lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dawid Weiss <dawid.we...@cs.put.poznan.pl>
Subject Re: [JENKINS] Lucene-Solr-4.x-Linux (32bit/jdk1.6.0_37) - Build # 3170 - Failure!
Date Wed, 12 Dec 2012 07:23:30 GMT
> [junit4:junit4] HEARTBEAT J0 PID(16801@serv1.sd-datasolutions.de): 2012-12-11T21:46:47,
stalled for 3189s at: TestStressReorder.testStressReorderVersions
> [junit4:junit4] JVM J0: stdout was not empty, see: /mnt/ssd/jenkins/workspace/Lucene-Solr-4.x-Linux/solr/build/solr-core/test/junit4-J0-20121211_203559_704.sysout
> [junit4:junit4] >>> JVM J0: stdout (verbatim) ----
> [junit4:junit4] java.lang.OutOfMemoryError: PermGen space
> [junit4:junit4] Dumping heap to /mnt/ssd/jenkins/workspace/Lucene-Solr-4.x-Linux/heapdumps/java_pid16801.hprof
...
>
> Interestingly it hung because it ran out of permgen. I have seen this in solr tests a
few times now. In all cases, later test fail on initializing, so they do nothing, only print
a permgen error and then the testing framework seems to get stuck. Dawid?

Uh, bummer.

I'm guessing that what's happening is the JVM runs out of permgen
space but the test never returns to the framework (the exception
probably happening in a forked thread, the test thread waiting for
some condition indefinitely) so the runner just sits there waiting
patiently for it to return or timeout.

I think we should lower the default timeout value and maybe just
override it for nightlies. Normal tests shouldn't really have a
timeout of 2 hours per suite or whatever it currently is.

Dawid

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message