www-builds mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kristian Waagan <krist...@apache.org>
Subject Re: Use of Lucene Zones Re: Hudson build failed on lucene.zones.apache.org with OOME
Date Wed, 24 Feb 2010 11:34:00 GMT
On 24.02.10 12:03, Justin Mason wrote:
> hi Grant --
>
> looking at the Derby configuration pages, they all seem to be
> tied to a node or pair of nodes:
>
> http://hudson.zones.apache.org/hudson/view/Derby/job/Derby-branch-10.5/configure
> Ubuntu (minerva, vesta)
>
> http://hudson.zones.apache.org/hudson/view/Derby/job/Derby-docs/configure
> Ubuntu (minerva, vesta)
>
> http://hudson.zones.apache.org/hudson/view/Derby/job/Derby-trunk/configure
> Ubuntu (minerva, vesta)
>
> http://hudson.zones.apache.org/hudson/view/Derby/job/Derby-trunk_suites.All/configure
> minerva only
>
> however I can see that Derby-trunk build in
> http://hudson.zones.apache.org/hudson/computer/lucene.zones.apache.org%20%28Solaris%2010%29/builds
> 15 hours ago.
>
> Did someone change the Derby-trunk job since then?

Hi Justin,

Yes I did, after Grant brought up this issue. I'm the one who created 
the jobs and put them on the Lucene zone in the first place.

Derby-trunk, Derby-branch-10.5, and Derby-docs used to run on the Lucene 
zone. The first two polling svn hourly (run time < 10 min), the last one 
polling svn twice a week (run time ~45 min).

The two last jobs (Derby-trunk_suites.All and Derby-trunk_clover) are 
more experimental jobs. The former runs the Derby test suite (takes 3 
hours+, grabs lock), whereas the latter was supposed to publish Clover 
reports. However, testing on a non-ASF machine showed that generating 
the HTML report took more than 26 hours, so this job won't be running in 
Hudson... I'm also afraid that the ASF servers won't be too happy 
serving 10 MB+ HTML files (report dir ends up at 2.6 GB).
I'm still investigating if I can configure Clover to do less work and 
generate a smaller report a lot faster (i.e. disable per test coverage 
and cutting down the number of instrumented classes, this task is on the 
back-burner for now).


-- 
Kristian

>
>
>
> On Tue, Feb 23, 2010 at 17:50, Grant Ingersoll<gsingers@apache.org>  wrote:
>> I'm a bit confused why Derby is using the Lucene Zone to begin with.  Doesn't Derby
have it's own Zone?  I don't necessarily have a problem with it, but it would be nice if the
Lucene PMC was asked before our resources are utilized.  I'm sure it was just assumed b/c
the Lucene Zone shows up in Hudson, but my understanding from earlier discussions was that
it was done out of convenience for Lucene (b/c that's where Hudson was first installed anyway)
and that other Zones would be setup if projects wanted to provide their own resources to Hudson
and that other build machines were provisioned to support the majority of projects
>>
>> Anyone else have any insight?
>>
>> -Grant
>>
>> On Feb 19, 2010, at 6:59 AM, Kristian Waagan wrote:
>>
>>> Hi,
>>>
>>> I don't know if this error requires action to be taken by an administrator, but
the last Derby-trunk build failed with the following stack trace:
>>>
>>> [Kristian Waagan] DERBY-4400: Document the process of producing Maven 2 artifacts
for Derby.
>>>
>>> Minor formatting changes.
>>> Fixed some typos.
>>>
>>> ------------------------------------------
>>> Failed to access build log
>>>
>>> hudson.util.IOException2: remote file operation failed: /export/home/hudson/hudson-slave/workspace/Derby-trunk
athudson.remoting.Channel@84bdd1:lucene.zones.apache.org  (Solaris 10)
>>>        at hudson.FilePath.act(FilePath.java:690)
>>>        at hudson.FilePath.act(FilePath.java:676)
>>>        at hudson.FilePath.toURI(FilePath.java:731)
>>>        at hudson.tasks.MailSender.createFailureMail(MailSender.java:256)
>>>        at hudson.tasks.MailSender.getMail(MailSender.java:133)
>>>        at hudson.tasks.MailSender.execute(MailSender.java:81)
>>>        at hudson.tasks.Mailer.perform(Mailer.java:99)
>>>        at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:19)
>>>        at hudson.model.AbstractBuild$AbstractRunner.perform(AbstractBuild.java:582)
>>>        at hudson.model.AbstractBuild$AbstractRunner.performAllBuildStep(AbstractBuild.java:563)
>>>        at hudson.model.AbstractBuild$AbstractRunner.performAllBuildStep(AbstractBuild.java:550)
>>>        at hudson.model.Build$RunnerImpl.post2(Build.java:152)
>>>        at hudson.model.AbstractBuild$AbstractRunner.post(AbstractBuild.java:528)
>>>        at hudson.model.Run.run(Run.java:1221)
>>>        at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:46)
>>>        at hudson.model.ResourceController.execute(ResourceController.java:88)
>>>        at hudson.model.Executor.run(Executor.java:122)
>>> Caused by: java.io.IOException: Remote call on lucene.zones.apache.org (Solaris
10) failed
>>>        at hudson.remoting.Channel.call(Channel.java:560)
>>>        at hudson.FilePath.act(FilePath.java:683)
>>>        ... 16 more
>>> Caused by: java.lang.OutOfMemoryError: GC overhead limit exceeded
>>>
>>>
>>> Am I correct in assuming this is the Hudson slave process, and not the process
building the code? (I further assume these are two separate processes)
>>> The changes for this build were only text changes to a README...
>>>
>>> I expect another build to be triggered later today, so we'll see if the problem
goes away or not.
>>>
>>>
>>> Regards,
>>> --
>>> Kristian
>>>
>>>
>>>
>>
>>
>>
>
>
>


Mime
View raw message