Open Stack

On 6.11.2017 10:52, Marios Andreou wrote:
> On Mon, Nov 6, 2017 at 11:09 AM, Marius Cornea <mariusc at redhat.com> wrote:
> 
>> On Sat, Nov 4, 2017 at 2:27 AM, Emilien Macchi <emilien at redhat.com> wrote:
>>> Since we've got promotion, we can now properly test upgrades from ocata
>> to pike.
>>> It's now failing for various reasons, as you can see on:
>>> https://review.openstack.org/#/c/500625/
>>>
>>> I haven't filled bug yet but this is the kind of thing I see now:
>>> http://logs.openstack.org/25/500625/20/check/legacy-tripleo-ci-centos-7-
>> scenario002-multinode-oooq-container-upgrades/62e7f14/
>> logs/undercloud/home/zuul/overcloud_upgrade_console.log.
>> txt.gz#_2017-11-04_00_14_17
>>
>> I think this is related to https://review.openstack.org/#/c/510577/
>> which introduced running os-net-config during the major upgrade
>> composable step. In case of environments without network isolation
>> /etc/os-net-config/config.json doesn't exist so the os-net-config
>> command fails. I filed https://bugs.launchpad.net/tripleo/+bug/1730328
>> to keep track of it.
>>
>>
> heh, beat me to it :) I was about to file that. Indeed from logs @ [0] you
> can see the step3 ansible-playbook failing for
> https://github.com/openstack/tripleo-heat-templates/blob/e463ca15fb2189fde7e7e2de136cfb2303d3171f/puppet/services/tripleo-packages.yaml#L56-L64
> 
> I had a poke at one of the other jobs too since there are apparently
> multiple issues - I found a different one
> for legacy-tripleo-ci-centos-7-containers-multinode-upgrades and filed
> https://bugs.launchpad.net/tripleo/+bug/1730349 for that. It looks like all
> the upgrade_tasks pass there but then fails on docker-puppet

I'm not sure if it's related to that ^ error in particular, but since we 
landed deploy/upgrade scenario separation [1], the upgrade job on Pike 
effectively started testing non-pacemaker to pacemaker upgrade, which 
won't work. Due to a chicken-and-egg issue with landing related patches 
we could not set the dependencies properly. There's a patch fixing this 
issue and making the Pike upgrade pacemaker-to-pacemaker [2]. This may 
not solve all the issues, but i think we need it merged to at least have 
a chance at a green result.

> 
> [0]
> http://logs.openstack.org/25/500625/20/check/legacy-tripleo-ci-centos-7-scenario002-multinode-oooq-container-upgrades/62e7f14/logs/subnode-2/var/log/messages.txt.gz#_Nov__4_00_13_55
[1] https://review.openstack.org/#/c/500552
[2] https://review.openstack.org/#/c/512305
> 
> thanks,
> 
> marios
> 
> 
>> I'm requesting some help from the upgrades squad, if they already saw
>>> the failures, etc. It would be great to have the jobs passing at some
>>> point, now the framework is in place and we had promotion.
>>>
>>> Thanks,
>>
>> --
>>> Emilien Macchi
>>
> 
> 
> 
> __________________________________________________________________________
> 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
>