www-builds mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Bayer <andrew.ba...@gmail.com>
Subject Re: Latest status on the hanging Jenkins slaves/builds
Date Tue, 27 May 2014 05:18:07 GMT
So far, at least, things are looking much better - the front page does take
a while to load initially, but I don't see any hanging jobs, which had been
cropping up in less time than this previously, and I don't see the polling
problems either...

A.

On Monday, May 26, 2014, Andrew Bayer <andrew.bayer@gmail.com> wrote:

> Downgraded and restarted - let's see how things go now...
>
> A.
>
>
> On Mon, May 26, 2014 at 11:32 AM, Andrew Bayer <andrew.bayer@gmail.com<javascript:_e(%7B%7D,'cvml','andrew.bayer@gmail.com');>
> > wrote:
>
>> I'm starting work on moving to 1.554.1 - I'd really prefer to go with
>> 1.532.3, but there are just enough installed plugins that are newer than
>> 1.532 that I'll play it safe.
>>
>> A.
>>
>>
>> On Mon, May 26, 2014 at 10:28 AM, Andrew Bayer <andrew.bayer@gmail.com<javascript:_e(%7B%7D,'cvml','andrew.bayer@gmail.com');>
>> > wrote:
>>
>>> That's what's happening across all the slaves with Maven jobs.
>>>
>>> A.
>>>
>>>
>>> On Monday, May 26, 2014, Olivier Lamy <olamy@apache.org<javascript:_e(%7B%7D,'cvml','olamy@apache.org');>>
>>> wrote:
>>>
>>>> Note I have disabled ubuntu 5&6 nodes. As jobs were hanging on those
>>>> nodes.
>>>> Especially when installing artifacts locally, someone to have a look
>>>> if there any io issues on those servers?
>>>>
>>>> On 26 May 2014 14:15, Olivier Lamy <olamy@apache.org> wrote:
>>>> > Go if you have a bit of time.
>>>> > I just finished to downgrade to jdk1.7.0_55 (master and ubuntu
>>>> slaves).
>>>> > Furthermore I have increase memory for all slaves (the slave process
>>>> > is now using -Xmx1024m for all ubuntu slaves)
>>>> >
>>>> > On 26 May 2014 11:16, Andrew Bayer <andrew.bayer@gmail.com> wrote:
>>>> >> Downgrading to the latest LTS definitely seems like a good idea.
Do
>>>> you want
>>>> >> to do that or should I?
>>>> >>
>>>> >> On May 23, 2014 9:51 PM, "Olivier Lamy" <olamy@apache.org>
wrote:
>>>> >>>
>>>> >>> Out of the trendy tendance to say this plugin is crap :-)
>>>> >>> What is the real technical reason?
>>>> >>> I don't see any in the jira issue.
>>>> >>> Why we don't try using LTS?
>>>> >>> The current master is plenty of changes especially in the remoting
>>>> >>> library.
>>>> >>> And revert back masters/slaves to java 1.7 ( I can do it Monday
).
>>>> >>>
>>>> >>> Cheers
>>>> >>> --
>>>> >>> Olivier
>>>> >>>
>>>> >>> On May 24, 2014 4:10 AM, "Andrew Bayer" <abayer@apache.org>
wrote:
>>>> >>>>
>>>> >>>> So I opened https://issues.jenkins-ci.org/browse/JENKINS-23098for
this,
>>>> >>>> and am currently putting together the support bundle mentioned
in
>>>> it. But it
>>>> >>>> looks like there's a very real chance that the Maven plugin/job
>>>> type is just
>>>> >>>> so thoroughly crap at scale that we need to get rid of it
and move
>>>> all jobs
>>>> >>>> to freestyle projects, using Maven build steps instead.
>>>> >>>>
>>>> >>>> If that turns out to be the only real option, it's going
to be an
>>>> ugly
>>>> >>>> transition, given that we have over 600 Maven projects from
many
>>>> different
>>>> >>>> projects, and there's no automatic way to transition those
>>>> projects. I can
>>>> >>>> probably come up with something that'll do a pretty decent
>>>> approximation of
>>>> >>>> that, but it wouldn't be perfect. We may want to consider
tossing
>>>> all the
>>>> >>>> Maven jobs, removing the Maven plugin, and having the projects
>>>> rebuild the
>>>> >>>> jobs themselves.
>>>> >>>>
>>>> >>>> A.
>>>> >
>>>> >
>>>> >
>>>> > --
>>>> > Olivier Lamy
>>>> > Ecetera: http://ecetera.com.au
>>>> > http://twitter.com/olamy | http://linkedin.com/in/olamy
>>>>
>>>>
>>>>
>>>> --
>>>> Olivier Lamy
>>>> Ecetera: http://ecetera.com.au
>>>> http://twitter.com/olamy | http://linkedin.com/in/olamy
>>>>
>>>
>>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message