[Box-Admins] Re: Jenkins (was: Re: [squeak-dev] Re: New Cog VMs available)

[Frank Shearar]

On 3 March 2015 at 14:44, Levente Uzonyi <leves at elte.hu> wrote:
> Moved to the box-admins list.
>
> So, I restarted jenkins, but now I see a few errors in the logs, and charts
> don't seem to work anymore. Any ideas what may be causing it?

I see "Unable to access X. You need to run the web container in the
headless mode. Add -Djava.awt.headless=true" to VM" (and I suspect
that's truncated). That's the guy, right? I've not seen it before!

The wisdom of StackOverflow and Jenkins Wiki ([1]) suggests installing
the ttf-dejavu package.

frank

[1] https://wiki.jenkins-ci.org/display/JENKINS/Jenkins+got+java.awt.headless+problem

> Levente
>
> On Tue, 3 Mar 2015, Levente Uzonyi wrote:
>
>> On Tue, 3 Mar 2015, Frank Shearar wrote:
>>
>>> Yep, this is the problem that I was trying to resolve the other week
>>> with Eliot. I don't think it's a Cog bug, I think it's a bug in our
>>> security settings or something.
>>>
>>> I had a look, things looked OK, but this is well beyond my
>>> rememberings of Linux security, hence why it's still broken.
>>
>>
>> Changes of the files in limits.d are only applied to new sessions. The
>> squeak.conf file was created after jenkins was started. So restarting
>> jenkins should fix the problem.
>>
>> Levente
>>
>>>
>>> frank
>>>
>>> On 2 March 2015 at 17:37, Levente Uzonyi <leves at elte.hu> wrote:
>>>>
>>>> I tried to run this on box3, and it just works as expected.
>>>> But the SqueakTrunk job hasn't produced any artifacts since build #1059.
>>>>
>>>> Levente
>>>>
>>>>
>>>> On Tue, 17 Feb 2015, Frank Shearar wrote:
>>>>
>>>>> On 16 February 2015 at 19:55, Eliot Miranda <eliot.miranda at gmail.com>
>>>>> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Mon, Feb 16, 2015 at 9:12 AM, Frank Shearar
>>>>>> <frank.shearar at gmail.com>
>>>>>> wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Reverted, but I still get the same error.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> What's the last good version?
>>>>>
>>>>>
>>>>>
>>>>> I'm not convinced it's a Cog issue: the last known good was 3164 [1]
>>>>> but the next build showed the "Operation not permitted" problem [2],
>>>>> also with 3164.
>>>>>
>>>>> [1] http://build.squeak.org/job/SqueakTrunk/990/console
>>>>> [2] http://build.squeak.org/job/SqueakTrunk/991/console
>>>>>
>>>>> frank
>>>>>
>>>>>>> frank
>>>>>>>
>>>>>>> On 15 February 2015 at 18:13, Eliot Miranda <eliot.miranda at gmail.com>
>>>>>>> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>> Hi Frank,
>>>>>>>>
>>>>>>>>      can you revert to 3253?  This one should be good.  I'm pretty
>>>>>>>> sure
>>>>>>>> the
>>>>>>>> 3254 VMs are subject to the same inlning bug that I mentioned after
>>>>>>>> the
>>>>>>>> latest VMMaker.oscog commit and am working on fixing now.
>>>>>>>>
>>>>>>>> On Sat, Feb 14, 2015 at 5:47 AM, Frank Shearar
>>>>>>>> <frank.shearar at gmail.com>
>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 12 February 2015 at 20:42, Eliot Miranda
>>>>>>>>> <eliot.miranda at gmail.com>
>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> http://www.mirandabanda.org/files/Cog/VM/VM.r3254/
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> These fix two important bugs in the Spur VMs, so Newspeakers
>>>>>>>>>> should
>>>>>>>>>> update,
>>>>>>>>>> and changes the default text encoding to UTF-8 on linux.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I updated CI with 3254, and now I see errors like this:
>>>>>>>>>
>>>>>>>>> spawning command 1 with timeout 1500 seconds:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> /var/lib/jenkins/workspace/SqueakTrunk/target/cog.r3254/coglinuxht/bin/squeak
>>>>>>>>> -vm-sound-null -vm-display-null
>>>>>>>>> "/var/lib/jenkins/workspace/SqueakTrunk/target/TrunkImage.image"
>>>>>>>>> ../update-image.st
>>>>>>>>> pthread_setschedparam failed: Operation not permitted
>>>>>>>>> Read e.g.
>>>>>>>>> http://www.mirandabanda.org/files/Cog/VM/VM.r3254/README.3254
>>>>>>>>>
>>>>>>>>> As it happens, we do have a /etc/security/limits.d/squeak.conf on
>>>>>>>>> build.squeak.org that looks like this:
>>>>>>>>>
>>>>>>>>> *       hard    rtprio  2
>>>>>>>>> *       soft    rtprio  2
>>>>>>>>>
>>>>>>>>> Thoughts?
>>>>>>>>>
>>>>>>>>> frank
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> best,
>>>>>>>> Eliot
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> best,
>>>>>> Eliot
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>
>>
>