[ https://issues.apache.org/jira/browse/SOLR-13032?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16707535#comment-16707535 ]
Hoss Man commented on SOLR-13032: --------------------------------- Jenkins has also seen testRedundantDeletes fail, which makes sense since they all follow the same logic pattern -- so if the pattern is flawed in one test method it will be flawed in of them. (In thetaphi/Lucene-Solr-master-MacOSX/4959, it was testRedundantDeletes that initially failed, and the nwhen trying to repro the failure testAdds failed instead. ---- i've started looking into this ... IIUC i think there is a simple race condition in how the tests try to {{waitForCommit}} that can result in an intermediate commit making the test think everything is "done" while there are still more docs in the tlog waiting to be committed, reading through the control flow to sanity check myself. > Harden MaxSizeAutoCommitTest.testAdds > ------------------------------------- > > Key: SOLR-13032 > URL: https://issues.apache.org/jira/browse/SOLR-13032 > Project: Solr > Issue Type: Sub-task > Security Level: Public(Default Security Level. Issues are Public) > Components: Tests > Reporter: Mark Miller > Priority: Major > -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org