15:00:22 #startmeeting third-party 15:00:24 Meeting started Mon Aug 1 15:00:22 2016 UTC and is due to finish in 60 minutes. The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:25 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:27 The meeting name has been set to 'third_party' 15:00:39 o/ 15:00:44 hello asselin 15:00:45 hi 15:00:48 how are you today? 15:00:53 hey lennyb 15:00:55 hi anteaya lennyb 15:02:05 does anyone have anything they would like to discuss today? 15:02:21 some tempest review for multi node scenario tests#link https://review.openstack.org/#/c/343294/ #link https://review.openstack.org/#/c/335447/ 15:02:32 hi 15:02:47 hey rhedlind 15:03:41 lennyb: have you asked anyone in the #openstack-qa channel for reviews on those patches? 15:03:54 anteaya: will do 15:04:42 thanks 15:04:52 since tempest is their repo 15:05:10 I think having them at least okay the direction at this stage is a good idea 15:05:37 any more on lennyb's tempest patches? 15:05:47 no 15:05:56 okay great, thanks lennyb 15:06:07 does anyone have anything else they would like to talk about today? 15:06:37 if anyone has time to help me with a CI issue, I could use some help 15:06:51 rhedlind: what issue? 15:07:29 Zuul and Jenkins are not syncing up. Zuul sees the trigger events coming in but it never kicks of any jobs in jenkins 15:07:46 what do you see in debug log? 15:08:10 did it work with sandbox-ci? 15:08:56 trying to find an example of a cinder event 15:08:57 do you have existing /var/run/zuul/zuul.pid ? 15:09:10 it used to be up and running openstack/cinder testing 15:11:18 one good thing to do is to check state of gearman worker queue, i.e. 'echo status | nc 4730'. It should show all the functions jenkins has registered with gearman 15:11:31 in my CI in such cases the issue was usually bad layout.yaml syntax . Did you validated layout.yaml ? 15:12:07 in your project-config you should have a tox.ini. tox -e zuul should be the command to validate 15:12:09 rhedlind: there are a lot of online yaml validators 15:12:09 oh, I forgot about the validation feature. I will try that right now 15:12:48 rhedlind: also check if zuul.pid exists 15:13:15 #link layout.yaml and jenkins job build configuration validation: http://git.openstack.org/cgit/openstack-infra/project-config-example/tree/tox.ini#n39 15:13:50 ok, I will take a look at this ideas. thanks! 15:15:32 any more on this topic? 15:15:40 * lennyb 1sec 15:15:59 no, I don't want to take up any more of your time on it. Thanks for the ideas! 15:16:18 rhedlind: also check if zuul is not locked in loop. I've noticed that from time to time. a lot of exceptions in zuul.log 15:16:27 * lennyb done 15:16:49 okay thanks rhedlind for the topic 15:17:17 does anyone have anything else they would like to discuss? 15:17:23 yes 15:17:56 go ahead lennyb 15:17:58 from time to time I've noticed that in case that gearman cant see a job to run or no valid slaves are avaliable 15:18:18 it comments on gerrit with NOT_REGISTERED job 15:18:39 yes 15:18:43 this issue is very hard to monitor for me. All I found as a solution is grep the log 15:18:50 ah 15:18:54 any ideas how to prevent it? 15:19:11 +1 on preventing it 15:19:17 or monitor it? 15:19:32 well infra used to get it when the jenkins output specified one node and the gearman was looking for another kind of node 15:19:56 we changed how nodes are specified and that has removed a lot of NOT_REGISTERED for us 15:20:10 and as for monitoring we have humans tell us when there is a problem 15:20:12 however 15:20:26 I can see you would not be able to rely on the same mechanism 15:20:37 it is still a problem when you restart zuul, and jenkins did not have time to register all jobs yet 15:20:40 anteaya: the last time human said it to me, our CI was disabled due to this issue :) 15:20:49 imo, zuul should not throw a fit every time it does not see a job :) 15:20:50 so as of right now, I personally don't know of any way to monitor or prevent it 15:21:12 lennyb: yes, right, then you need a better way other than humans 15:21:25 maybe we can disable commenting in such cases 15:21:34 okay well perhaps you can draft an email to the infra list lennyb 15:21:40 lennyb: I was thinking patching zuul to prevent it 15:21:47 mmedvede +1 15:21:54 and we can discuss it and inlude the rest of the infra folks and other third party operators 15:22:16 mmedvede: let's raise some awareness first so your patch will get some attention 15:22:17 anteaya: ok, I will drop an email 15:22:32 since folks will be aware of your situation and hence the motivation for the patch 15:22:35 I was thinking local patch, not going into community 15:22:38 does that sound reasonable? 15:22:47 mmedvede: okay well that is your choice 15:22:50 but I'll see if it is possible 15:22:56 but upstream might be willing to consider it 15:22:58 thanks 15:23:07 mmedvede: we are community if it's good for you it will be good for me as well :) 15:23:16 :) 15:23:19 right :) 15:23:39 do we have any more on this matter? 15:23:49 lennyb: if I find a way, I'll push a patch. It was more of an annoyance for me with "NOT_REGISTERED" 15:24:03 mmedvede ok 15:24:44 * lennyb finished with this issue 15:24:55 thank you for bringing it up for discussion 15:25:11 does anyone have any other item they would like to discuss today? 15:26:27 does anyone have any objection to me closing the meeting? 15:27:13 thanks everyone for you kind attendance and participation today 15:27:20 enjoy the rest of your day/evening 15:27:27 see you all next week 15:27:29 thank you 15:27:31 #endmeeting