prateekm opened a new pull request, #1654:
URL: https://github.com/apache/samza/pull/1654
Part 1 of 2. Follow up PR #2 to restore side input stores using Blob Store
backups coming soon.
Symptoms:
1. Side input stores are uploaded twice when using Blob Store State Backend.
2. Store-level Gauges (but not Timers) in BlobStoreBackupManagerMetrics are
broken for side input stores.
3. Task level Gauges in BlobStoreBackupManagerMetrics have incorrect value
(count twice for side input stores).
Cause:
1. StorageConfig#getStoreNames() returns side input stores twice in the
list.
2. BlobStoreBackupManager does not dedup storesToBackup list.
3. PR #1223 makes the duplicate-registration behavior between Gauges and
Timers inconsistent.
Changes:
1. Fixed StorageConfig#getStoreNames() to dedup store names.
2. Added defensive dedup in BlobStoreBackupManager.
3. Changed store level metrics initialization in
BlobStoreBackupManagerMetrics to computeIfAbsent instead of putIfAbsent to
avoid overwriting-yet-returning-old-Gauges in case of duplicate store names.
Tests:
Added unit tests for StorageConfig to verify deduping.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
