You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@reef.apache.org by Markus Weimer <ma...@weimo.de> on 2015/08/06 17:43:40 UTC
Ubuntu build server git complaints
Hi,
CI is the gift that keeps on giving. Most recently, we've seen a whole
bunch of build failures involving git, e.g.[0]. Usually, they are
"easily fixed" when the contributor rebases their branch upon current
master.
I wonder why we don't see the same on the windows side? Could it be that
we hosed some file name capitalizations, which Windows conveniently ignores?
Markus
[0]: https://builds.apache.org/job/Reef-pull-request-ubuntu/250/console
Re: Ubuntu build server git complaints
Posted by Brian Cho <ch...@gmail.com>.
Andrew:
After login on the top-right, the Rebuild option appears on the left
sidebar. It may be that login is only available to committers though.
Markus:
It looks like the “just try another Rebuild” workaround isn’t working very
well for recent builds. We’ve also noticed some general insanity. In [1]
the only change was to a script file, but Jenkins decided that changes from
February were relevant.
Changes
1. [REEF-118 <https://issues.apache.org/jira/browse/REEF-118>] Add
Shimoga library for elastic group communication. (commit:
45f731809cada3fd72af2c3d588811d36abb642e
<https://github.com/apache/incubator-reef/commit/45f731809cada3fd72af2c3d588811d36abb642e>)
(detail
<https://builds.apache.org/job/Reef-pull-request-ubuntu/244/changes#detail0>
)
2. move tron.math from shimoga examples to utils.math and use it for all
examples. Have not moved stuff into reef-examples and reef-io yet (commit:
3364cb59593568e6e51802b0e9be3986ce11198d
<https://github.com/apache/incubator-reef/commit/3364cb59593568e6e51802b0e9be3986ce11198d>)
(detail
<https://builds.apache.org/job/Reef-pull-request-ubuntu/244/changes#detail1>
)
[1] https://builds.apache.org/job/Reef-pull-request-ubuntu/244/
On Fri, Aug 7, 2015 at 5:17 AM, Andrew Chung <af...@gmail.com> wrote:
> Hi Brian,
>
> Would you share how to rebuild? I was not able to find the rebuild button
> in the UI.
>
> Thanks,
> Andrew
>
> On Thu, Aug 6, 2015 at 9:04 AM, Markus Weimer <ma...@weimo.de> wrote:
>
> > On 2015-08-06 08:48, Brian Cho wrote:
> > > Markus, About the symptoms -- I didn't even have to rebase for [0]
> > > (and for others as well). I just clicked rebuild on 250 and it
> > > produced a good 251.
> >
> > Good to know. I'll do that next time.
> >
> > > I don't really have a good idea about the diagnosis though. I wonder
> > > if our git repo size slows things down?
> >
> > Maybe. Do we clone each time we build? The repo is ~32MB right now.
> >
> > The issue raised on failed build, e.g. 250, seem pseudo-legit:
> >
> > > Caused by: hudson.plugins.git.GitException: Command "git checkout -b
> > master origin/master" returned status code 1:
> > > stdout: lang/cs/Org.Apache.REEF.Common/Avro/AvroHttpSerializer.cs:
> needs
> > merge
> > > lang/cs/Org.Apache.REEF.Driver/Evaluator/IEvaluatorRequest.cs: needs
> > merge
> > > lang/cs/Org.Apache.REEF.Driver/IDriverStarted.cs: needs merge
> > >
> >
> lang/cs/Org.Apache.REEF.Evaluator/Exceptions/EvaluatorConfigurationParseException.cs:
> > needs merge
> > >
> >
> lang/cs/Org.Apache.REEF.IMRU/Examples/MapperCount/IntSumReduceFunction.cs:
> > needs merge
> > > lang/cs/Org.Apache.REEF.Network/Group/Driver/Impl/MessageType.cs: needs
> > merge
> > > lang/cs/Org.Apache.REEF.Wake/Remote/Parameters/TcpPortRangeSeed.cs:
> > needs merge
> >
> > Markus
> >
>
Re: Ubuntu build server git complaints
Posted by Andrew Chung <af...@gmail.com>.
Hi Brian,
Would you share how to rebuild? I was not able to find the rebuild button
in the UI.
Thanks,
Andrew
On Thu, Aug 6, 2015 at 9:04 AM, Markus Weimer <ma...@weimo.de> wrote:
> On 2015-08-06 08:48, Brian Cho wrote:
> > Markus, About the symptoms -- I didn't even have to rebase for [0]
> > (and for others as well). I just clicked rebuild on 250 and it
> > produced a good 251.
>
> Good to know. I'll do that next time.
>
> > I don't really have a good idea about the diagnosis though. I wonder
> > if our git repo size slows things down?
>
> Maybe. Do we clone each time we build? The repo is ~32MB right now.
>
> The issue raised on failed build, e.g. 250, seem pseudo-legit:
>
> > Caused by: hudson.plugins.git.GitException: Command "git checkout -b
> master origin/master" returned status code 1:
> > stdout: lang/cs/Org.Apache.REEF.Common/Avro/AvroHttpSerializer.cs: needs
> merge
> > lang/cs/Org.Apache.REEF.Driver/Evaluator/IEvaluatorRequest.cs: needs
> merge
> > lang/cs/Org.Apache.REEF.Driver/IDriverStarted.cs: needs merge
> >
> lang/cs/Org.Apache.REEF.Evaluator/Exceptions/EvaluatorConfigurationParseException.cs:
> needs merge
> >
> lang/cs/Org.Apache.REEF.IMRU/Examples/MapperCount/IntSumReduceFunction.cs:
> needs merge
> > lang/cs/Org.Apache.REEF.Network/Group/Driver/Impl/MessageType.cs: needs
> merge
> > lang/cs/Org.Apache.REEF.Wake/Remote/Parameters/TcpPortRangeSeed.cs:
> needs merge
>
> Markus
>
Re: Ubuntu build server git complaints
Posted by Byung-Gon Chun <bg...@gmail.com>.
I think we clone each time we build. The network bw from the jenkins server
seems to be small.
The current timeout for git operations (clone, checkout) is 10 minutes. We
should increase it.
On Fri, Aug 7, 2015 at 1:04 AM, Markus Weimer <ma...@weimo.de> wrote:
> On 2015-08-06 08:48, Brian Cho wrote:
> > Markus, About the symptoms -- I didn't even have to rebase for [0]
> > (and for others as well). I just clicked rebuild on 250 and it
> > produced a good 251.
>
> Good to know. I'll do that next time.
>
> > I don't really have a good idea about the diagnosis though. I wonder
> > if our git repo size slows things down?
>
> Maybe. Do we clone each time we build? The repo is ~32MB right now.
>
> The issue raised on failed build, e.g. 250, seem pseudo-legit:
>
> > Caused by: hudson.plugins.git.GitException: Command "git checkout -b
> master origin/master" returned status code 1:
> > stdout: lang/cs/Org.Apache.REEF.Common/Avro/AvroHttpSerializer.cs: needs
> merge
> > lang/cs/Org.Apache.REEF.Driver/Evaluator/IEvaluatorRequest.cs: needs
> merge
> > lang/cs/Org.Apache.REEF.Driver/IDriverStarted.cs: needs merge
> >
> lang/cs/Org.Apache.REEF.Evaluator/Exceptions/EvaluatorConfigurationParseException.cs:
> needs merge
> >
> lang/cs/Org.Apache.REEF.IMRU/Examples/MapperCount/IntSumReduceFunction.cs:
> needs merge
> > lang/cs/Org.Apache.REEF.Network/Group/Driver/Impl/MessageType.cs: needs
> merge
> > lang/cs/Org.Apache.REEF.Wake/Remote/Parameters/TcpPortRangeSeed.cs:
> needs merge
>
> Markus
>
--
Byung-Gon Chun
Re: Ubuntu build server git complaints
Posted by Markus Weimer <ma...@weimo.de>.
On 2015-08-06 08:48, Brian Cho wrote:
> Markus, About the symptoms -- I didn't even have to rebase for [0]
> (and for others as well). I just clicked rebuild on 250 and it
> produced a good 251.
Good to know. I'll do that next time.
> I don't really have a good idea about the diagnosis though. I wonder
> if our git repo size slows things down?
Maybe. Do we clone each time we build? The repo is ~32MB right now.
The issue raised on failed build, e.g. 250, seem pseudo-legit:
> Caused by: hudson.plugins.git.GitException: Command "git checkout -b master origin/master" returned status code 1:
> stdout: lang/cs/Org.Apache.REEF.Common/Avro/AvroHttpSerializer.cs: needs merge
> lang/cs/Org.Apache.REEF.Driver/Evaluator/IEvaluatorRequest.cs: needs merge
> lang/cs/Org.Apache.REEF.Driver/IDriverStarted.cs: needs merge
> lang/cs/Org.Apache.REEF.Evaluator/Exceptions/EvaluatorConfigurationParseException.cs: needs merge
> lang/cs/Org.Apache.REEF.IMRU/Examples/MapperCount/IntSumReduceFunction.cs: needs merge
> lang/cs/Org.Apache.REEF.Network/Group/Driver/Impl/MessageType.cs: needs merge
> lang/cs/Org.Apache.REEF.Wake/Remote/Parameters/TcpPortRangeSeed.cs: needs merge
Markus
Re: Ubuntu build server git complaints
Posted by Brian Cho <ch...@gmail.com>.
Markus,
About the symptoms -- I didn't even have to rebase for [0] (and for others as well). I just clicked rebuild on 250 and it produced a good 251.
I don't really have a good idea about the diagnosis though. I wonder if our git repo size slows things down?
Thanks,Brian
On Thu, Aug 6, 2015 at 8:44 AM -0700, "Markus Weimer" <ma...@weimo.de> wrote:
Hi,
CI is the gift that keeps on giving. Most recently, we've seen a whole
bunch of build failures involving git, e.g.[0]. Usually, they are
"easily fixed" when the contributor rebases their branch upon current
master.
I wonder why we don't see the same on the windows side? Could it be that
we hosed some file name capitalizations, which Windows conveniently ignores?
Markus
[0]: https://builds.apache.org/job/Reef-pull-request-ubuntu/250/console