You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by Apache Hudson Server <hu...@hudson.apache.org> on 2011/04/08 04:34:49 UTC

[HUDSON] Lucene-trunk - Build # 1523 - Failure

Build: https://hudson.apache.org/hudson/job/Lucene-trunk/1523/

1 tests failed.
REGRESSION:  org.apache.lucene.index.TestNRTThreads.testNRTThreads

Error Message:
Some threads threw uncaught exceptions!

Stack Trace:
junit.framework.AssertionFailedError: Some threads threw uncaught exceptions!
	at org.apache.lucene.util.LuceneTestCase$LuceneTestCaseRunner.runChild(LuceneTestCase.java:1232)
	at org.apache.lucene.util.LuceneTestCase$LuceneTestCaseRunner.runChild(LuceneTestCase.java:1160)
	at org.apache.lucene.util.LuceneTestCase.tearDown(LuceneTestCase.java:521)




Build Log (for compile errors):
[...truncated 11839 lines...]



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


RE: [HUDSON] Lucene-trunk - Build # 1523 - Failure

Posted by Steven A Rowe <sa...@syr.edu>.
AFAICT, these (Oracle/Sun-only) JVM parameters were introduced in 1.6 (that's what this parameters list says: http://blogs.sun.com/watt/resource/jvm-options-list.html) - we tell Jenkins to use 1.6 for Lucene/Solr testing, so this isn't an issue in practice, I guess.

Hopefully 1.5 JVMs, and non-Oracle/Sun JVMs, won't choke on these unknown parameters.

Steve

> -----Original Message-----
> From: Michael McCandless [mailto:lucene@mikemccandless.com]
> Sent: Friday, April 08, 2011 3:55 PM
> To: dev@lucene.apache.org
> Cc: Chris Hostetter; Apache Hudson Server
> Subject: Re: [HUDSON] Lucene-trunk - Build # 1523 - Failure
> 
> On Fri, Apr 8, 2011 at 2:36 PM, Chris Hostetter
> <ho...@fucit.org> wrote:
> >
> > : OOME on this one... we really need the "dump heap on OOME" JRE command
> > : line set...
> >
> > Ant runs the tests in forked JVMs right? so thta should just be a
> > build.xml change.
> 
> OK I tried that, with the patch below (applies to 3.x), and then
> provoked an OOME and it works great, though I think this is Sun
> (Oracle!) JRE specific... which is OK for now (we use Oracle JRE on
> Jenkins right?), but if we want to rotate JREs in the future this
> won't work...
> 
> The problem is... the resulting dump is large (mine was ~400 MB).  We
> can specify a location for the dump (-XX:HeapDumpPath=/some/path)... I
> think we should somehow remove them after a few days?  How much disk
> space can we use up?
> 
> Patch:
> 
> Index: solr/build.xml
> ===================================================================
> --- solr/build.xml	(revision 1089906)
> +++ solr/build.xml	(working copy)
> @@ -464,6 +464,7 @@
>           <jvmarg line="${dir.prop}"/>
>        -->
>        <jvmarg line="${args}"/>
> +      <jvmarg line="-XX:+HeapDumpOnOutOfMemoryError"/>
> 
>        <formatter classname="${junit.details.formatter}"
> usefile="false" if="junit.details"/>
>        <classpath refid="test.run.classpath"/>
> Index: lucene/common-build.xml
> ===================================================================
> --- lucene/common-build.xml	(revision 1089906)
> +++ lucene/common-build.xml	(working copy)
> @@ -488,6 +488,7 @@
>  	      </assertions>
> 
>  	      <jvmarg line="${args}"/>
> +	      <jvmarg value="-XX:+HeapDumpOnOutOfMemoryError"/>
> 
>  	      <!-- allow tests to control debug prints -->
>  	      <sysproperty key="tests.verbose" value="${tests.verbose}"/>
> 
> Mike
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org


Re: [HUDSON] Lucene-trunk - Build # 1523 - Failure

Posted by Michael McCandless <lu...@mikemccandless.com>.
On Fri, Apr 8, 2011 at 2:36 PM, Chris Hostetter
<ho...@fucit.org> wrote:
>
> : OOME on this one... we really need the "dump heap on OOME" JRE command
> : line set...
>
> Ant runs the tests in forked JVMs right? so thta should just be a
> build.xml change.

OK I tried that, with the patch below (applies to 3.x), and then
provoked an OOME and it works great, though I think this is Sun
(Oracle!) JRE specific... which is OK for now (we use Oracle JRE on
Jenkins right?), but if we want to rotate JREs in the future this
won't work...

The problem is... the resulting dump is large (mine was ~400 MB).  We
can specify a location for the dump (-XX:HeapDumpPath=/some/path)... I
think we should somehow remove them after a few days?  How much disk
space can we use up?

Patch:

Index: solr/build.xml
===================================================================
--- solr/build.xml	(revision 1089906)
+++ solr/build.xml	(working copy)
@@ -464,6 +464,7 @@
          <jvmarg line="${dir.prop}"/>
       -->
       <jvmarg line="${args}"/>
+      <jvmarg line="-XX:+HeapDumpOnOutOfMemoryError"/>

       <formatter classname="${junit.details.formatter}"
usefile="false" if="junit.details"/>
       <classpath refid="test.run.classpath"/>
Index: lucene/common-build.xml
===================================================================
--- lucene/common-build.xml	(revision 1089906)
+++ lucene/common-build.xml	(working copy)
@@ -488,6 +488,7 @@
 	      </assertions>
 	
 	      <jvmarg line="${args}"/>
+	      <jvmarg value="-XX:+HeapDumpOnOutOfMemoryError"/>
 	
 	      <!-- allow tests to control debug prints -->
 	      <sysproperty key="tests.verbose" value="${tests.verbose}"/>

Mike

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


Re: [HUDSON] Lucene-trunk - Build # 1523 - Failure

Posted by Chris Hostetter <ho...@fucit.org>.
: OOME on this one... we really need the "dump heap on OOME" JRE command
: line set...

Ant runs the tests in forked JVMs right? so thta should just be a 
build.xml change.


-Hoss

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


Re: [HUDSON] Lucene-trunk - Build # 1523 - Failure

Posted by Michael McCandless <lu...@mikemccandless.com>.
OOME on this one... we really need the "dump heap on OOME" JRE command
line set...

Mike

http://blog.mikemccandless.com

On Thu, Apr 7, 2011 at 10:34 PM, Apache Hudson Server
<hu...@hudson.apache.org> wrote:
> Build: https://hudson.apache.org/hudson/job/Lucene-trunk/1523/
>
> 1 tests failed.
> REGRESSION:  org.apache.lucene.index.TestNRTThreads.testNRTThreads
>
> Error Message:
> Some threads threw uncaught exceptions!
>
> Stack Trace:
> junit.framework.AssertionFailedError: Some threads threw uncaught exceptions!
>        at org.apache.lucene.util.LuceneTestCase$LuceneTestCaseRunner.runChild(LuceneTestCase.java:1232)
>        at org.apache.lucene.util.LuceneTestCase$LuceneTestCaseRunner.runChild(LuceneTestCase.java:1160)
>        at org.apache.lucene.util.LuceneTestCase.tearDown(LuceneTestCase.java:521)
>
>
>
>
> Build Log (for compile errors):
> [...truncated 11839 lines...]
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>
>

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