[openstack-dev] [tripleo] Pacemaker + containers CI

Jiří Stránský jistr at redhat.com
Wed Aug 30 08:26:49 UTC 2017


On 29.8.2017 17:12, Emilien Macchi wrote:
> On Tue, Aug 29, 2017 at 2:14 AM, Jiří Stránský <jistr at redhat.com> wrote:
> [...]
>> the CI for containerized deployments with Pacemaker is close! In fact, it
>> works [1][2] (but there are pending changes to merge).
> 
> Really good news, thanks for the update!
> 
>> The way it's proposed in gerrit currently is to switch the
>> centos-7-containers-multinode job (featureset010) to deploy with Pacemaker.
>> What do you think about making this switch as a first step? [...]
> 
> I'm ok with the idea

No -1s yet, so i removed WIP status of [4].

> as long as
> gate-tripleo-ci-centos-7-containers-multinode-upgrades-nv keep working
> fine.

That's a different featureset so we can control it independently from 
the basic deployment job. It might be good to switch this one to 
Pacemaker too, if we can solve the current timeout issues and perhaps 
have some spare wall time.

Non-pacemaker containers are still CI'd by OVB job, so the upgrade job 
(currently still non-Pacemaker) shouldn't be more vulnerable even if we 
switch the multinode job to Pacemaker.

> Deploying Pacemaker on a single node environment is not optimal but
> already cover a bunch of code which is good.
> 
>> Later it would be nice to get a proper clustering test with 3 controllers.
>> Should we try and switch the centos-7-ovb-ha-oooq job to deploy containers
>> on master and stable/pike? (Probably by adding a new job that only runs on
>> master + Pike, and making the old ovb-ha-oooq only run upto Ocata, to keep
>> the OVB capacity demands unchanged?) I'd be +1 on that since containers are
>> the intended way of deploying Pike and beyond. WDYT?
> 
> It's actually a good start to our discussion at the PTG:
> https://etherpad.openstack.org/p/tripleo-ptg-queens-ci-related-topics
> (we have a session on Wednesday morning about CI topics, please make
> sure you can join!)
> 
> I think in Queens, we'll run container-only jobs, even for OVB.
> That said, I think OVB coverage in Queens will be very useful to try
> HA with 3 controllers (containerized) and the baremetal services
> coverage will only run on Pike, Ocata and Newton.
> 
> That way, we would have:
> 
> Queens:
> - multinode jobs covering basic HA scenario, single node but still
> useful to test a good part of the code
> - OVB jobs covering production environment and hopefully spot issues
> we wouldn't see with multinode jobs
> 
> Pike, Ocata, Newton:
> no change on OVB job
> 
> (note it's a proposal, not a statement)

Yea focusing the CI changes towards containerized mainly on Queens+ 
could be fine too. The frequency of patches going into stable/pike will 
be dropping as it gains stability, so time spent on CI enhancements 
might indeed be better focused on Queens+. We can always adjust if that 
doesn't prove to be the case.

> 
> [...]
>> [3] https://review.openstack.org/498474
> 
> approved
> 
> [...]
> 
> Thanks,
> 




More information about the OpenStack-dev mailing list