[HBASE-28129] Do not retry refreshSources when region server is already stopping

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.6.0, 2.4.18, 2.5.6, 3.0.0-beta-1
• Component/s: Replication, rpc
• Hadoop Flags: Reviewed
• Assignee: Duo Zhang
• Reporter: Duo Zhang
• Created: 03/Oct/23 12:17
• Updated: 08/Oct/23 05:47
• Resolved: 07/Oct/23 08:10

Description

Found this when debugging HBASE-28114, it is designed to retry forever but when aborting a region server, we will stop the rpc client, and it will retry a lot of times and output a lot of logs, which even cause the Jenkins job OOM…

I think we should throw the exception out directly and give up retry when aborting the region server.

==== Updated ====

It turns out that the problem is the retry when initialize new shipper failed. We should stop retrying when region server is already stopped.

Issue Links

• links to: GitHub Pull Request #5453