail(&fenced_pool_idxs) {
+            res.detail = detail;
+        }
        return Ok(res);
    }

+    pause_before_heal_bucket_volume_mutation().await;
+
    let mut operation_error = errs
        .iter()
        .filter_map(|err| match err {
@@ -1315,26 +1520,35 @@ pub(crate) async fn heal_bucket_local_on_disks(
    if opts.remove && !bucket.starts_with(disk::RUSTFS_META_BUCKET) && !is_all_buckets_not_found(&errs) {
        let mut futures = Vec::new();
        for (index, disk) in disks.iter().enumerate() {
-            if matches!(errs[index].as_ref(), Some(Error::DiskNotFound | Error::VolumeNotFound)) {
+            if fenced_disks[index] || matches!(errs[index].as_ref(), Some(Error::DiskNotFound | Error::VolumeNotFound)) {
                continue;
            }
            let Some(disk) = disk.clone() else {
                continue;
            };
            let bucket = bucket.to_string();
+            let mutation_disk = disk.clone();
            futures.push(async move {
-                if let Some(err) = injected_heal_bucket_operation_error(&bucket, index, HealBucketOperation::Delete) {
-                    return (index, Err(err));
-                }
-                (index, disk.delete_volume(&bucket, false).await)
+                let result = run_heal_bucket_volume_mutation(&disk, pool_meta, || async move {
+                    if let Some(err) = injected_heal_bucket_operation_error(&bucket, index, HealBucketOperation::Delete) {
+                        return Err(err);
+                    }
+                    mutation_disk.delete_volume(&bucket, false).await
+                })
+                .await;
+                (index, result)
            });
        }

        for (index, result) in join_all(futures).await {
            match result {
-                Ok(()) | Err(Error::VolumeNotFound) => {
+                Ok(None) | Err(Error::VolumeNotFound) => {
                    after_state.write().await[index] = DriveState::Missing.to_string();
                }
+                Ok(Some(pool_idx)) => {
+                    fenced_pool_idxs.insert(pool_idx);
+                    after_state.write().await[index] = fenced_decommission_drive_state().to_string();
+                }
                Err(Error::VolumeNotEmpty) => {
                    warn!(
                        bucket,
@@ -1365,30 +1579,38 @@ pub(crate) async fn heal_bucket_local_on_disks(
            let bs_clone = before_state.clone();
            futures.push(async move {
                if bs_clone.read().await[idx] == DriveState::Missing.to_string() {
-                    let Some(disk) = disk.as_ref() else {
-                        return (idx, Some(Error::DiskNotFound));
+                    let Some(disk) = disk else {
+                        return (idx, Err(Error::DiskNotFound));
                    };
-
-                    if let Some(err) = injected_heal_bucket_operation_error(&bucket, idx, HealBucketOperation::Make) {
-                        return (idx, Some(err));
-                    }
-                    match disk.make_volume(&bucket).await {
-                        Ok(()) | Err(Error::VolumeExists) => return (idx, None),
-                        Err(err) => return (idx, Some(err)),
-                    }
+                    let mutation_disk = disk.clone();
+                    let result = run_heal_bucket_volume_mutation(&disk, pool_meta, || async move {
+                        if let Some(err) = injected_heal_bucket_operation_error(&bucket, idx, HealBucketOperation::Make) {
+                            return Err(err);
+                        }
+                        match mutation_disk.make_volume(&bucket).await {
+                            Ok(()) | Err(Error::VolumeExists) => Ok(()),
+                            Err(err) => Err(err),
+                        }
+                    })
+                    .await;
+                    return (idx, result);
                }
-                (idx, None)
+                (idx, Ok(None))
            });
        }

        for (index, result) in join_all(futures).await {
            match result {
-                None => {
+                Ok(None) => {
                    if before_state.read().await[index] == DriveState::Missing.to_string() {
                        after_state.write().await[index] = DriveState::Ok.to_string();
                    }
                }
-                Some(err) => {
+                Ok(Some(pool_idx)) => {
+                    fenced_pool_idxs.insert(pool_idx);
+                    after_state.write().await[index] = fenced_decommission_drive_state().to_string();
+                }
+                Err(err) => {
                    after_state.write().await[index] = match &err {
                        Error::DiskNotFound => DriveState::Offline.to_string(),
                        _ => DriveState::Corrupt.to_string(),
@@ -1409,6 +1631,10 @@ pub(crate) async fn heal_bucket_local_on_disks(
        });
    }

+    if let Some(detail) = heal_bucket_fence_detail(&fenced_pool_idxs) {
+        res.detail = detail;
+    }
+
    match operation_error {
        Some(err) => Err(err),
        None => Ok(res),
@@ -1426,6 +1652,7 @@ async fn clone_drives() -> Vec<Option<DiskStore>> {
 #[cfg(test)]
 mod tests {
    use super::*;
+    use crate::core::pools::{PoolDecommissionInfo, PoolStatus};
    use crate::disk::WalkDirOptions;
    use crate::disk::disk_store::LocalDiskWrapper;
    use crate::disk::endpoint::Endpoint;
@@ -1599,6 +1826,23 @@ mod tests {
        disks
    }

+    fn heal_bucket_pool_meta(suspended_pool: Option<usize>) -> PoolMeta {
+        PoolMeta {
+            pools: (0..2)
+                .map(|pool_idx| PoolStatus {
+                    id: pool_idx,
+                    cmd_line: format!("pool-{pool_idx}"),
+                    last_update: ::time::OffsetDateTime::UNIX_EPOCH,
+                    decommission: (suspended_pool == Some(pool_idx)).then(|| PoolDecommissionInfo {
+                        start_time: Some(::time::OffsetDateTime::UNIX_EPOCH),
+                        ..Default::default()
+                    }),
+                })
+                .collect(),
+            ..Default::default()
+        }
+    }
+
    fn test_remote_peer(addr: &str) -> RemotePeerS3Client {
        RemotePeerS3Client {
            pools: Some(vec![0]),
@@ -1910,6 +2154,127 @@ mod tests {
        reset_local_disk_test_state().await;
    }

+    #[tokio::test]
+    #[serial]
+    async fn heal_bucket_rechecks_decommission_before_recreating_volume() {
+        reset_local_disk_test_state().await;
+
+        let temp_dir = TempDir::new().expect("create temp dir for bucket-heal fence regression");
+        let disks = init_test_local_disks_for_pools(&temp_dir, &[(0, 1), (1, 1)], "heal-bucket-mutation-fence").await;
+        let bucket = "fenced-recreate-bucket";
+        disks[0]
+            .make_volume(bucket)
+            .await
+            .expect("active pool should start with the bucket volume");
+
+        let pool_meta = Arc::new(RwLock::new(heal_bucket_pool_meta(None)));
+        let barrier = install_heal_bucket_pre_mutation_barrier();
+        let heal = tokio::spawn({
+            let disks = disks.clone();
+            let pool_meta = pool_meta.clone();
+            async move {
+                heal_bucket_local_on_disks_with_pool_meta(
+                    bucket,
+                    &HealOpts {
+                        recreate: true,
+                        ..Default::default()
+                    },
+                    disks.into_iter().map(Some).collect(),
+                    Some(pool_meta.as_ref()),
+                    &[],
+                )
+                .await
+            }
+        });
+
+        barrier.wait_until_paused().await;
+        pool_meta.write().await.pools[1].decommission = Some(PoolDecommissionInfo {
+            start_time: Some(::time::OffsetDateTime::UNIX_EPOCH),
+            ..Default::default()
+        });
+        barrier.release();
+
+        let result = heal
+            .await
+            .expect("bucket-heal task should join")
+            .expect("suspended pool should be reported as skipped");
+        assert!(result.detail.contains("skipped") && result.detail.contains('1'));
+        assert!(matches!(disks[1].stat_volume(bucket).await, Err(Error::VolumeNotFound)));
+
+        reset_local_disk_test_state().await;
+    }
+
+    #[tokio::test]
+    #[serial]
+    async fn heal_bucket_dispatch_fence_blocks_stale_active_peer_state() {
+        reset_local_disk_test_state().await;
+
+        let temp_dir = TempDir::new().expect("create temp dir for stale bucket-heal peer regression");
+        let disks = init_test_local_disks_for_pools(&temp_dir, &[(0, 1), (1, 1)], "heal-bucket-dispatch-fence").await;
+        let bucket = "dispatch-fenced-bucket";
+        disks[0]
+            .make_volume(bucket)
+            .await
+            .expect("active pool should start with the bucket volume");
+        let stale_pool_meta = RwLock::new(heal_bucket_pool_meta(None));
+
+        let result = heal_bucket_local_on_disks_with_pool_meta(
+            bucket,
+            &HealOpts {
+                recreate: true,
+                ..Default::default()
+            },
+            disks.iter().cloned().map(Some).collect(),
+            Some(&stale_pool_meta),
+            &[1],
+        )
+        .await
+        .expect("dispatch fence should override stale active peer metadata");
+
+        assert!(result.detail.contains("skipped") && result.detail.contains('1'));
+        assert!(matches!(disks[1].stat_volume(bucket).await, Err(Error::VolumeNotFound)));
+
+        reset_local_disk_test_state().await;
+    }
+
+    #[tokio::test]
+    #[serial]
+    async fn heal_bucket_keeps_suspended_pool_volume_on_remove() {
+        reset_local_disk_test_state().await;
+
+        let temp_dir = TempDir::new().expect("create temp dir for bucket-heal delete fence regression");
+        let disks = init_test_local_disks_for_pools(&temp_dir, &[(0, 1), (1, 1)], "heal-bucket-delete-fence").await;
+        let bucket = "fenced-remove-bucket";
+        for disk in &disks {
+            disk.make_volume(bucket)
+                .await
+                .expect("bucket volume should exist before heal");
+        }
+        let pool_meta = RwLock::new(heal_bucket_pool_meta(Some(1)));
+
+        let result = heal_bucket_local_on_disks_with_pool_meta(
+            bucket,
+            &HealOpts {
+                remove: true,
+                ..Default::default()
+            },
+            disks.iter().cloned().map(Some).collect(),
+            Some(&pool_meta),
+            &[],
+        )
+        .await
+        .expect("suspended pool should be skipped during bucket-volume removal");
+
+        assert!(result.detail.contains("skipped") && result.detail.contains('1'));
+        assert!(matches!(disks[0].stat_volume(bucket).await, Err(Error::VolumeNotFound)));
+        disks[1]
+            .stat_volume(bucket)
+            .await
+            .expect("suspended pool bucket volume must not be deleted");
+
+        reset_local_disk_test_state().await;
+    }
+
    #[tokio::test]
    #[serial]
    async fn heal_bucket_local_dry_run_reports_discovered_drive_states() {
@@ -2123,6 +2488,32 @@ mod tests {
        assert!(partial.recreate);
    }

+    #[test]
+    fn heal_bucket_rpc_envelope_preserves_legacy_compatibility_fail_closed() {
+        let opts = HealOpts {
+            recreate: true,
+            pool: Some(2),
+            ..Default::default()
+        };
+        let encoded = encode_heal_bucket_rpc_options(opts, &[1, 2]).expect("encode bucket-heal RPC envelope");
+
+        assert!(
+            serde_json::from_str::<HealOpts>(&encoded).is_err(),
+            "an old peer must reject the nested request instead of ignoring its dispatch fence"
+        );
+        let (decoded, fenced_pools) =
+            decode_heal_bucket_rpc_options(&encoded).expect("new peer should decode bucket-heal RPC envelope");
+        assert!(decoded.recreate);
+        assert_eq!(decoded.pool, Some(2));
+        assert_eq!(fenced_pools, vec![1, 2]);
+
+        let legacy = serde_json::to_string(&opts).expect("encode legacy HealOpts");
+        let (decoded, fenced_pools) = decode_heal_bucket_rpc_options(&legacy).expect("new peer should accept a legacy request");
+        assert!(decoded.recreate);
+        assert_eq!(decoded.pool, Some(2));
+        assert!(fenced_pools.is_empty());
+    }
+
    #[tokio::test]
    async fn test_make_bucket_reduces_quorum_by_pool_participants() {
        let peer_sys = S3PeerSys {
@@ -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 {
@@ -19,6 +19,7 @@ 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 std::collections::BTreeSet;
 use tracing::trace;

 const LOG_COMPONENT_ECSTORE: &str = "ecstore";
@@ -291,7 +292,45 @@ impl ECStore {

    #[instrument(skip(self))]
    pub(super) async fn handle_heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem> {
-        let res = self.peer_sys.heal_bucket(bucket, opts).await?;
+        let mut fenced_pools = BTreeSet::new();
+        {
+            let pool_meta = self.pool_meta.read().await;
+            fenced_pools.extend((0..pool_meta.pools.len()).filter(|pool_idx| pool_meta.is_suspended(*pool_idx)));
+            if let Some(pool_idx) = opts.pool {
+                if pool_idx >= pool_meta.pools.len() {
+                    return Err(invalid_heal_pool_index(pool_idx, pool_meta.pools.len()));
+                }
+                if pool_meta.is_suspended(pool_idx) {
+                    let complete = pool_meta.pools[pool_idx]
+                        .decommission
+                        .as_ref()
+                        .is_some_and(|decommission| decommission.complete);
+                    return Err(if complete {
+                        StorageError::InvalidArgument(
+                            "heal".to_string(),
+                            "pool".to_string(),
+                            format!("heal pool {pool_idx} has completed decommission"),
+                        )
+                    } else {
+                        Error::SlowDown
+                    });
+                }
+            }
+        }
+
+        let dispatch_fenced_pools = fenced_pools.iter().copied().collect::<Vec<_>>();
+        let mut res = self
+            .peer_sys
+            .heal_bucket_with_fence(bucket, opts, &dispatch_fenced_pools)
+            .await?;
+        {
+            let pool_meta = self.pool_meta.read().await;
+            fenced_pools.extend((0..pool_meta.pools.len()).filter(|pool_idx| pool_meta.is_suspended(*pool_idx)));
+        }
+        if !fenced_pools.is_empty() {
+            let pools = fenced_pools.iter().map(usize::to_string).collect::<Vec<_>>().join(", ");
+            res.detail = format!("skipped: bucket-volume heal fenced on decommission-suspended pool(s): {pools}");
+        }

        Ok(res)
    }
@@ -857,6 +896,64 @@ mod tests {
        }
    }

+    #[tokio::test]
+    async fn scoped_heal_bucket_blocks_before_dispatch_when_pool_is_suspended() {
+        let mut store = minimal_heal_store().await;
+        store.pool_meta = RwLock::new(PoolMeta {
+            pools: vec![
+                PoolStatus {
+                    id: 0,
+                    cmd_line: "pool-0".to_string(),
+                    last_update: OffsetDateTime::UNIX_EPOCH,
+                    decommission: None,
+                },
+                PoolStatus {
+                    id: 1,
+                    cmd_line: "pool-1".to_string(),
+                    last_update: OffsetDateTime::UNIX_EPOCH,
+                    decommission: Some(PoolDecommissionInfo {
+                        start_time: Some(OffsetDateTime::UNIX_EPOCH),
+                        ..Default::default()
+                    }),
+                },
+            ],
+            ..Default::default()
+        });
+
+        let err = store
+            .handle_heal_bucket(
+                "bucket",
+                &HealOpts {
+                    pool: Some(1),
+                    ..Default::default()
+                },
+            )
+            .await
+            .expect_err("suspended pool must be blocked before bucket-heal fan-out");
+        assert_eq!(err, Error::SlowDown);
+
+        store.pool_meta.write().await.pools[1]
+            .decommission
+            .as_mut()
+            .expect("decommission state should exist")
+            .complete = true;
+        let err = store
+            .handle_heal_bucket(
+                "bucket",
+                &HealOpts {
+                    pool: Some(1),
+                    ..Default::default()
+                },
+            )
+            .await
+            .expect_err("completed pool must remain fenced from bucket heal");
+        assert!(
+            matches!(err, StorageError::InvalidArgument(_, ref field, ref reason)
+                if field == "pool" && reason.contains("completed decommission")),
+            "unexpected completed-pool error: {err:?}"
+        );
+    }
+
    #[tokio::test]
    #[serial_test::serial]
    async fn unscoped_heal_object_suspended_owner_semantics() {
@@ -3194,7 +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() {
@@ -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);
@@ -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]
@@ -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.
@@ -80,7 +80,6 @@
 | object_lambda_test | 16 | 🌙 |
 | object_lock | 34 |  |
 | overwrite_cleanup_regression_test | 1 |  |
-| policy | 6 |  |
 | presigned_negative_test | 7 | ✅ |
 | protocols | 16 | 🌙 |
 | quota_test | 14 |  |
@@ -100,4 +99,4 @@
 | tls_hot_reload_test | 1 | ✅ |
 | version_id_regression_test | 10 | ✅ |

-**Total listed: 581 tests across 83 modules · PR smoke: 163 tests / 36 modules · merge/main full: 459 tests / 74 modules · nightly replication: 55 tests · nightly cluster faults: 28 tests / 7 modules · nightly protocols: 16 tests** · updated 2026-08-23.
+**Total listed: 575 tests across 82 modules · PR smoke: 163 tests / 36 modules · merge/main full: 453 tests / 73 modules · nightly replication: 55 tests · nightly cluster faults: 28 tests / 7 modules · nightly protocols: 16 tests** · updated 2026-08-23.
@@ -336,13 +336,13 @@ 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"] }
@@ -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(())
 }

@@ -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) };
@@ -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);
@@ -17,9 +17,8 @@ use crate::storage::storage_api::rpc_consumer::node_service::contract::bucket::{
    BucketOptions, DeleteBucketOptions, MakeBucketOptions,
 };
 use crate::storage::storage_api::rpc_consumer::node_service::{
-    DiskError, StoragePeerS3ClientExt as _, reload_bucket_metadata, remove_bucket_metadata,
+    DiskError, StoragePeerS3ClientExt as _, decode_heal_bucket_rpc_options, reload_bucket_metadata, remove_bucket_metadata,
 };
-use rustfs_common::heal_channel::HealOpts;
 use rustfs_protos::proto_gen::node_service::*;
 use tonic::{Request, Response, Status};
 use tracing::debug;
@@ -239,7 +238,7 @@ impl NodeService {
    ) -> Result<Response<HealBucketResponse>, Status> {
        debug!("heal bucket");
        let request = request.into_inner();
-        let options = match serde_json::from_str::<HealOpts>(&request.options) {
+        let (options, fenced_pools) = match decode_heal_bucket_rpc_options(&request.options) {
            Ok(options) => options,
            Err(err) => {
                return Ok(Response::new(HealBucketResponse {
@@ -249,7 +248,11 @@ impl NodeService {
            }
        };

-        match self.local_peer.heal_bucket(&request.bucket, &options).await {
+        match self
+            .local_peer
+            .heal_bucket_with_fence(&request.bucket, &options, &fenced_pools)
+            .await
+        {
            Ok(_) => Ok(Response::new(HealBucketResponse {
                success: true,
                error: None,
@@ -239,6 +239,7 @@ pub(crate) mod rpc_consumer {
    }

    pub(crate) mod node_service {
+        pub(crate) use super::super::ecstore_rpc::decode_heal_bucket_rpc_options;
        pub(crate) use super::super::storage_contracts::{
            SCANNER_ACTIVITY_LEGACY_PROTOCOL_VERSION, SCANNER_ACTIVITY_PREVIOUS_PROTOCOL_VERSION,
        };
@@ -520,10 +521,10 @@ pub(crate) mod ecstore_rpc {
    pub(crate) use rustfs_ecstore::api::rpc::{
        KMS_SIGNAL_SUBSYSTEM, LocalPeerS3Client, PEER_RESTDRY_RUN, PEER_RESTSIGNAL, PEER_RESTSUB_SYS, PeerRestClient,
        PeerS3Client, SERVICE_SIGNAL_REFRESH_CONFIG, SERVICE_SIGNAL_RELOAD_DYNAMIC, TONIC_RPC_PREFIX,
-        check_and_record_signed_rpc_nonce, normalize_tonic_rpc_audience, sign_ns_scanner_capability, sign_put_file_capability,
-        sign_tonic_rpc_response_proof, tonic_boot_epoch_challenge, tonic_boot_epoch_response_headers,
-        tonic_rpc_auth_failure_reason, verify_put_file_auth_trailer, verify_rpc_signature, verify_tonic_canonical_body_digest,
-        verify_tonic_mutation_body_digest, verify_tonic_rpc_signature_with_bootstrap,
+        check_and_record_signed_rpc_nonce, decode_heal_bucket_rpc_options, normalize_tonic_rpc_audience,
+        sign_ns_scanner_capability, sign_put_file_capability, sign_tonic_rpc_response_proof, tonic_boot_epoch_challenge,
+        tonic_boot_epoch_response_headers, tonic_rpc_auth_failure_reason, verify_put_file_auth_trailer, verify_rpc_signature,
+        verify_tonic_canonical_body_digest, verify_tonic_mutation_body_digest, verify_tonic_rpc_signature_with_bootstrap,
    };
    #[cfg(test)]
    pub(crate) use rustfs_ecstore::api::rpc::{
@@ -1412,6 +1413,12 @@ pub(crate) trait StoragePeerS3ClientExt {
        bucket: &str,
        opts: &rustfs_common::heal_channel::HealOpts,
    ) -> DiskResult<rustfs_madmin::heal_commands::HealResultItem>;
+    async fn heal_bucket_with_fence(
+        &self,
+        bucket: &str,
+        opts: &rustfs_common::heal_channel::HealOpts,
+        fenced_pools: &[usize],
+    ) -> DiskResult<rustfs_madmin::heal_commands::HealResultItem>;
    async fn make_bucket(&self, bucket: &str, opts: &contract::bucket::MakeBucketOptions) -> DiskResult<()>;
    async fn list_bucket(&self, opts: &contract::bucket::BucketOptions) -> DiskResult<Vec<contract::bucket::BucketInfo>>;
    async fn delete_bucket(&self, bucket: &str, opts: &contract::bucket::DeleteBucketOptions) -> DiskResult<()>;
@@ -1431,6 +1438,15 @@ impl StoragePeerS3ClientExt for LocalPeerS3Client {
        ecstore_rpc::PeerS3Client::heal_bucket(self, bucket, opts).await
    }

+    async fn heal_bucket_with_fence(
+        &self,
+        bucket: &str,
+        opts: &rustfs_common::heal_channel::HealOpts,
+        fenced_pools: &[usize],
+    ) -> DiskResult<rustfs_madmin::heal_commands::HealResultItem> {
+        ecstore_rpc::PeerS3Client::heal_bucket_with_fence(self, bucket, opts, fenced_pools).await
+    }
+
    async fn make_bucket(&self, bucket: &str, opts: &contract::bucket::MakeBucketOptions) -> DiskResult<()> {
        ecstore_rpc::PeerS3Client::make_bucket(self, bucket, opts).await
    }
Author	SHA1	Message	Date
overtrue	360f773f69	Merge remote-tracking branch 'origin/main' into overtrue/fix-1912-heal-bucket-fence # Conflicts: # crates/ecstore/src/store/heal.rs	2026-08-23 04:40:48 +08:00
overtrue	c422425431	fix(ecstore): fence bucket heal during decommission	2026-08-23 04:22:12 +08:00