[openstack-dev] [TripleO] Upstream backwards compatibility job for Newton oooq

mathieu bultel mbultel at redhat.com
Tue Jan 17 16:33:05 UTC 2017


On 01/17/2017 05:19 PM, Emilien Macchi wrote:
> On Tue, Jan 17, 2017 at 10:57 AM, mathieu bultel <mbultel at redhat.com> wrote:
>> On 01/17/2017 04:42 PM, Emilien Macchi wrote:
>>> On Tue, Jan 17, 2017 at 9:34 AM, mathieu bultel <mbultel at redhat.com> wrote:
>>>> Hi Adriano
>>>>
>>>> On 01/17/2017 03:05 PM, Adriano Petrich wrote:
>>>>
>>>> So I want to make a backwards compatibility job upstream so from last scrum
>>>> I got the feeling that we should not be adding more stuff to the
>>>> experimental jobs due to lack of resources (and large queues)
>>>>
>>>> What kind of "test" do you want to add ?
>>>> I ask because since few days we have upstream an upgrade job that does:
>>>> master UC -> deploying a Newton OC with Newton OC + tht stable/newton ->
>>>> then upgrade the OC to master with tht master branch.
>>>> It's sounds like a "small backward compatibility" validation, but I'm not
>>>> sure if it's cover what you need.
>>> While I understand what is the idea, I don't see the use case.
>>> In which case you want to deploy a old version of overcloud by using a
>>> recent undercloud?
>>> Why don't use deploy a stable undercloud to deploy a stable overcloud?
>> From my side, the use case is the major OC upgrade. We don't want to
>> test the major upgrade of the undercloud (since a job already exist),
>> only overcloud, that's why we start by a "master" undercloud, and that
>> save us from unwanted/unrelated issues due to the UC upgrade and reduce
>> the duration of the job.
> ok so your use-case is CI focused. Good to know.
> Another question for you then, have we count the time needed to
> upgrade an undercloud? I'm doing it quite oftent and it doesn't take
> more than 15 min for me. Thoughts?
Yes it's approximately this duration. Actually it depend on the hardware
and the network connectivity, but saying 15/20 minutes avg sounds
reasonable.
>
>>>> Is that so? I was thinking about using nonha-multinode-oooq that seems to be
>>>> working.
>>>>
>>>> Is that allright to add this new job or should I wait until we get more
>>>> resource and do ci.centos for now, or any idea on where to do this is also
>>>> welcome.
>>>>
>>>>
>>>> Cheers,
>>>>    Adriano
>>>>
>>>>
>>>> __________________________________________________________________________
>>>> OpenStack Development Mailing List (not for usage questions)
>>>> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>>
>>>>
>>>>
>>>> __________________________________________________________________________
>>>> OpenStack Development Mailing List (not for usage questions)
>>>> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>>
>>>
>>
>> __________________________________________________________________________
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>




More information about the OpenStack-dev mailing list