[openstack-dev] [ironic] [third-party-ci] pkvmci ironic job breakage details
Michael Turek
mjturek at linux.vnet.ibm.com
Mon Apr 17 19:33:25 UTC 2017
On 04/17/2017 02:25 PM, Matt Riedemann wrote:
> On 4/14/2017 10:51 AM, Michael Turek wrote:
>> Hey ironic-ers,
>>
>> So our third party CI job for ironic has been, and remains, broken. I
>> was able to do some investigation today and here's a summary of what
>> we're seeing. I'm hoping someone might know the root of the problem.
>>
>> For reference, please see this paste and the logs of the job that I was
>> working in:
>> http://paste.openstack.org/show/606564/
>> https://dal05.objectstorage.softlayer.net/v1/AUTH_3d8e6ecb-f597-448c-8ec2-164e9f710dd6/pkvmci/ironic/25/454625/10/check-ironic/tempest-dsvm-ironic-agent_ipmitool/0520958/
>>
>>
>>
>> I've redacted the credentials in the ironic node-show for obvious
>> reasons but rest assured they are properly set. These commands are run
>> while
>> '/opt/stack/new/ironic/devstack/lib/ironic:wait_for_nova_resources' is
>> looping.
>>
>> Basically, the ironic hypervisor for the node doesn't appear. As well,
>> none of the node's properties make it to the hypervisor stats.
>>
>> Some more strangeness is that the 'count' value from the 'openstack
>> hypervisor stats show'. Though no hypervisors appear, the count is still
>> 1. Since the run was broken, I decided to delete node-0 (about 3-5
>> minutes before the run failed) and see if it updated the count. It did.
>>
>> Does anyone have any clue what might be happening here? Any advice would
>> be appreciated!
>>
>> Thanks,
>> mjturek
>>
>>
>> __________________________________________________________________________
>>
>> 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
>
> See:
>
> http://lists.openstack.org/pipermail/openstack-dev/2017-April/115486.html
>
Thanks Matt,
Unfortunately doesn't seem to be the fix.
I did a quick test run of the job and ran "nova-manage cell_v2
discover_hosts --verbose" manually while ironic:wait_for_nova_resources
was looping (where we eventually fail). This fixes the issue of the
hypervisor not appearing, but the resources associated with the
hypervisor (vcpus, memory_mb, etc) remain 0.
mjturek
More information about the OpenStack-dev
mailing list