On 29/12/2011 18:54, Thorsten Scherler wrote:
On Thu, 2011-12-29 at 18:07 +0100, Francesco Chicchiriccò wrote:
On 29/12/2011 15:50, Thorsten Scherler wrote:
On Thu, 2011-12-29 at 14:01 +0000, Apache Jenkins Server wrote:
The Apache Jenkins build system has built Cocoon-trunk (build #122)

Status: Failure

Check console output at https://builds.apache.org/job/Cocoon-trunk/122/ to view the results.
Hmm, weird. Not sure what is wrong with Jenkins, but in my system I
actually had to set MAVEN_OPTS="-Xmx256m -XX:PermSize=128m
-XX:MaxPermSize=128m" to start it.sh and build successful.

Maybe same prob for J but not sure since the NPE could be about
anything. Let us see the next build.
Uhm, build 123 went wrong as well, with same meaningless NPE :-(

Any idea?
The only one regarding the MAVEN_OPTS. 

...but https://builds.apache.org/job/Cocoon-trunk/123/consoleFull 
[trunk] $ /home/hudson/tools/java/latest1.6/bin/java -Xmx768m -Xms768m
-client -XX:MaxPermSize=256m

It actually gives the build more Perm then I do. :(

It builds fine on my box and the one of my colleague. Further we have setup a jenkins for our current project that as well builds c3 as part of the main project and that is working fine as well (for all commits since 24.11 or r1203644).

The only thing that I see is that is doing and svn up on the same files between 122 and 123 (in the latter there a couple more but is 122 +).
Further it does not make sense that 122 is failing in the parent, since it does not have even changes in parent/pom.xml (At revision 1225536)

Further in 123 it goes to r1225592 and getting the same files at before (even more), where it should heve done only:
thorsten@mcKenny:~/src/apache/c3$ svnu
Updated to revision 1225602.

Shot in the dark some underlying os issue, being it disk full or something similar.

If this can be any useful, this morning another (non ASF) Jenkins instance started failing builds with NPE and apparently the same stacktrace [1]; in this case nothing has changed in sources since yesterday evening when we had latest successful build. Argh.

After restarting a couple of time, I've been able to notice that - in my case - the actual failing was during PMD plugin execution, for a missing dependency. Now it's back working.

Maybe we should ping infra?



[1] http://jenkins.syncope-idm.org/job/Syncope%20branch%200_7_X%20(GNU%20Linux)/18/org.syncope$syncope/console
[2] https://issues.jenkins-ci.org/browse/JENKINS-12056
Francesco Chicchiriccò

Apache Cocoon Committer and PMC Member