[HDFS-16948] Update log of BlockManager#chooseExcessRedundancyStriped when EC internal block is moved by balancer

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.3.4
• Fix Version/s: 3.4.0
• Component/s: erasure-coding
• Labels: pull-request-available
• Target Version/s: 3.4.0
• Hadoop Flags: Reviewed
• Assignee: Ruinan Gu
• Reporter: Ruinan Gu
• Created: 12/Mar/23 17:34
• Updated: 28/Jan/24 07:13
• Resolved: 23/Mar/23 09:01

Description

This is a follow-up of HDFS-16179.When the EC internal block is moved by the balancer, it will trigger chooseExcessRedundancyStriped( ), and the parameter delNodeHint is not null.So the nonExcess list will be modified in processChosenExcessRedundancy( ), normally the size of nonExcess will be reduced to the expected EC group size, that's why the annoying log "excess types chose ... is empty." will be printed.

Issue Links

• fixes: HDFS-16179 Update loglevel for BlockManager#chooseExcessRedundancyStriped to avoid too much logs
• links to: GitHub Pull Request #5474