[openstack-dev] [tripleo] Ocata to Pike upgrade job is working as of today.
sathlang at redhat.com
Wed Jan 17 15:52:46 UTC 2018
Emilien Macchi <emilien at redhat.com> writes:
> Nice work Sofer!
> I have a few questions, see inline.
> On Wed, Jan 17, 2018 at 4:46 AM, Sofer Athlan-Guyot <sathlang at redhat.com> wrote:
>> So join us (upgrade squad) to celebrate the working ocata->pike upgrade
>> job, without any depends-on whatsoever.
> w00t, really good work!
>> We would like it to be voting as soon as possible. It has been a
>> rather consuming task to revive that forgotten but important jobs, and
>> the only way for it to not drift into oblivion again is to have it
> The last time I asked how we could make RDO jobs voting, it was
> something in Software Factory to enable, but I'm not sure about the
> I'm ok to make them voting, as long as:
> - they don't timeout or reach the timeout limit (which isn't the case
> now, 2h27 is really good)
> - they proved to be stable during some time
> - they're part of the promotion pipeline so gate can't break easily
> BTW the process is documented here:
Oki reading that, thanks for the pointer.
>> Eventually, let’s thanks rdo-cloud people for their support (especially
>> David Manchado), James Slagle for Traas and Alfredo Moralejo for his
>> constant availability to answer our questions.
>>  https://review.openstack.org/#/c/532791/, look for «gate-tripleo-ci-centos-7-containers-multinode-upgrades-pike»
>>  https://github.com/slagle/traas … the repo we use -> https://github.com/sathlan/traas (so many pull requests to make that it would be cool for it to be an openstack project … :))
> Please rebase https://github.com/slagle/traas/pull/9 and address
> James's comments, so we can avoid using a fork.
So yeah, the situation with traas is not so good as we have a 70 patches
difference with the main repo. The pull/9 may be abandoned (jfrancoa
will see if it’s still required) in favor of the what’s in
https://github.com/sathlan/traas . We’re going to work toward pulling
in slage/traas the main patches to avoid a fork.
> Also, any plan to run tempest instead of pingtest?
Not yet. We could cycle back to that at a later point in time, unless
it’s a requirement. Currently I’ve added it to our backlog.
A last point is that we may add ceph in the mix based on what it’s done
in scenario 001 and 004.
> Emilien Macchi
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
More information about the OpenStack-dev