[Gluster-devel] logs/cores for smoke failures

Nigel Babu nigelb at redhat.com
Fri Sep 30 10:30:55 UTC 2016


On Tue, Sep 27, 2016 at 01:13:15PM +0530, Nigel Babu wrote:
> On Tue, Sep 27, 2016 at 12:52:45PM +0530, Pranith Kumar Karampuri wrote:
> > On Tue, Sep 27, 2016 at 12:39 PM, Nigel Babu <nigelb at redhat.com> wrote:
> >
> > > On Tue, Sep 27, 2016 at 12:00:40PM +0530, Pranith Kumar Karampuri wrote:
> > > > On Tue, Sep 27, 2016 at 11:20 AM, Nigel Babu <nigelb at redhat.com> wrote:
> > > >
> > > > > These are gbench failures rather than smoke failures. If you know how
> > > to
> > > > > debug dbench failures, please add comments on the bug and I'll get you
> > > the
> > > > > logs you need.
> > > > >
> > > >
> > > > Oh, we can't archive the logs like we do for regression runs?
> > >
> > > We don't log anything for smoke tests. Perhaps we should. Would you care to
> > > send a patch for smoke.sh[1] so we log the appropriate files?
> > >
> >
> > hmm... I see that gluster is launched normally so it should log fine. I
> > guess I didn't understand the question.
>
> The regression runs log quite a lot of things in line with what they test. The
> smoke test runs just two things - posix compliance tests and dbench. Dbench is
> the bit that's failing for us. The dbench output is printed onto the screen if
> it fails.
>
> What logs do you want to add to smoke jobs? If you want Gluster logs, write the
> patch to get those cleared before the start of smoke test, archived afterward,
> and dumped into /archives?

Smoke tests now have gluster logs. I intend to have gluster statedump output as
well, but I need more to get that correctly.

--
nigelb


More information about the Gluster-devel mailing list