No heapdump, but the log will tell you what happened: https://builds.apache.org/job/Lucene-Solr-NightlyTests-6.x/174/consoleText
At the end you see this: [junit4] ERROR: JVM J0 ended with an exception: Quit event not received from the forked process? This may indicate JVM crash or runner bugs. And indeed, when you look at the logs, this is what it has to say: [junit4] <<< JVM J0: EOF ---- [junit4] JVM J0: stderr was not empty, see: /x1/jenkins/jenkins-slave/workspace/Lucene-Solr-NightlyTests-6.x/checkout/solr/build/solr-core/test/temp/junit4-J0-20161012_135246_796.syserr [junit4] >>> JVM J0 emitted unexpected output (verbatim) ---- [junit4] [junit4] Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "OldIndexDirectoryCleanupThreadForCore-awholynewstresscollection_collection1_2_shard3_replica2" [junit4] WARN: Unhandled exception in event serialization. -> java.lang.OutOfMemoryError: GC overhead limit exceeded These are the types or exceptions that are very difficult to guard against -- GC overhead limit simply happens when the JVM is low on memory and GC can't keep up with cleaning up. Here it happened on event serialization back to the master process controlling JUnit execution. The runner has to have enough room to serialize events, otherwise the results will be hard to predict or control. Dawid P.S. The entire log dump contains more stack dumps, but these seem to be dumped from Solr -- perhaps there is an obscured reason for running so low on memory there. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
