[BugFix] Fix several problem for cluster snapshot backup #55177
+214
−158
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![starrocks-cr[bot]](https://avatars.githubusercontent.com/in/404334?s=60&v=4){kind=small}
|
|
||
| automatedSnapshotSvName = data.getAutomatedSnapshotSvName(); | ||
| automatedSnapshot = data.getAutomatedSnapshot(); | ||
| historyAutomatedSnapshotJobs = data.getHistoryAutomatedSnapshotJobs(); | ||
| } | ||
|
|
||
| @Override |
There was a problem hiding this comment.
The most risky bug in this code is:
The use of ClusterSnapshotUtils.clearAllAutomatedSnapshotFromRemote(null) which may result in unintended deletions if a null argument is used incorrectly, indicating an unsafe invocation.
You can modify the code like this:
public void setAutomatedSnapshotOff(AdminSetAutomatedSnapshotOffStmt stmt) {
setAutomatedSnapshotOff();
ClusterSnapshotLog log = new ClusterSnapshotLog();
log.setDropSnapshot(AUTOMATED_NAME_PREFIX);
GlobalStateMgr.getCurrentState().getEditLog().logClusterSnapshotLog(log);
try {
// Ensure the method correctly targets intended snapshots without using null.
String snapshotName = (automatedSnapshot != null) ? automatedSnapshot.getSnapshotName() : null;
if (snapshotName != null) {
ClusterSnapshotUtils.clearAllAutomatedSnapshotFromRemote(snapshotName);
}
} catch (StarRocksException e) {
LOG.warn("Cluster Snapshot delete failed, err msg: {}", e.getMessage());
}
}This ensures that only defined snapshot names are passed to avoid accidental deletions or errors.
| if (path.getName().startsWith(ClusterSnapshotMgr.AUTOMATED_NAME_PREFIX)) { | ||
| HdfsUtil.deletePath(status.path, brokerDesc); | ||
| } | ||
| } | ||
| } | ||
| } |
There was a problem hiding this comment.
The most risky bug in this code is:
Potentially unsafe deletion of files without confirmation.
You can modify the code like this:
public static void clearAllAutomatedSnapshotFromRemote(String notAllowDelete) throws StarRocksException {
StorageVolume sv = GlobalStateMgr.getCurrentState().getClusterSnapshotMgr().getAutomatedSnapshotSv();
BrokerDesc brokerDesc = new BrokerDesc(sv.getProperties());
String snapshotImageRootPath = getSnapshotImagePath(sv, "");
List<TBrokerFileStatus> statuses =
HdfsUtil.listPath(snapshotImageRootPath + ClusterSnapshotMgr.AUTOMATED_NAME_PREFIX + "*",
false, sv.getProperties());
for (TBrokerFileStatus status : statuses) {
Path path = new Path(status.path);
if (notAllowDelete != null && path.getName().equals(notAllowDelete)) {
continue;
}
if (path.getName().startsWith(ClusterSnapshotMgr.AUTOMATED_NAME_PREFIX)) {
LOG.warn("Attempting to delete snapshot file: " + status.path); // Add logging before deletion
// Consider adding a confirmation mechanism or revisiting the conditions here to ensure safe deletion.
HdfsUtil.deletePath(status.path, brokerDesc);
}
}
}Explanation: The method clearAllAutomatedSnapshotFromRemote potentially deletes snapshots based on a prefix pattern without explicit confirmation or secondary checks. Adding logging provides a trace for deleted paths, and considering a confirmation mechanism can guard against accidental data loss.
| if (!createStarMgrImageRet.first) { | ||
| errMsg = "checkpoint failed for starMgr image: " + createStarMgrImageRet.second; | ||
| break; | ||
| } | ||
| } | ||
| LOG.info("Finished create image for starMgr image, version: {}", consistentIds.second); | ||
|
|
There was a problem hiding this comment.
The most risky bug in this code is:
There is a possibility of bypassing the creation of FE or starMgr images if their journal IDs are equal to or greater than the checkpoint journal IDs. This might be unintended behavior and could lead to missing checkpoints.
You can modify the code like this:
protected void runCheckpointScheduler() {
job.logJob();
Pair<Long, Long> getFEIdsRet = feController.getCheckpointJournalIds();
long feImageJournalId = getFEIdsRet.first;
long feCheckpointJournalId = consistentIds.first;
// Ensure checkpoint is created even if IDs are equal
Pair<Boolean, String> createFEImageRet = feController.runCheckpointControllerWithIds(feImageJournalId, feCheckpointJournalId);
if (!createFEImageRet.first) {
errMsg = "checkpoint failed for FE image: " + createFEImageRet.second;
break;
}
LOG.info("Finished create image for FE image, version: {}", consistentIds.first);
Pair<Long, Long> getStarMgrIdsRet = starMgrController.getCheckpointJournalIds();
long starMgrImageJournalId = getStarMgrIdsRet.first;
long starMgrCheckpointJournalId = consistentIds.second;
// Ensure checkpoint is created even if IDs are equal
Pair<Boolean, String> createStarMgrImageRet = starMgrController.runCheckpointControllerWithIds(starMgrImageJournalId, starMgrCheckpointJournalId);
if (!createStarMgrImageRet.first) {
errMsg = "checkpoint failed for starMgr image: " + createStarMgrImageRet.second;
break;
}
LOG.info("Finished create image for starMgr image, version: {}", consistentIds.second);
}
Signed-off-by: srlch <[email protected]>
srlch
force-pushed
the
fix_cluster_snapshot_backup
branch
from
429e5b9 to
50c592c
Compare
Signed-off-by: srlch <[email protected]>
Signed-off-by: srlch <[email protected]>
Signed-off-by: srlch <[email protected]>
|
[Java-Extensions Incremental Coverage Report]✅ pass : 0 / 0 (0%) |
[FE Incremental Coverage Report]❌ fail : 81 / 121 (66.94%) file detail
|
[BE Incremental Coverage Report]❌ fail : 0 / 3 (00.00%) file detail
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What I'm doing:
Fix several problem for cluster snapshot backup:
cluster_snapshotssys tableif the checkpointJournanlid is greater than image version
isUnFinishedStatefunctionsnapshot is created.
resetLastUnFinishedAutomatedSnapshotJobto reset the unfinished jobif fe restart or leader change.
Fixes #issue
What type of PR is this:
Does this PR entail a change in behavior?
If yes, please specify the type of change:
Checklist:
Bugfix cherry-pick branch check: