smengcl opened a new pull request, #11646: URL: https://github.com/apache/ozone/pull/11646
Generated-by: Claude Code (Opus 5.5) ## What changes were proposed in this pull request? `TestSCMHAManagerImpl#testRemoveSCM` intermittently times out waiting for the removed SCM to terminate, since HDDS-16922 added a 10 second wait for it. The removed follower only learns it was removed when its Ratis election timeout fires and the pre-vote returns `NOT_IN_CONF`. Ratis `FollowerState` sleeps a full random election timeout (5000 to 5200 ms with the default `ozone.scm.ha.ratis.leader.election.timeout` of 5 s) per iteration, and the last append from the leader usually arrives during the first sleep. So detection takes up to about 10.4 s, which overruns the wait. In the failed run the follower became candidate with `lastRpcElapsedTime:10137ms`. This PR lowers the SCM Ratis leader election timeout in the test to 1 s instead of increasing the wait. That bounds the removed follower's shutdown at about 2.4 s, well within the existing 10 s wait, and cuts the class run time from about 25 s to 9 s. ## What is the link to the Apache JIRA https://issues.apache.org/jira/browse/HDDS-17180 ## How was this patch tested? - `mvn -pl :hdds-server-scm test -Dtest=TestSCMHAManagerImpl` locally 5 times: all passed, `testRemoveSCM` took 2.2 to 2.3 s (10.2 to 10.4 s before the change). - 10x10 [flaky-test-check](https://github.com/smengcl/hadoop-ozone/actions/runs/39483766335) on this branch: passed 100/100. 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
