Hi,
>From https://builds.apache.org/job/PreCommit-HBASE-Build/3727/console (for
HBASE-7384) :

/home/jenkins/tools/maven/latest/bin/mvn clean test -DskipTests
-DHBasePatchProcess >
/home/jenkins/jenkins-slave/workspace/PreCommit-HBASE-Build/patchprocess/trunkJavacWarnings.txt
2>&1
Trunk compilation is broken?

I think this was due to the following check:

  if [[ $? != 0 ]] ; then
    echo "Trunk compilation is broken?"
    cleanupAndExit 1
  fi

I ran this command locally (with patch from HBASE-7384):

mvn clean test -DskipTests -DHBasePatchProcess > javacWarnings.txt

Here is the tail of javacWarnings.txt:

[INFO] Reactor Summary:
[INFO]
[INFO] HBase ............................................. SUCCESS [1.350s]
[INFO] HBase - Common .................................... SUCCESS [3.329s]
[INFO] HBase - Protocol .................................. SUCCESS [9.576s]
[INFO] HBase - Client .................................... SUCCESS [0.571s]
[INFO] HBase - Hadoop Compatibility ...................... SUCCESS [0.559s]
[INFO] HBase - Hadoop One Compatibility .................. SUCCESS [0.843s]
[INFO] HBase - Server .................................... SUCCESS [14.312s]
[INFO] HBase - Hadoop Two Compatibility .................. SUCCESS [2.103s]
[INFO] HBase - Integration Tests ......................... SUCCESS [1.080s]
[INFO] HBase - Examples .................................. SUCCESS [2.084s]
[INFO]
------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO]
------------------------------------------------------------------------
[INFO] Total time: 36.231s
[INFO] Finished at: Thu Dec 27 18:56:06 PST 2012

$? was 1.
I think the check should be improved so that Hadoop QA can continue running
test suite.

Cheers

Do you know why maven returned 1? It seems that the build was successful.

On Thu, Dec 27, 2012 at 7:04 PM, Ted Yu <[EMAIL PROTECTED]> wrote:

> should

What's your maven version?
I tried locally, with maven 3.0.4, $? is 0
I had a go on maven bugs, but I didn't find anything related.
On Fri, Dec 28, 2012 at 4:12 AM, Enis Söztutar <[EMAIL PROTECTED]> wrote:

> Do you know why maven returned 1? It seems that the build was successful.
>
> On Thu, Dec 27, 2012 at 7:04 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
>
> > should
>

I was using:
Apache Maven 3.0.3 (r1075438; 2011-02-28 09:31:09-0800)

I think the $? should be logged by test-patch.sh so that we can understand
what happened on hadoopx.

One other enhancement is to log the name of machine where the test suite
runs. This would be useful to identify faulty server where several tests
encountered OOME.

Cheers

On Thu, Dec 27, 2012 at 11:54 PM, Nicolas Liochon <[EMAIL PROTECTED]> wrote:

> What's your maven version?
> I tried locally, with maven 3.0.4, $? is 0
> I had a go on maven bugs, but I didn't find anything related.
>
>
> On Fri, Dec 28, 2012 at 4:12 AM, Enis Söztutar <[EMAIL PROTECTED]> wrote:
>
> > Do you know why maven returned 1? It seems that the build was successful.
> >
> > On Thu, Dec 27, 2012 at 7:04 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
> >
> > > should
> >
>

Ok, I tried on another machine with the 3.0.3 installed, I got the $? == 0
as well so it seems that it's something that is not reproducible on all
computers...

On Fri, Dec 28, 2012 at 3:38 PM, Ted Yu <[EMAIL PROTECTED]> wrote:

> I was using:
> Apache Maven 3.0.3 (r1075438; 2011-02-28 09:31:09-0800)
>
> I think the $? should be logged by test-patch.sh so that we can understand
> what happened on hadoopx.
>
> One other enhancement is to log the name of machine where the test suite
> runs. This would be useful to identify faulty server where several tests
> encountered OOME.
>
> Cheers
>
> On Thu, Dec 27, 2012 at 11:54 PM, Nicolas Liochon <[EMAIL PROTECTED]>
> wrote:
>
> > What's your maven version?
> > I tried locally, with maven 3.0.4, $? is 0
> > I had a go on maven bugs, but I didn't find anything related.
> >
> >
> > On Fri, Dec 28, 2012 at 4:12 AM, Enis Söztutar <[EMAIL PROTECTED]>
> wrote:
> >
> > > Do you know why maven returned 1? It seems that the build was
> successful.
> > >
> > > On Thu, Dec 27, 2012 at 7:04 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
> > >
> > > > should
> > >
> >
>

It turned out to be environment issue on hadoop9 and hadoop8.

I have narrowed machine selection for Hadoop QA

Thanks

On Friday, December 28, 2012, Nicolas Liochon wrote:

> Ok, I tried on another machine with the 3.0.3 installed, I got the $? == 0
> as well so it seems that it's something that is not reproducible on all
> computers...
>
>
>
>
>
> On Fri, Dec 28, 2012 at 3:38 PM, Ted Yu <[EMAIL PROTECTED]<javascript:;>>
> wrote:
>
> > I was using:
> > Apache Maven 3.0.3 (r1075438; 2011-02-28 09:31:09-0800)
> >
> > I think the $? should be logged by test-patch.sh so that we can
> understand
> > what happened on hadoopx.
> >
> > One other enhancement is to log the name of machine where the test suite
> > runs. This would be useful to identify faulty server where several tests
> > encountered OOME.
> >
> > Cheers
> >
> > On Thu, Dec 27, 2012 at 11:54 PM, Nicolas Liochon <[EMAIL PROTECTED]<javascript:;>
> >
> > wrote:
> >
> > > What's your maven version?
> > > I tried locally, with maven 3.0.4, $? is 0
> > > I had a go on maven bugs, but I didn't find anything related.
> > >
> > >
> > > On Fri, Dec 28, 2012 at 4:12 AM, Enis Söztutar <[EMAIL PROTECTED]<javascript:;>
> >
> > wrote:
> > >
> > > > Do you know why maven returned 1? It seems that the build was
> > successful.
> > > >
> > > > On Thu, Dec 27, 2012 at 7:04 PM, Ted Yu <[EMAIL PROTECTED]<javascript:;>>
> wrote:
> > > >
> > > > > should
> > > >
> > >
> >
>