Skip to content

Cleanup BlobStoreRepository Abort and Failure Handling #46208

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 8 commits into from
Sep 2, 2019
Merged

Cleanup BlobStoreRepository Abort and Failure Handling #46208

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
6e4bd15
bck
original-brownbear Aug 31, 2019
f44fa90
step 1
original-brownbear Aug 31, 2019
9e5465f
Merge remote-tracking branch 'elastic/master' into rationalize-blobst…
original-brownbear Aug 31, 2019
74b152d
bck
original-brownbear Aug 31, 2019
d1845f6
Merge remote-tracking branch 'elastic/master' into rationalize-blobst…
original-brownbear Sep 1, 2019
bf9e608
dont mess with store refcnt
original-brownbear Sep 1, 2019
fbf90ab
Merge remote-tracking branch 'elastic/master' into rationalize-blobst…
original-brownbear Sep 2, 2019
c09a8f6
CR: add comment
original-brownbear Sep 2, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 12 additions & 1 deletion server/src/main/java/org/elasticsearch/repositories/blobstore/BlobStoreRepository.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -110,6 +110,7 @@
import java.util.Optional;
import java.util.Set;
import java.util.concurrent.Executor;
import java.util.concurrent.atomic.AtomicBoolean;
import java.util.stream.Collectors;

import static org.elasticsearch.index.snapshots.blobstore.BlobStoreIndexShardSnapshot.FileInfo.canonicalName;
Expand Down Expand Up @@ -1048,17 +1049,27 @@ public void snapshotShard(Store store, MapperService mapperService, SnapshotId s
final GroupedActionListener<Void> filesListener =
new GroupedActionListener<>(allFilesUploadedListener, indexIncrementalFileCount);
final Executor executor = threadPool.executor(ThreadPool.Names.SNAPSHOT);
// Flag to signal that the snapshot has been aborted/failed so we can stop any further blob uploads from starting
final AtomicBoolean alreadyFailed = new AtomicBoolean();
for (BlobStoreIndexShardSnapshot.FileInfo snapshotFileInfo : filesToSnapshot) {
executor.execute(new ActionRunnable<>(filesListener) {
@Override
protected void doRun() {
try {
snapshotFile(snapshotFileInfo, indexId, shardId, snapshotId, snapshotStatus, store);
if (alreadyFailed.get() == false) {
snapshotFile(snapshotFileInfo, indexId, shardId, snapshotId, snapshotStatus, store);
}
Copy link
Member

@tlrx tlrx Sep 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we call filesListener.onFailure() with a IndexShardSnapshotFailedException and an appropriate comment in case the upload is skipped because of the alreadyFailed flag?

Copy link
Contributor Author

@original-brownbear original-brownbear Sep 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we do that, we get all those execeptions added to the overall exception again, precisely what I tried avoiding here. Note that you could theoretically have a pretty large number of files here.

Copy link
Member

@tlrx tlrx Sep 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right - then at least a trace log would be helpful?

Copy link
Contributor Author

@original-brownbear original-brownbear Sep 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To be honest, I can't see what it would be useful for?
We'll get some indication that things were aborted anyway from the files that were actually aborted to set the flag and we will not update the blob index-N in the blob as a result of failing here as well -> seems like the information on what precise files were never uploaded is completely irrelevant?

Copy link
Member

@tlrx tlrx Sep 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK

(Monday morning reviews, you know)

filesListener.onResponse(null);
} catch (IOException e) {
throw new IndexShardSnapshotFailedException(shardId, "Failed to perform snapshot (index files)", e);
}
}

@Override
public void onFailure(Exception e) {
alreadyFailed.set(true);
super.onFailure(e);
}
});
}
} catch (Exception e) {
Expand Down