hadoop-yarn-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wangda Tan <wheele...@gmail.com>
Subject Re: YARN precommit busted?
Date Tue, 03 Jan 2017 21:11:16 GMT
Hi Sean,

I was trying to investigate the issue yesterday as well. I only found
existing JIRA to track the issue is on windows:
https://groups.google.com/forum/#!topic/jenkinsci-issues/BgcrxQqObj4.

I cannot find HADOOP/YARN JIRA, please feel free to create one.

Thanks,
Wangda


On Tue, Jan 3, 2017 at 1:04 PM, Sean Busbey <busbey@cloudera.com> wrote:

> historically that git clean failure thing was a problem where an test
> purposefully set bad permissions to simulate a disk failure that
> resulted in a ro filesystem. I had thought all of the hadoop-related
> precommit jobs had special handling in place to account for this test
> now.
>
> Is there a JIRA in HADOOP or YARN tracking this issue yet? I can chase
> down the problem pretty fast if it's just that old problem.
>
> On Tue, Jan 3, 2017 at 2:38 PM, Wangda Tan <wheeleast@gmail.com> wrote:
> > Thanks Ted!
> >
> > On Tue, Jan 3, 2017 at 12:14 PM, Ted Yu <yuzhihong@gmail.com> wrote:
> >
> >> With continuous failed QA runs on H5 and H6, I have to take them out
> again.
> >>
> >> The INFRA has bounced these two machines. Unfortunately that didn't
> solve
> >> the issue.
> >>
> >> On Fri, Dec 30, 2016 at 2:42 PM, Ted Yu <yuzhihong@gmail.com> wrote:
> >>
> >> > Put back H5 and H6 since the INFRA ticket has been resolved.
> >> > Excluded H9.
> >> >
> >> > On Fri, Dec 30, 2016 at 2:36 PM, Arun Suresh <arun.suresh@gmail.com>
> >> > wrote:
> >> >
> >> >> It looks like jobs that land on H9 are all getting stuck at:
> >> >>
> >> >> ...
> >> >>
> >> >> Untagged images:
> >> >> Attempting to remove docker image sha256:
> 4d4f8bf49bcfb5cea087e6bd6008e8
> >> 04f5f28a994b2dc513c2e5c4017679f8f7
> >> >>
> >> >>
> >> >> Can we exclude H9 as well (or maybe fix it somehow) ?
> >> >>
> >> >> Cheers
> >> >> -Arun
> >> >>
> >> >>
> >> >> On Wed, Dec 21, 2016 at 12:22 PM, Ted Yu <yuzhihong@gmail.com>
> wrote:
> >> >>
> >> >>> Exclusion has already been done.
> >> >>>
> >> >>> Precommit build #14423 has completed.
> >> >>>
> >> >>> On Wed, Dec 21, 2016 at 12:02 PM, Arun Suresh <
> arun.suresh@gmail.com>
> >> >>> wrote:
> >> >>>
> >> >>>> Actually I found H6 working yesterday and H5 bad,
> >> >>>>
> >> >>>> >  consider excluding H6 from Precommit builds until this
problem
> >> >>>> Can someone point me to how this can be done (I am assuming
this is
> >> >>>> some sort of exclusion list) before kicking of Jenkins job
?
> >> >>>>
> >> >>>> Cheers
> >> >>>> -Arun
> >> >>>>
> >> >>>>
> >> >>>>
> >> >>>> On Wed, Dec 21, 2016 at 9:09 AM, Ted Yu <yuzhihong@gmail.com>
> wrote:
> >> >>>>
> >> >>>>> All these builds took place on :
> >> >>>>>
> >> >>>>> https://builds.apache.org/computer/H6
> >> >>>>>
> >> >>>>> Maybe log an INFRA ticket so that some admin can check
what
> happens
> >> to
> >> >>>>> hadoop-common-project/hadoop-common/target/test/data/4
> >> >>>>>
> >> >>>>> Meanwhile, consider excluding H6 from Precommit builds
until this
> >> >>>>> problem
> >> >>>>> is solved.
> >> >>>>>
> >> >>>>> On Wed, Dec 21, 2016 at 9:02 AM, Sangjin Lee <sjlee@apache.org>
> >> wrote:
> >> >>>>>
> >> >>>>> > It appears that the YARN precommit builds are broken
for the
> most
> >> >>>>> part. I
> >> >>>>> > see quick failures with the following message:
> >> >>>>> >
> >> >>>>> > FATAL: Command "git clean -fdx" returned status code
1:
> >> >>>>> > stdout:
> >> >>>>> > stderr: warning: failed to remove
> >> >>>>> > hadoop-common-project/hadoop-common/target/test/data/4
> >> >>>>> > hudson.plugins.git.GitException
> >> >>>>> > <http://stacktrace.jenkins-ci.org/search?query=hudson.
> >> >>>>> > plugins.git.GitException>:
> >> >>>>> > Command "git clean -fdx" returned status code 1:
> >> >>>>> >
> >> >>>>> >
> >> >>>>> > Any ideas what might be going on?
> >> >>>>> >
> >> >>>>> > https://builds.apache.org/job/PreCommit-YARN-Build/
> >> >>>>> >
> >> >>>>> > Thanks,
> >> >>>>> > Sangjin
> >> >>>>> >
> >> >>>>>
> >> >>>>
> >> >>>>
> >> >>>
> >> >>
> >> >
> >>
>
>
>
> --
> busbey
>

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