[Gluster-infra] NetBSD slaves

Vijay Bellur vbellur at redhat.com
Tue Jun 9 16:07:02 UTC 2015

On 06/09/2015 09:33 PM, Niels de Vos wrote:
> On Fri, Jun 05, 2015 at 09:35:19PM +0200, Emmanuel Dreyfus wrote:
>> Vijay Bellur <vbellur at redhat.com> wrote:
>>> Rebooting is not working effectively for slaves like slave74. I have
>>> created 3 new VMs nbslave7{d..f} for load balancing NetBSD regression
>>> queue. If necessary we can spin a few ephemeral VMs over the weekend to
>>> drain the queue.
>> FWIW I planned nbslave7g after nbslave7f :-)
>> When it does not reboot, a peek at the console may help. I guess this is
>> a fsck problem.
> Just wondering if the VMs were created, and not only Jenkins Slaves?
> These Jenkins slaves were all using nbslave71.cloud.gluster.org:
>   - nbslave7f.cloud.gluster.org
>   - nbslave7e.cloud.gluster.org
>   - nbslave7d.cloud.gluster.org
> Running multiple smoke/regression tests at the same time on one VM will
> surely result in very strange problems.
> I've disabled the above Jenkins Slaves. If VMs have been created, please
> update the configuration of these slaves to point to the right hostname.

Thanks, Niels! This certainly is my bad and apologies for all the 
inconvenience caused.

This certainly does explain the baffling behavior of regression runs 
getting overrun and confirm my borked understanding of creating slave VMs.


More information about the Gluster-infra mailing list