[Gluster-devel] Running Vagrant tests on the CentOS CI (WAS: Re: 3.7.9 update)

[Kaushal M]

On Tue, Mar 15, 2016 at 6:34 PM, Jeff Darcy <jdarcy at redhat.com> wrote:
>> Yes, Right now we get the machine from duffy instantly. If it is not instant,
>> it means no machines are in ready state and we will have to wait few more
>> minutes (~5 minutes) to get them.
>
> If machines are available to be allocated instantly, does that mean they're
> being re-used across allocations?  Or are there just so many extra machines
> in the pool that there's time to re-provision in the background before they
> get re-used again?

The have a set of machines on hot-standby, which are ready to be
allocated when requested.
Machines are provisioned in the background and added to this set.
If the number of requested machines exceeds the set, then the extra
systems will provisioned on demand, which should take about ~5
minutes.

>
>> But I think vagrant is very useful anywhere else (in non CentOS CI infra).
>> But I wasn't sure *in* CentOS CI, since the machines required for distaf are
>> available via duffy. So I wasn't sure where vagrant would fit in.
>>
>> But if the plan is to run the existing regression tests by spawning vms
>> inside the machines provided by duffy, then I think it can be done. Although
>> someone will have to try that out see if it provides any advantage.
>
> I think the advantages would be two-fold.
>
> (a) The environment within a vagrant box is one we can control more than
> might be possible within duffy.
>
> (b) It's an environment that we can replicate with certainty in other
> host environments where duffy doesn't exist.
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-devel