[HBASE-26947] Implement a special TestAppender to limit the size of test output

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.5.0, 3.0.0-alpha-3
• Component/s: logging, test
• Hadoop Flags: Reviewed
• Assignee: Duo Zhang
• Reporter: Duo Zhang
• Created: 12/Apr/22 11:04
• Updated: 22/Apr/22 19:32
• Resolved: 21/Apr/22 15:30

Description

Recently our build nodes were out of space several times. I've checked some of them, all are because of logging in an infinite loop which makes a very big log output.

I've already opened several issues for this, such as HBASE-26858, HBASE-26941, HBASE-26946.

So in general, I think we should have a way to limit the total size of a log output. So when we do something wrong in code it will not eat all the space soon and bring down our ci.

Issue Links

• is related to: HBASE-26941 LocalHBaseCluster.waitOnRegionServer should not call join while interrupted
• is related to: HBASE-26946 TestRetainAssignmentOnRestartSplitWithoutZk infinite loop
• is related to: HBASE-26858 Refactor TestMasterRegionOnTwoFileSystems to avoid dead loop
• is related to: HBASE-26936 project Jenkins nodes need a fail-safe clean up of surefire deferred output
• links to: GitHub Pull Request #4340