[HBASE-24574] Procedure V2 - Distributed WAL Splitting => LOGGING

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 3.0.0-alpha-1, 2.3.0
• Component/s: wal
• Hadoop Flags: Reviewed
• Assignee: Michael Stack
• Reporter: Michael Stack
• Created: 16/Jun/20 18:37
• Updated: 02/Jul/20 14:07
• Resolved: 18/Jun/20 15:54

Description

The procedure-based distributed WAL splitter uses the old SplitLogManager class. It registers it as a chore even though the chore #run immediately skips out. Meantime we log SLM ran in #0ms frequently. For example...

...
2020-06-17 17:48:30,567 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
2020-06-17 17:48:31,567 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
2020-06-17 17:48:32,570 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
2020-06-17 17:48:33,567 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
2020-06-17 17:48:34,569 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
2020-06-17 17:48:35,569 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
2020-06-17 17:48:36,568 DEBUG [localhost:16000.splitLogManager..Chore.1] hbase.ScheduledChore: SplitLogManager Timeout Monitor execution time: 0 ms.
...

The new pv2 WAL splitter introduces new procedures that log their class name only w/o detail on log that is being split or the server that is running the split.

Issue Links

• is related to: HBASE-24585 Failed start recovering crash in standalone mode if procedure-based distributed WAL split & hbase.wal.split.to.hfile=true
• links to: GitHub Pull Request #1912