[HBASE-26861] Fix flaky TestSnapshotFromMaster.testSnapshotHFileArchiving

• Type: Task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.5.0, 3.0.0-alpha-3, 2.4.12
• Component/s: snapshots, test
• Hadoop Flags: Reviewed
• Assignee: Duo Zhang
• Reporter: Duo Zhang
• Created: 18/Mar/22 07:19
• Updated: 19/Mar/22 20:12
• Resolved: 18/Mar/22 16:51

Description

After investigating, finally I found that the problem is we just use tryLock in SnapshotFileCache.getUnreferencedFiles. If we failed to grab lock, we will skip cleaning of the archive files.

We should add some retries in the test to make it more stable.

Issue Links

• links to: GitHub Pull Request #4242