[Gluster-devel] NetBSD regression tests not Initializing...

[Niels de Vos]

On Tue, Jul 07, 2015 at 06:04:44PM +0200, Niels de Vos wrote:
> On Tue, Jul 07, 2015 at 07:13:53PM +0530, Kaushal M wrote:
> > I've taken this slave and one other offline and am rebooting it.
> 
> Reminder that you do not need to take teh system offline for rebooting.
> I normally follow these steps to get hung systems back functional:
> 
> 1. verify stuck job, NFS unmount related?
> 2. open http://build.gluster.org/view/Infra/job/reboot-vm/build
> 3. login on Jenkins
> 4. start the reboot-vm job for the stuck system
> 5. wait until the job finished
> 6. click the "abort" [x] link on the stuck job
> 7. retrigger the job after aborting has been done (reload page)
> 
> These hangs do not seem to happen on tests from the master branch
> anymore, only on release-3.7. I think this is a confirmation that the
> reference counting for auth-cache structures in gluster/nfs is a working
> solution.
> 
> We should backport these changes:
> 
> - nfs: add a gf_lock_t for the auth_cache->cache_dict
>   http://review.gluster.org/11021
> 
> - core: add "gf_ref_t" for common refcounting structures
>   http://review.gluster.org/11022
>   (already done through http://review.gluster.org/11421)
> 
> - nfs: refcount each auth_cache_entry and related data_t
>   http://review.gluster.org/11023
> 
> - refcount: correct the documentation
>   http://review.gluster.org/11328
> 
> 
> I'll try to send backports later this week (maybe Thursday?), unless
> someone else beats me to it. Please reply to this thread if you file a
> bug for this and send some backports.

The above backports have been posted. These should prevent the
Gluster/NFS crashes in the regression tests, and therefor prevent the
hanging of NetBSD on unmounting NFS (when the NFS-server died).

Please check these patches, and merge them when ready:

   http://review.gluster.org/#/q/status:open+project:glusterfs+branch:release-3.7+topic:bug-1242515

Thanks,
Niels

> 
> Thanks,
> Niels
> 
> > 
> > On Tue, Jul 7, 2015 at 6:44 PM, Kotresh Hiremath Ravishankar
> > <khiremat at redhat.com> wrote:
> > > Hi Emmanuel,
> > >
> > > We are seeing these issues again on nbslave7h.cloud.gluster.org
> > > http://build.gluster.org/job/rackspace-netbsd7-regression-triggered/7974/console
> > >
> > > Thanks and Regards,
> > > Kotresh H R
> > >
> > > ----- Original Message -----
> > >> From: "Emmanuel Dreyfus" <manu at netbsd.org>
> > >> To: "Kotresh Hiremath Ravishankar" <khiremat at redhat.com>, "Gluster Devel" <gluster-devel at gluster.org>
> > >> Sent: Sunday, July 5, 2015 12:52:23 AM
> > >> Subject: Re: [Gluster-devel] NetBSD regression tests not Initializing...
> > >>
> > >> Kotresh Hiremath Ravishankar <khiremat at redhat.com> wrote:
> > >>
> > >> > Any help is appreciated.
> > >>
> > >> nbslave72 was sick indeed: it refused SSH connexions. I rebooted it and
> > >> retiggered your change, but it went on another machine.
> > >>
> > >> --
> > >> Emmanuel Dreyfus
> > >> http://hcpnet.free.fr/pubz
> > >> manu at netbsd.org
> > >>
> > > _______________________________________________
> > > Gluster-devel mailing list
> > > Gluster-devel at gluster.org
> > > http://www.gluster.org/mailman/listinfo/gluster-devel
> > _______________________________________________
> > Gluster-devel mailing list
> > Gluster-devel at gluster.org
> > http://www.gluster.org/mailman/listinfo/gluster-devel



> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-devel