river-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Patricia Shanahan <p...@acm.org>
Subject Re: Build failed in Hudson: River-trunk-QA #3
Date Tue, 31 Aug 2010 14:38:28 GMT
Not specific, not yet. Maybe related to lookup, events and the like.

However, given Peter's comment about RemoteEvent being involved in the 
problem he was working, this may be a different manifestation, with 
different timings, of the same problem. Maybe we put this on aside until 
Peter checks in a fix for that, and see if it goes away.

Patricia


Jonathan Costers wrote:
> Hi Patricia
> Since Peter is still probably sleeping, I will adjust the logging levels for
> the QA suite in a cpl hours and have the QA harness generate HTML reports
> too.
> Did you have any specific loggers and/or log levels in mind?
> Thanks
> Jonathan
> 
> 2010/8/31 Peter Firmstone <jini@zeus.net.au>
> 
>> Haven't reached the point of failure yet, but I'm sure I'll find it soon.
>>
>> Cheers,
>>
>> Peter.
>>
>>    [java] -----------------------------------------
>>    [java]
>>    [java] # of tests started   = 40
>>    [java] # of tests completed = 40
>>    [java] # of tests passed    = 40
>>
>>    [java] # of tests failed    = 0
>>    [java]
>>    [java] -----------------------------------------
>>    [java]
>>    [java]    Date finished:
>>    [java]       Tue Aug 31 20:30:26 EST 2010
>>    [java]    Time elapsed:
>>    [java]       5200 seconds
>>    [java]
>>
>> BUILD SUCCESSFUL
>> Total time: 86 minutes 46 seconds
>>
>>
>>
>> Peter Firmstone wrote:
>>
>>> Thanks Patricia,
>>>
>>> I can update the qa.logging file temporarily later, I'm working on an
>>> earlier build at them moment though.
>>>
>>> I'm making some progress, with the incremental changes. I'm updating small
>>> sections of code, followed by clean builds and test, runs, it's taking some
>>> time.
>>>
>>> This could be a particularly tricky bug to nail down.
>>>
>>> Regards,
>>>
>>> Peter.
>>>
>>> Patricia Shanahan wrote:
>>>
>>>> Unfortunately, it turns out my failure is not the same as the Hudson
>>>> failure, so I have no way of making progress without more data.
>>>>
>>>> Is it possible to get logs from Hudson runs? Can somebody with Hudson
>>>> access collect more data?
>>>>
>>>> Patricia
>>>>
>>>>
>>>> On 8/30/2010 3:24 PM, Patricia Shanahan wrote:
>>>>
>>>>> The com/sun/jini/test/spec/lookupdiscovery/MulticastMonitorAllChange.td
>>>>> failure reproduces in my VirtualBox/Ubuntu environment. I checked out
>>>>> the latest revision, and get 10 failures on 10 tries.
>>>>>
>>>>> Is there any information you would like me to collect and report from
>>>>> it, always assuming data collection does not make the failure go away?
>>>>>
>>>>> I'm also running a servicediscovery test in another VirtualBox. I'll
>>>>> report results when that finishes.
>>>>>
>>>>> Patricia
>>>>>
>>>>>
>>>>> On 8/30/2010 1:28 PM, Jonathan Costers wrote:
>>>>>
>>>>>> Im not able to reproduce the
>>>>>> com/sun/jini/test/spec/lookupdiscovery/MulticastMonitorAllChange.td
>>>>>> failure
>>>>>> either on my machine ...
>>>>>> I just started a new build on Hudson. Hopefully it can reproduced
>>>>>> there ...
>>>>>>
>>>>>> 2010/8/30 Jonathan Costers<jonathan.costers@googlemail.com>
>>>>>>
>>>>>>  Did you try to run the servicediscovery category? I now get 15 test
>>>>>>> failures consistently ... see attached report.
>>>>>>>
>>>>>>> 2010/8/30 Peter Firmstone<jini@zeus.net.au>
>>>>>>>
>>>>>>> I've noticed that among the tests there are inconsistencies related
to
>>>>>>>
>>>>>>>> RemoteEvent's, sometimes more events than expected are recieved
>>>>>>>> (multiple of
>>>>>>>> two) and other time no event is received when expected. I'll
post
>>>>>>>> some
>>>>>>>> detailed test results later today.
>>>>>>>>
>>>>>>>>
>>>>>>>> Peter Firmstone wrote:
>>>>>>>>
>>>>>>>>  Are there any details, I can't seem to replicate it?
>>>>>>>>> As I get time this week, it might take a little longer,
but I'll be
>>>>>>>>> working from a known stable state, slowly adding the
changes, until
>>>>>>>>> I
>>>>>>>>> discover the failure.
>>>>>>>>>
>>>>>>>>> The issue seems to be Event based, as are the other issues
that are
>>>>>>>>> occurring.
>>>>>>>>>
>>>>>>>>> Regards,
>>>>>>>>>
>>>>>>>>> Peter.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> [java] -----------------------------------------
>>>>>>>>> [java]
>>>>>>>>> [java] # of tests started = 497
>>>>>>>>> [java] # of tests completed = 497
>>>>>>>>> [java] # of tests skipped = 22
>>>>>>>>> [java] # of tests passed = 497
>>>>>>>>> [java] # of tests failed = 0
>>>>>>>>> [java]
>>>>>>>>> [java] -----------------------------------------
>>>>>>>>> [java]
>>>>>>>>> [java] Date finished:
>>>>>>>>> [java] Tue Aug 31 03:38:24 EST 2010
>>>>>>>>> [java] Time elapsed:
>>>>>>>>> [java] 15780 seconds
>>>>>>>>> [java]
>>>>>>>>>
>>>>>>>>> BUILD SUCCESSFUL
>>>>>>>>> Total time: 291 minutes 59 seconds
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Patricia Shanahan wrote:
>>>>>>>>>
>>>>>>>>>  Jonathan Costers wrote:
>>>>>>>>>>  OK, this one is ligitimate ...
>>>>>>>>>>> The changes that were committed yesterday apparently
cause a
>>>>>>>>>>> single QA
>>>>>>>>>>> test
>>>>>>>>>>> to fail:
>>>>>>>>>>>
>>>>>>>>>>> [java] com/sun/jini/test/spec/
>>>>>>>>>>> lookupdiscovery/MulticastMonitorAllChange.td
>>>>>>>>>>> [java] Test Failed: Test Failed:
>>>>>>>>>>> com.sun.jini.qa.harness.TestException:
>>>>>>>>>>> change failed -- waited 870 seconds (14 minutes)
-- 3 change
>>>>>>>>>>> event(s)
>>>>>>>>>>> expected, 0 change event(s) received
>>>>>>>>>>>
>>>>>>>>>>> Note that this QA run did not include "servicediscovery"
nor any
>>>>>>>>>>> other
>>>>>>>>>>> new
>>>>>>>>>>> test categories.
>>>>>>>>>>> The same tests were run as were run when the
QA run (build #1)
>>>>>>>>>>> passed
>>>>>>>>>>> yesterday.
>>>>>>>>>>>
>>>>>>>>>>> Any chance this can get some more attention?
>>>>>>>>>>> IMHO, getting this fixed is our top priority
right now.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>  I agree with the priority within River, but
I am spending the
>>>>>>>>>> next
>>>>>>>>>> few
>>>>>>>>>> hours on something even higher priority - going ceramic
painting
>>>>>>>>>> with a
>>>>>>>>>> friend. I'll check the mailing list when I get back.
If nothing
>>>>>>>>>> else
>>>>>>>>>> happens, I'll look into it this afternoon or evening.
>>>>>>>>>>
>>>>>>>>>> If you have time, could you check whether it is a
solid failure or
>>>>>>>>>> intermittent?
>>>>>>>>>>
>>>>>>>>>> Patricia
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>
>>>>
>>>
> 


Mime
View raw message