[openstack-dev] [nova] readout from Philly Operators Meetup

Robert Collins robertc at robertcollins.net
Wed Mar 11 21:18:34 UTC 2015


On 12 March 2015 at 10:09, Sean Dague <sean at dague.net> wrote:
> On 03/11/2015 11:26 AM, Clark Boylan wrote:
>> On Wed, Mar 11, 2015, at 05:59 AM, Sean Dague wrote:
>>> Galera Upstream Testing
>>> -----------------------
>>>
>>> The majority of deploys run with Galera MySQL. There was a question
>>> about whether or not we could get that into upstream testing pipeline
>>> as that's the common case.
>>>
>> It should be possible to run a two node galera cluster with garbd in the
>> current two node test environment. Or run an extra mysql on one of the
>> two nodes for a total of three database daemons (likely the compute node
>> as the controller already has a bunch of memory stealing daemons
>> running).
>>
>> These aiopcpu (all in one plus compute node) tests run against
>> devstack-gate, devstack, tempest, nova, and neutron's experimental
>> queues and can be triggered by leaving a `check experimental` comment on
>> changes belonging to any of these projects. If we need to add the job to
>> the experimental queues of more projects we can do that too.
>
> Can we sanely run a 2 node galera in a single node devstack? It might
> make for a more interesting default there given that time and again this
> seems to be the deployment strategy for nearly everyone, and there are
> some edge case differences that Jay Pipes has highlighted that would be
> nice to be more in front of everyone.

Yes. Its just a matter of memory and applying the effort to set it up.

-Rob


-- 
Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Converged Cloud



More information about the OpenStack-dev mailing list