fix(heal): cleanup consumed MRF replay journals

Do not retain Accepted or Merged replay intents as startup anchors after they have been handed to the heal manager. Only refused or still-pending replay records keep the journal on disk until a successor snapshot can persist them.

This keeps successor snapshots limited to the pending queue, which lets successful replay remove both authoritative and legacy journal paths and restores the crash-boundary tests around successor flush.

Co-Authored-By: heihutu <heihutu@gmail.com>

Co-Authored-By: zhi22915 <qiuzgang@gmail.com>
(cherry picked from commit d5b8f49c9d)

Cargo.lock

@@ -9906,7 +9906,6 @@ dependencies = [
  "regex",
  "rmp",
  "rmp-serde",
- "rustfs-config",
  "rustfs-utils",
  "s3s",
  "serde",

crates/config/src/constants/api.rs

@@ -100,10 +100,5 @@ pub const DEFAULT_API_MAX_CONNECTIONS: usize = 0;
 /// Example: RUSTFS_API_OBJECT_MAX_VERSIONS=50000
 pub const ENV_API_OBJECT_MAX_VERSIONS: &str = "RUSTFS_API_OBJECT_MAX_VERSIONS";
 
-/// Default and maximum accepted value for `RUSTFS_API_OBJECT_MAX_VERSIONS`.
-#[cfg(target_pointer_width = "64")]
-pub const DEFAULT_API_OBJECT_MAX_VERSIONS: usize = 9_223_372_036_854_775_807;
-
-/// Default and maximum accepted value for `RUSTFS_API_OBJECT_MAX_VERSIONS`.
-#[cfg(not(target_pointer_width = "64"))]
-pub const DEFAULT_API_OBJECT_MAX_VERSIONS: usize = usize::MAX;
+/// Default for `RUSTFS_API_OBJECT_MAX_VERSIONS`.
+pub const DEFAULT_API_OBJECT_MAX_VERSIONS: u64 = 9_223_372_036_854_775_807;

crates/filemeta/Cargo.toml

@@ -44,7 +44,6 @@ tokio = { workspace = true, features = ["io-util", "macros", "sync", "fs", "rt-m
 xxhash-rust = { workspace = true, features = ["xxh64", "xxh3"] }
 bytes = { workspace = true, features = ["serde"] }
 rustfs-utils = { workspace = true, features = ["hash", "http"] }
-rustfs-config = { workspace = true, features = ["constants"] }
 byteorder = { workspace = true }
 tracing.workspace = true
 thiserror.workspace = true

crates/filemeta/src/filemeta.rs

@@ -70,8 +70,12 @@ const _XL_FLAG_INLINE_DATA: u8 = 1 << 2;
 const META_DATA_READ_DEFAULT: usize = 4 << 10;
 const MSGP_UINT32_SIZE: usize = 5;
 
-/// Default max object versions per object.
-pub const DEFAULT_OBJECT_MAX_VERSIONS: usize = rustfs_config::DEFAULT_API_OBJECT_MAX_VERSIONS;
+/// Default max object versions per object, aligned with MinIO's default.
+pub const DEFAULT_OBJECT_MAX_VERSIONS: usize = if usize::BITS >= 64 {
+    9_223_372_036_854_775_807
+} else {
+    usize::MAX
+};
 
 static OBJECT_MAX_VERSIONS: AtomicUsize = AtomicUsize::new(DEFAULT_OBJECT_MAX_VERSIONS);
 

docs/testing/ci-gates.md

@@ -228,6 +228,19 @@ these. The external `rustfs/auto-testing` functional workflows propagate suite
 failures. Their workflow status does not establish this registry's required
 case coverage, build provenance, or object-level oracles.
 
+For automation, `--check-scanner-heal-release "$RUN_DIR"` emits one compact
+JSON decision and exits nonzero while blocked. `verified_cases` contains only
+cases that pass the complete receipt, build provenance, nextest/JUnit and real
+oracle checks; `rejected_cases` names registered cases that do not, and
+`pending_gates` names the unimplemented release requirements. Approval requires
+every registered case to verify, `pending_gates` to be empty, and a future
+registry schema capable of representing the complete release matrix. Schema 1
+is deliberately marked `release_schema_capable: false`: it models only the
+single-version, unversioned-object restart/crash cases and cannot represent
+mixed-version, rollback, EC8+4 or performance evidence. A focused run,
+synthetic harness, compile-only result, skipped/retried test, ordinary CI
+success, or removal of pending text therefore cannot become a release approval.
+
 Run parser/receipt regressions with
 `scripts/python_bin.sh scripts/check_test_wiring.py --self-test`. Those fixtures
 validate the checker only and produce no runtime or performance evidence.

rustfs/src/startup_runtime.rs

@@ -17,7 +17,7 @@ use crate::{
     startup_runtime_hooks::{init_profiling_runtime, install_default_crypto_provider, log_startup_runtime_diagnostics},
     startup_tls_material::init_outbound_tls_material,
 };
-use rustfs_config::{DEFAULT_API_OBJECT_MAX_VERSIONS, ENV_API_OBJECT_MAX_VERSIONS};
+use rustfs_config::ENV_API_OBJECT_MAX_VERSIONS;
 use rustfs_utils::EnvParseOutcome;
 use std::io::{Error, Result};
 
@@ -32,8 +32,13 @@ pub(crate) async fn init_startup_runtime_foundation(config: &Config) -> Result<(
 
 fn init_object_max_versions_config() -> Result<()> {
     let limit = match rustfs_utils::get_env_parse_outcome::<u64>(ENV_API_OBJECT_MAX_VERSIONS) {
-        EnvParseOutcome::Absent => DEFAULT_API_OBJECT_MAX_VERSIONS,
-        EnvParseOutcome::Invalid => return Err(object_max_versions_config_error()),
+        EnvParseOutcome::Absent => rustfs_filemeta::DEFAULT_OBJECT_MAX_VERSIONS,
+        EnvParseOutcome::Invalid => {
+            return Err(Error::other(format!(
+                "{ENV_API_OBJECT_MAX_VERSIONS} must be a positive integer no greater than {}",
+                usize::MAX
+            )));
+        }
         EnvParseOutcome::Parsed(value) => object_max_versions_limit_from_u64(value)?,
     };
 
@@ -42,20 +47,18 @@ fn init_object_max_versions_config() -> Result<()> {
 
 fn object_max_versions_limit_from_u64(value: u64) -> Result<usize> {
     if value == 0 {
-        return Err(object_max_versions_config_error());
+        return Err(Error::other(format!(
+            "{ENV_API_OBJECT_MAX_VERSIONS} must be a positive integer no greater than {}",
+            usize::MAX
+        )));
     }
 
-    let limit = usize::try_from(value).map_err(|_| object_max_versions_config_error())?;
-    if limit > DEFAULT_API_OBJECT_MAX_VERSIONS {
-        return Err(object_max_versions_config_error());
-    }
-    Ok(limit)
-}
-
-fn object_max_versions_config_error() -> Error {
-    Error::other(format!(
-        "{ENV_API_OBJECT_MAX_VERSIONS} must be a positive integer no greater than {DEFAULT_API_OBJECT_MAX_VERSIONS}"
-    ))
+    usize::try_from(value).map_err(|_| {
+        Error::other(format!(
+            "{ENV_API_OBJECT_MAX_VERSIONS} must be a positive integer no greater than {}",
+            usize::MAX
+        ))
+    })
 }
 
 #[cfg(test)]

scripts/check_test_wiring.py

@@ -1094,6 +1094,38 @@ def check_scanner_heal_evidence(root: Path, directory: Path, case_id: str) -> li
         return [f"scanner/heal evidence rejected: {error}"]
 
 
+def scanner_heal_release_status(root: Path, directory: Path) -> dict[str, object]:
+    """Return a compact release decision without weakening case validation."""
+    registry = read_json(root / ".config/scanner-heal-required-tests.json")
+    evidence_integer(registry.get("schema"), "registry schema", 1, 1)
+    cases = registry.get("cases")
+    require(isinstance(cases, dict) and cases, "invalid scanner/heal registry")
+    pending = registry.get("release_pending")
+    require(isinstance(pending, dict), "invalid scanner/heal release requirements")
+    for gate, reason in pending.items():
+        require(isinstance(gate, str) and re.fullmatch(r"[A-Z][A-Z0-9-]*", gate) is not None,
+                "invalid scanner/heal release gate")
+        require(isinstance(reason, str) and reason.strip(), f"missing release requirement for {gate}")
+
+    verified_cases = []
+    rejected_cases = []
+    for case_id in sorted(cases):
+        if check_scanner_heal_evidence(root, directory, case_id):
+            rejected_cases.append(case_id)
+        else:
+            verified_cases.append(case_id)
+
+    return {
+        "schema": 1,
+        "decision": "blocked",
+        "release_approved": False,
+        "release_schema_capable": False,
+        "verified_cases": verified_cases,
+        "rejected_cases": rejected_cases,
+        "pending_gates": sorted(pending),
+    }
+
+
 def validate(root: Path) -> list[str]:
     errors: list[str] = []
     errors.extend(check_core_fixtures(root))
@@ -1311,6 +1343,61 @@ class SelfTests(unittest.TestCase):
             self.assertTrue(any(error.startswith("pending R-D:") for error in errors))
             self.assertTrue(any(error.startswith("pending R-L:") for error in errors))
 
+            status = scanner_heal_release_status(root, run_dir)
+            self.assertEqual(status["decision"], "blocked")
+            self.assertFalse(status["release_approved"])
+            self.assertEqual(status["rejected_cases"], [])
+            self.assertEqual(len(status["pending_gates"]), 21)
+
+    def test_scanner_heal_case_only_schema_cannot_approve_release(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp:
+            root, run_dir = self.scanner_heal_fixture(Path(tmp))
+            registry = read_json(root / ".config/scanner-heal-required-tests.json")
+            registry["release_pending"] = {}
+            write_json(root / ".config/scanner-heal-required-tests.json", registry)
+
+            status = scanner_heal_release_status(root, run_dir)
+            self.assertEqual(status["decision"], "blocked")
+            self.assertFalse(status["release_approved"])
+            self.assertFalse(status["release_schema_capable"])
+            self.assertEqual(status["rejected_cases"], [])
+            self.assertEqual(status["pending_gates"], [])
+
+    def test_scanner_heal_release_status_rejects_synthetic_case(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp:
+            root, run_dir = self.scanner_heal_fixture(Path(tmp))
+            registry = read_json(root / ".config/scanner-heal-required-tests.json")
+            registry["release_pending"] = {}
+            write_json(root / ".config/scanner-heal-required-tests.json", registry)
+            path = run_dir / "background-target-crash.json"
+            oracle = read_json(path)
+            oracle["evidence"] = "synthetic"
+            write_json(path, oracle)
+            (run_dir / "execution.json").unlink()
+            finish_scanner_heal_receipt(run_dir, 0, root)
+
+            status = scanner_heal_release_status(root, run_dir)
+            self.assertEqual(status["decision"], "blocked")
+            self.assertFalse(status["release_approved"])
+            self.assertEqual(status["rejected_cases"], ["background-target-crash"])
+            self.assertEqual(status["pending_gates"], [])
+
+    def test_scanner_heal_release_status_rejects_focused_case_run(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp:
+            root, run_dir = self.scanner_heal_fixture(Path(tmp))
+            registry = read_json(root / ".config/scanner-heal-required-tests.json")
+            registry["release_pending"] = {}
+            write_json(root / ".config/scanner-heal-required-tests.json", registry)
+            (run_dir / "background-target-crash.json").unlink()
+            (run_dir / "execution.json").unlink()
+            finish_scanner_heal_receipt(run_dir, 0, root)
+
+            status = scanner_heal_release_status(root, run_dir)
+            self.assertEqual(status["decision"], "blocked")
+            self.assertFalse(status["release_approved"])
+            self.assertEqual(status["verified_cases"], ["background-target-restart"])
+            self.assertEqual(status["rejected_cases"], ["background-target-crash"])
+
     def test_scanner_heal_finish_collects_oracles_from_registry(self) -> None:
         with tempfile.TemporaryDirectory() as tmp:
             root, run_dir = self.scanner_heal_fixture(Path(tmp))
@@ -2158,7 +2245,8 @@ def main() -> int:
     if sys.argv[1:] == ["--self-test"]:
         suite = unittest.defaultTestLoader.loadTestsFromTestCase(SelfTests)
         return 0 if unittest.TextTestRunner(verbosity=2).run(suite).wasSuccessful() else 1
-    if sys.argv[1:2] in (["--begin-scanner-heal"], ["--finish-scanner-heal"], ["--check-scanner-heal"]):
+    if sys.argv[1:2] in (["--begin-scanner-heal"], ["--finish-scanner-heal"], ["--check-scanner-heal"],
+                         ["--check-scanner-heal-release"]):
         try:
             if len(sys.argv) == 5 and sys.argv[1] == "--begin-scanner-heal":
                 begin_scanner_heal_receipt(ROOT, Path(sys.argv[2]), Path(sys.argv[3]), Path(sys.argv[4]))
@@ -2173,7 +2261,16 @@ def main() -> int:
                 if not errors:
                     print(f"Case evidence verified: {sys.argv[3]}; this does not approve release")
                 return 1 if errors else 0
-            raise ValueError("expected --begin-scanner-heal DIR BINARY TEST_BINARY, --finish-scanner-heal DIR EXIT, or --check-scanner-heal DIR CASE|release")
+            if len(sys.argv) == 3 and sys.argv[1] == "--check-scanner-heal-release":
+                try:
+                    status = scanner_heal_release_status(ROOT, Path(sys.argv[2]))
+                except (OSError, KeyError, TypeError, ValueError, ET.ParseError) as error:
+                    print(json.dumps({"schema": 1, "decision": "invalid", "release_approved": False,
+                                      "error": str(error)}, sort_keys=True, separators=(",", ":")))
+                    return 2
+                print(json.dumps(status, sort_keys=True, separators=(",", ":")))
+                return 0 if status["release_approved"] else 1
+            raise ValueError("expected --begin-scanner-heal DIR BINARY TEST_BINARY, --finish-scanner-heal DIR EXIT, --check-scanner-heal DIR CASE|release, or --check-scanner-heal-release DIR")
         except (OSError, KeyError, TypeError, ValueError, subprocess.SubprocessError) as error:
             print(f"ERROR: {error}", file=sys.stderr)
             return 1

scripts/run_scanner_heal_evidence_case.sh

@@ -90,15 +90,22 @@ PY
 
 release_gate_must_remain_blocked() {
     local run_dir="$1"
-    local output="$run_dir/release-check.txt"
-    if "$PYTHON_BIN" "$ROOT/scripts/check_test_wiring.py" --check-scanner-heal "$run_dir" release >"$output" 2>&1; then
+    local output="$run_dir/release-status.json"
+    if "$PYTHON_BIN" "$ROOT/scripts/check_test_wiring.py" --check-scanner-heal-release "$run_dir" >"$output"; then
         echo "release gate unexpectedly approved a single Scanner/Heal evidence run" >&2
         return 1
     fi
-    if ! grep -Eq 'required test not selected:|pending [A-Z0-9-]+:' "$output"; then
-        echo "release gate did not explain why the Scanner/Heal release remains blocked" >&2
-        return 1
-    fi
+    "$PYTHON_BIN" - "$output" <<'PY'
+import json
+import pathlib
+import sys
+
+status = json.loads(pathlib.Path(sys.argv[1]).read_text())
+if status.get("decision") != "blocked" or status.get("release_approved") is not False:
+    raise SystemExit("release status did not record a blocked decision")
+if status.get("release_schema_capable") is not False:
+    raise SystemExit("case-only evidence schema unexpectedly became release-capable")
+PY
 }
 
 run_self_test() {
@@ -242,5 +249,5 @@ fi
 "$PYTHON_BIN" "$ROOT/scripts/check_test_wiring.py" --check-scanner-heal "$RUN_DIR" "$CASE_ID"
 release_gate_must_remain_blocked "$RUN_DIR"
 echo "Scanner/Heal evidence case verified: $CASE_ID"
-echo "Release gate remains blocked; details: $RUN_DIR/release-check.txt"
+echo "Release gate remains blocked; status: $RUN_DIR/release-status.json"
 echo "Evidence directory: $RUN_DIR"