[
https://issues.apache.org/jira/browse/SOLR-15032?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17926972#comment-17926972
]
David Smiley commented on SOLR-15032:
-------------------------------------
Looking at TestPullReplicaErrorHandling.testCloseHooksDeletedOnReconnect like
in [develocity
here|https://develocity.apache.org/scans/tests?search.relativeStartTime=P90D&search.rootProjectNames=solr-root&search.timeZoneId=America%2FNew_York&tests.container=org.apache.solr.cloud.TestPullReplicaErrorHandling&tests.test=testCloseHooksDeletedOnReconnect],
it's rather flaky and more flaky lately. Basically every failure is
{{expected:<5> but was:<3>}}
In the same failing run on my machine today, also saw
testCantConnectToPullReplica in this suite fail: {{Underlying core creation
failed while creating collection:
pull_replica_error_handling_test_close_hooks_deleted_on_reconnec}}
> Race condition in TestPullReplicaErrorHandling
> ----------------------------------------------
>
> Key: SOLR-15032
> URL: https://issues.apache.org/jira/browse/SOLR-15032
> Project: Solr
> Issue Type: Test
> Components: Tests
> Reporter: Mike Drob
> Priority: Major
> Labels: race-condition
>
> See discussion at
> https://github.com/apache/lucene-solr/pull/2115#discussion_r534445545
> There is a race condition in two tests in TestPullReplicaErrorHandling where
> we expire a ZK session, then wait for a node down and a node up. It's
> possible that the node recovers before we even start waiting for the first
> down.
> Better would be to set a watch on the live-node that we're about to expire,
> and wait to see the delete before checking for the node to come back up.
> cc: [~tflobbe]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
