fix(ecstore): fence deletes against decommission commits (#6363)

* fix(ecstore): fence deletes against decommission commits

* fix(ecstore): preserve decommission target write locks

* fix(ecstore): preserve delete markers during source cleanup

* fix(ecstore): route batch delete markers to active pools

* fix(ecstore): preserve batch delete pool errors

* fix(ecstore): retain source-set lock during cleanup

* test(ecstore): exercise decommission delete fences

* test(ecstore): finish decommission delete fence scenario

* fix(ecstore): reuse fixed fence for reverse decommission

* fix(ecstore): fence decommission commit loss

* fix(ecstore): annotate batch delete fallback

* test(ecstore): fix decommission fence fixtures

* fix(ecstore): unblock decommission delete fences

* fix(ecstore): preserve distributed decommission set locks

* fix(ecstore): match decommission lock backend domain

* test(ecstore): align decommission fence barriers

* fix(ecstore): satisfy delete fence lint checks

* fix(rebalance): preserve access-denied delete errors
This commit is contained in:
Zhengchao An
2026-08-23 00:40:38 +08:00
committed by GitHub
parent 3ddf1a81ac
commit 7c1a76dfd9
13 changed files with 2606 additions and 134 deletions
+55 -24
View File
@@ -858,6 +858,7 @@ const EVENT_DISK_LOCAL_DIRECT_IO_FALLBACK: &str = "disk_local_direct_io_fallback
#[cfg(target_os = "linux")]
const EVENT_DISK_LOCAL_URING_LATCH_OFF: &str = "disk_local_uring_latch_off";
const EVENT_DISK_LOCAL_DELETE_FAILED: &str = "disk_local_delete_failed";
const EVENT_DISK_LOCAL_DELETE_ROLLBACK_FAILED: &str = "disk_local_delete_rollback_failed";
const EVENT_DISK_LOCAL_CHECK_PARTS: &str = "disk_local_check_parts";
const EVENT_DISK_LOCAL_ACCESS_FAILED: &str = "disk_local_access_failed";
const EVENT_DISK_LOCAL_VOLUME_SETUP_FAILED: &str = "disk_local_volume_setup_failed";
@@ -6106,6 +6107,43 @@ impl LocalDisk {
Ok((bytes, modtime))
}
async fn write_missing_delete_marker(
&self,
volume: &str,
path: &str,
fi: FileInfo,
object_dir: &Path,
xl_path: &Path,
rollback_dir: Option<Uuid>,
) -> Result<()> {
if let Some(rollback_dir) = rollback_dir {
let rollback_path = object_dir.join(rollback_dir.to_string());
fs::create_dir_all(&rollback_path).await.map_err(to_file_error)?;
fs::write(rollback_path.join(DELETE_MARKER_ROLLBACK_FILE), [])
.await
.map_err(to_file_error)?;
}
if let Err(err) = self.write_metadata("", volume, path, fi).await {
if let Some(rollback_dir) = rollback_dir
&& let Err(restore_err) = restore_delete_rollback(object_dir, xl_path, rollback_dir, &self.publication_root).await
{
warn!(
event = EVENT_DISK_LOCAL_DELETE_ROLLBACK_FAILED,
component = LOG_COMPONENT_ECSTORE,
subsystem = LOG_SUBSYSTEM_DISK_LOCAL,
result = "failed",
volume,
path,
rollback_dir = %rollback_dir,
error = ?restore_err,
"Disk local delete rollback failed"
);
}
return Err(err);
}
Ok(())
}
async fn delete_versions_internal(&self, volume: &str, path: &str, fis: &[FileInfo], opts: &DeleteOptions) -> Result<()> {
let volume_dir = self.io_get_bucket_path(volume)?;
let xlpath = self.io_get_object_path(volume, format!("{path}/{STORAGE_FORMAT_FILE}").as_str())?;
@@ -6123,7 +6161,20 @@ impl LocalDisk {
return restore_metadata_backup(object_dir, &xlpath, rollback_dir, &self.publication_root).await;
}
let (data, _) = self.read_all_data_with_dmtime(volume, volume_dir.as_path(), &xlpath).await?;
let (data, _) = match self.read_all_data_with_dmtime(volume, volume_dir.as_path(), &xlpath).await {
Ok(data) => data,
Err(DiskError::FileNotFound) => {
// `deleted` alone can be an explicit marker purge; only
// `mark_deleted` may create metadata that was not present.
let Some(delete_marker) = fis.iter().find(|fi| fi.deleted && fi.mark_deleted).cloned() else {
return Err(DiskError::FileNotFound);
};
return self
.write_missing_delete_marker(volume, path, delete_marker, object_dir, &xlpath, opts.old_data_dir)
.await;
}
Err(err) => return Err(err),
};
if data.is_empty() {
return Err(DiskError::FileNotFound);
@@ -10422,29 +10473,9 @@ impl DiskAPI for LocalDisk {
}
if fi.deleted && force_del_marker {
if let Some(rollback_dir) = rollback_dir {
let rollback_path = file_path.join(rollback_dir.to_string());
fs::create_dir_all(&rollback_path).await.map_err(to_file_error)?;
fs::write(rollback_path.join(DELETE_MARKER_ROLLBACK_FILE), [])
.await
.map_err(to_file_error)?;
}
if let Err(err) = self.write_metadata("", volume, path, fi).await {
if let Some(rollback_dir) = rollback_dir
&& let Err(restore_err) =
restore_delete_rollback(file_path.as_path(), &xl_path, rollback_dir, &self.publication_root).await
{
warn!(
volume,
path,
rollback_dir = %rollback_dir,
error = ?restore_err,
"failed to restore metadata after delete marker commit error"
);
}
return Err(err);
}
return Ok(());
return self
.write_missing_delete_marker(volume, path, fi, file_path.as_path(), &xl_path, rollback_dir)
.await;
}
return if fi.version_id.is_some() {