<div dir="auto">The performance is not great because of "rdo-ci-slave01" from which Ansible runs on.<div dir="auto"><br></div><div dir="auto">We all know that node has performance problems (especially i/o).</div><div dir="auto">For example, a promote job [1] will take 1 hour and 4 minutes while the equivalent generic job [2] (ran on a cloudslave) will finish in about 35 minutes.</div><div dir="auto"><br></div><div dir="auto">I mean, it takes rdo-ci-slave01 more than five (5!) minutes to just bootstrap the job (clone weirdo, virtualenv with ara, ansible, shade and initialize ara).</div><div dir="auto">The same thing takes less than 30 seconds on a cloudslave.</div><div dir="auto"><br></div><div dir="auto">[1]: <a href="https://ci.centos.org/job/weirdo-master-promote-packstack-scenario001/1080/">https://ci.centos.org/job/weirdo-master-promote-packstack-scenario001/1080/</a></div><div dir="auto">[2]: <a href="https://ci.centos.org/view/rdo/view/weirdo/job/weirdo-generic-packstack-scenario001/515/">https://ci.centos.org/view/rdo/view/weirdo/job/weirdo-generic-packstack-scenario001/515/</a></div><div dir="auto"><br><div data-smartmail="gmail_signature" dir="auto">David Moreau Simard<br>Senior Software Engineer | Openstack RDO<br><br>dmsimard = [irc, github, twitter]</div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Apr 21, 2017 4:22 AM, "Alfredo Moralejo Alonso" <<a href="mailto:amoralej@redhat.com">amoralej@redhat.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Fri, Apr 21, 2017 at 2:40 AM, David Moreau Simard <<a href="mailto:dms@redhat.com">dms@redhat.com</a>> wrote:<br>
> WeIRDO jobs were tested manually on the rdo-ci-slave01 (promote slave)<br>
> on which the jobs would not run successfully yesterday.<br>
><br>
> Everything now looks good after untangling the update issue from<br>
> yesterday and WeIRDO promote jobs have been switched to rdo-cloud.<br>
><br>
<br>
Nice!, I've seen weirdo jobs in<br>
<a href="https://ci.centos.org/view/rdo/view/promotion-pipeline/job/rdo_trunk-promote-master-current-tripleo/44/" rel="noreferrer" target="_blank">https://ci.centos.org/view/<wbr>rdo/view/promotion-pipeline/<wbr>job/rdo_trunk-promote-master-<wbr>current-tripleo/44/</a><br>
ran in RDO Cloud with pretty good performance, they seems to run<br>
slower than jobs running in dusty servers in ci.centos but faster that<br>
the rest of servers.<br>
<br>
I'll keep an eye on it too to find out if there is any abnormal behavior.<br>
<br>
<br>
> I'll be monitoring this closely but let me know if you see any problems.<br>
><br>
> David Moreau Simard<br>
> Senior Software Engineer | Openstack RDO<br>
><br>
> dmsimard = [irc, github, twitter]<br>
><br>
><br>
> On Thu, Apr 20, 2017 at 12:26 AM, David Moreau Simard <<a href="mailto:dms@redhat.com">dms@redhat.com</a>> wrote:<br>
>> Hi,<br>
>><br>
>> There's been a few updates worth mentioning and explaining to a wider<br>
>> audience as far as RDO is concerned on the <a href="http://ci.centos.org" rel="noreferrer" target="_blank">ci.centos.org</a> environment.<br>
>><br>
>> First, please note that all packages on the five RDO slaves have been<br>
>> updated to the latest version.<br>
>> We had not yet updated to 7.3.<br>
>><br>
>> The rdo-ci-slave01 node (the "promotion" slave) ran into some issues<br>
>> that took some time to fix, EPEL was enabled and it picked up python<br>
>> packages it shouldn't have.<br>
>> Things seem to be back in order now but some jobs might have failed in<br>
>> a weird way, triggering them again should be fine.<br>
>><br>
>> Otherwise, all generic WeIRDO jobs are now running on OpenStack<br>
>> virtual machines provided by the RDO Cloud.<br>
>> This is provided by using the "rdo-virtualized" slave tags.<br>
>> The "rdo-promote-virtualized" tag will be used for the weirdo promote<br>
>> jobs once we're sure there's no more issues running them on the<br>
>> promotion slave.<br>
>><br>
>> These tags are designed to work with WeIRDO jobs only for the time<br>
>> being, please contact me if you'd like to run virtualized workloads<br>
>> from <a href="http://ci.centos.org" rel="noreferrer" target="_blank">ci.centos.org</a>.<br>
>><br>
>> This amounts to around 35 less jobs per day running on Duffy<br>
>> <a href="http://ci.centos.org" rel="noreferrer" target="_blank">ci.centos.org</a> hardware in total on a typical day (including generic<br>
>> weirdo jobs and promote weirdo jobs).<br>
>><br>
>> I've re-shuffled the capacity around a bit, considering we've now<br>
>> freed significant capacity for bare-metal based TripleO jobs.<br>
>> The slave threads are now as follows:<br>
>> - rdo-ci-slave01: 12 threads (up from 11), tagged with "rdo-promote"<br>
>> and "rdo-promote-virtualized"<br>
>> - rdo-ci-cloudslave01: 6 threads (up from 4), tagged with "rdo"<br>
>> - rdo-ci-cloudslave02: 6 threads (up from 4), tagged with "rdo"<br>
>> - rdo-ci-cloudslave03: 8 threads (up from 4), tagged with "rdo-virtualized"<br>
>> - rdo-ci-cloudslave04: 8 threads (down from 15), tagged with "rdo-virtualized"<br>
>><br>
>> There is a specific reason why cloudslave03 and cloudslave04 amount to<br>
>> 16 threads between the two, it is to match the quota we have been<br>
>> given in terms of capacity at RDO cloud.<br>
>> The threads will be used to artificially limit the amount of jobs run<br>
>> against the cloud concurrently without needing to implement queueing<br>
>> on our end.<br>
>><br>
>> You'll otherwise notice the net effect for the "rdo" and "rdo-promote"<br>
>> tag isn't much, at least for the time being, it's very much the same<br>
>> since I've re-allocated cloudslave03 to load balance virtualized jobs.<br>
>> However, jobs are likely to be more reliable and faster now that they<br>
>> won't have to retry for nodes because we're less likely to hit<br>
>> rate-limiting.<br>
>><br>
>> I'll monitor the situation over the next few days and bump the numbers<br>
>> if everything is looking good.<br>
>> That said, I'd like to hear about your feedback if you feel things are<br>
>> looking better and if we are running into "out of inventory" errors<br>
>> less often.<br>
>><br>
>> Let me know if you have any questions,<br>
>><br>
>> David Moreau Simard<br>
>> Senior Software Engineer | Openstack RDO<br>
>><br>
>> dmsimard = [irc, github, twitter]<br>
</blockquote></div></div>