[Gluster-devel] Netbsd build failure

[Raghavendra Talur]

On Aug 21, 2015 10:22, "Avra Sengupta" <asengupt at redhat.com> wrote:
>
> + Adding Vijaikumar
>
>
> On 08/20/2015 04:19 PM, Niels de Vos wrote:
>>
>> On Thu, Aug 20, 2015 at 03:05:56AM -0400, Susant Palai wrote:
>>>
>>> Hi,
>>>   I tried running netbsd regression twice on a patch. And twice it
failed at the same point. Here is the error:
>>>
>>> <snip>
>>> Build GlusterFS
>>> ***************
>>>
>>> + '/opt/qa/build.sh'
>>>   File "/usr/pkg/lib/python2.7/site.py", line 601
>>>     [2015-08-19 05:45:06.N]:++++++++++
G_LOG:./tests/basic/quota-anon-fd-nfs.t: TEST: 85 ! fd_write 3 content
++++++++++

Hypothesis with two things I know..

1. Ravishankar recently added G_LOG support to test setup where before any
TEST line is executed in a .t file we append every log in gluster log dir
with a special line as seen above,  it shows the test script name,  test
number,  and actual command.

2. I recently changed the function to redirect it's stderr (FD 2) to dev
null.

3. We always see log lines from this particular script in python files etc.
And it is trying to do a FD write on 3.

Some assumption with FD number in shell is wrong some where. Putting it out
so that anybody who had better idea can debug faster, I will look into it
too.

Thanks,
Raghavendra Talur

>
> This particular test is currently in bad test and I believe Vijaikumar is
looking into it. Could you please make sure if there is any other
failure(apart from this), which is failing the regression runs.
>>>
>>>            ^
>>> SyntaxError: invalid token
>>> + RET=1
>>> + '[' 1 '!=' 0 ']'
>>> + exit 1
>>> Build step 'Ex?cuter un script shell' marked build as failure
>>> Finished: FAILURE
>>> </snip>
>>>
>>>  Requesting you to take look into it.
>>
>> Which Jenkins slave was this? Got a link to the job that failed?
>>
>> This looks again like a NetBSD slave where logs from regression tests
>> are overwriting random files. The /usr/pkg/lib/python2.7/site.py file
>> should be valid Python, and not contain these logs...
>>
>> Anyone has any ideas why this happens?
>>
>> Thanks,
>> Niels
>>
>>
>>
>> _______________________________________________
>> Gluster-devel mailing list
>> Gluster-devel at gluster.org
>> http://www.gluster.org/mailman/listinfo/gluster-devel
>
>
>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-devel/attachments/20150821/f859dc2b/attachment.html>