[ https://issues.apache.org/jira/browse/CASSANDRA-16679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17600968#comment-17600968 ]
Ekaterina Dimitrova edited comment on CASSANDRA-16679 at 9/6/22 8:25 PM: ------------------------------------------------------------------------- {quote}The progression of things looks pretty normal, and if there *was* some issue with shutdown/startup for node 4, I would expect 1800+ runs across the three branches to expose it. In the end, I think the original failure report was so long ago that we should fix what appears to be the (reproducible on trunk, etc.) failure and open a new Jira if somehow this test ever gives us more problems {quote} Agreed, thanks for checking. You and David fixed a lot of things in the start/stop order for in-jvm tests framework so I guess this got fixed in time. {quote}The original Python dtest does not make any assertions around the hint-related metrics, so I wouldn't expect it to have the issue my patch addresses. {quote} (y) So now that we have already support for vnodes if we want to match the python dtests, _test_hintedhandoff_decom_ has a mark {_}no_vnode{_}, we need to do the same for the in-jvm version of the test (the original test was added before vnodes support for in-jvm framework was added) The rest LGTM was (Author: e.dimitrova): {quote}The progression of things looks pretty normal, and if there *was* some issue with shutdown/startup for node 4, I would expect 1800+ runs across the three branches to expose it. In the end, I think the original failure report was so long ago that we should fix what appears to be the (reproducible on trunk, etc.) failure and open a new Jira if somehow this test ever gives us more problems {quote} Agreed, thanks for checking. You and David fixed a lot of things in the start/stop order for in-jvm tests framework so I guess this got fixed in time. {quote}The original Python dtest does not make any assertions around the hint-related metrics, so I wouldn't expect it to have the issue my patch addresses. {quote} (y) So now that we have already support for vnodes if we want to match the python dtests, test_hintedhandoff_decom has a mark no_vnode, we need to do the same for the in-jvm version (the original test was added before vnodes support for in-jvm framework was added) The rest LGTM > HintedHandoffAddRemoveNodesTest is failing > ------------------------------------------ > > Key: CASSANDRA-16679 > URL: https://issues.apache.org/jira/browse/CASSANDRA-16679 > Project: Cassandra > Issue Type: Bug > Components: CI > Reporter: Ekaterina Dimitrova > Assignee: Caleb Rackliffe > Priority: Normal > Fix For: 4.0.x, 4.1-beta, 4.x > > Time Spent: 50m > Remaining Estimate: 0h > > https://ci-cassandra.apache.org/job/Cassandra-devbranch/785/testReport/junit/org.apache.cassandra.distributed.test/HintedHandoffAddRemoveNodesTest/shouldStreamHintsDuringDecommission/ > Java 8 > [https://app.circleci.com/pipelines/github/adelapena/cassandra/464/workflows/888e47fb-e432-47f2-97df-c34a0d33753a/jobs/4104] > > and Java 11 > > [https://app.circleci.com/pipelines/github/adelapena/cassandra/464/workflows/ce0f0690-e488-4b5b-ab82-ce92a2f336d8/jobs/4105] -- This message was sent by Atlassian Jira (v8.20.10#820010) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org