fix(storage): harden ODM and scanner publication (#7187)

* fix(storage): harden ODM and scanner publication

* fix(app): simplify absent SSE configuration matching

* test(heal): settle PUT rename tails before disk-wipe fixtures

* fix(ecstore): remove duplicate local rename implementation

Keep the canonical commit module after concurrent storage changes merged.
The control-write and rollback changes are already present there.

Co-Authored-By: heihutu <heihutu@gmail.com>
Co-Authored-By: zhi22915 <qiuzgang@gmail.com>

* fix(ci): satisfy new clippy lints

* style(scanner): order merged test imports

* fix(scanner): invalidate bucket work after namespace completion

* fix(scanner): fence cached snapshots by scan execution

---------

Co-authored-by: houseme <housemecn@gmail.com>
Co-authored-by: heihutu <heihutu@gmail.com>
Co-authored-by: zhi22915 <qiuzgang@gmail.com>
This commit is contained in:
Zhengchao An
2026-09-05 21:47:12 +08:00
committed by GitHub
parent 447f3c704b
commit e2a921bc16
35 changed files with 2252 additions and 274 deletions
+31 -5
View File
@@ -1616,7 +1616,7 @@ where
// Refresh the storage-owned movement snapshot before reading background
// heal state. A missing heal object yields an in-memory default; do not
// let that default influence a cycle while publication is blocked.
if storeapi.scanner_data_usage_publication_blocked().await {
if storeapi.scanner_data_movement_pause_status().await.paused {
mark_scan_cycle_idle(cycle_info, &mut cycle_metrics_guard).await;
return ScannerCycleOutcome::Deferred(ScannerCycleDeferReason::DataMovement);
}
@@ -1816,6 +1816,19 @@ where
let publication_defer_reason = publication_defer_reason
.or(remote_lease_defer_reason)
.or(remote_lease_fence_defer_reason);
// A PUT tail can finish between the walk and lease acquisition without
// changing the movement epoch accepted by those leases. Re-prove the
// namespace baseline only after every peer has granted publication.
let post_lease_activity_defer_reason = if publication_defer_reason.is_none()
&& remote_publication_leases.is_some()
&& let Ok(result) = &scan_result
&& result.status == ScannerCycleStatus::Complete
{
scanner_post_lease_activity_defer_reason(result.activity_digest(), probe_scanner_activity(storeapi.as_ref(), true).await)
} else {
None
};
let publication_defer_reason = publication_defer_reason.or(post_lease_activity_defer_reason);
// Include reasons discovered while acquiring or validating remote leases.
let publication_deferred = publication_defer_reason.is_some();
let budget_elapsed = cycle_budget.budget_elapsed() && !ctx.is_cancelled();
@@ -3240,6 +3253,21 @@ where
}
}
fn scanner_post_lease_activity_defer_reason(
expected_digest: Option<[u8; 32]>,
activity: Result<ScannerActivitySnapshot, String>,
) -> Option<ScannerCycleDeferReason> {
match activity {
Ok(snapshot)
if scanner_activity_allows_usage_publication(&snapshot)
&& expected_digest == Some(scanner_activity_snapshot_digest(&snapshot)) =>
{
None
}
Ok(_) | Err(_) => Some(ScannerCycleDeferReason::ActivityBaselineUnavailable),
}
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
enum ScannerCyclePreCommitOutcome {
RecoverCacheCycle(u64),
@@ -3428,13 +3456,11 @@ use cycle_state::*;
use leadership::*;
use usage_store::*;
#[cfg(test)]
pub(crate) use activity::scanner_activity_snapshot_digest;
pub use activity::scanner_topology_digest;
pub(crate) use activity::{
ScannerActivitySnapshot, ScannerDirtyUsageAcknowledgement, probe_scanner_activity, scanner_activity_allows_usage_publication,
scanner_activity_dirty_usage_state_for_host, scanner_activity_publication_lease_targets, scanner_activity_structural_digest,
scanner_dirty_usage_acknowledgements,
scanner_activity_dirty_usage_state_for_host, scanner_activity_publication_lease_targets, scanner_activity_snapshot_digest,
scanner_activity_structural_digest, scanner_dirty_usage_acknowledgements,
};
pub(crate) use activity::{ScannerCycleOutcome, scanner_cycle_outcome_with_pending_maintenance};
pub use backlog::{