[ https://issues.apache.org/jira/browse/SOLR-9544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15509878#comment-15509878 ]
Alan Woodward commented on SOLR-9544: ------------------------------------- I guess the simplest way to fix this is to add some retry-and-timeout logic to ObjectReleaseTracker.clearObjectTrackerAndCheckEmpty()? > ObjectReleaseTracker can false-fail on late asynchronous closing resources > -------------------------------------------------------------------------- > > Key: SOLR-9544 > URL: https://issues.apache.org/jira/browse/SOLR-9544 > Project: Solr > Issue Type: Bug > Security Level: Public(Default Security Level. Issues are Public) > Reporter: Alan Woodward > > SolrTestCaseJ4 assumes that, once its embedded CoreContainer has shutdown, it > can check the ObjectReleaseTracker and ensure that all cores are closed. > However, if the test has kicked off some asynchronous core reloads then this > assumption doesn't necessarily hold, particularly on slow machines. > See http://jenkins.thetaphi.de/job/Lucene-Solr-6.x-Windows/466/ for an > example failure. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org