| *** nilashishc has joined #zuul | 03:10 | |
| *** nilashishc has quit IRC | 03:15 | |
| *** nilashishc has joined #zuul | 03:17 | |
| *** nilashishc has quit IRC | 03:49 | |
| *** nilashishc has joined #zuul | 03:56 | |
| *** pcaruana has joined #zuul | 04:06 | |
| *** pcaruana has quit IRC | 04:23 | |
| *** nilashishc has quit IRC | 04:32 | |
| *** nilashishc has joined #zuul | 04:57 | |
| *** nilashishc has quit IRC | 05:41 | |
| *** jimi|ansible has joined #zuul | 05:42 | |
| *** pcaruana has joined #zuul | 05:51 | |
| *** nilashishc has joined #zuul | 05:54 | |
| *** jpena|off is now known as jpena | 07:40 | |
| *** zigo has joined #zuul | 07:47 | |
| *** electrofelix has joined #zuul | 08:30 | |
| *** nilashishc has quit IRC | 10:27 | |
| *** nilashishc has joined #zuul | 10:28 | |
| *** jpena is now known as jpena|lunch | 11:30 | |
| *** rfolco has joined #zuul | 12:21 | |
| *** jpena|lunch is now known as jpena | 12:31 | |
| *** nilashishc has quit IRC | 13:18 | |
| *** _ari_ has joined #zuul | 13:44 | |
| *** nilashishc has joined #zuul | 13:48 | |
| *** samccann has joined #zuul | 13:49 | |
| *** nilashishc has quit IRC | 13:52 | |
| *** nilashishc has joined #zuul | 13:58 | |
| pabelanger | SpamapS: tobiash: do you happen to use unlabel property in (check) pipelines? For example, on start unlabel the 'gate' label used to enqueue into gate pipeline. To force a review again by the user. | 14:42 |
|---|---|---|
| *** jimi|ansible has quit IRC | 14:50 | |
| corvus | tristanC: i chatted with mordred in person at ansiblefest and we agree https://review.openstack.org/606740 is a good solution. left some comments. we should also actually document that we publish that tarball. :) | 14:53 |
| tobiash | pabelanger: no, didn't use it yet | 14:59 |
| pabelanger | tobiash: k, you are gating both on code review approval, and some label for github? | 15:00 |
| pabelanger | The issue i am trying to figure out, is if a new patchset is uploaded, to remove the 'gate' label when the PR is back in check, to avoid somebody pushing up a new patch and keeping the 'gate' label | 15:01 |
| pabelanger | unlabel works, however if because we also use /recheck to add something into check, zuul will run unlabel there too. Removing 'gate' label | 15:02 |
| pabelanger | s/if because/if | 15:02 |
| tobiash | pabelanger: yes, we gate on code review and the 'merge | 15:05 |
| tobiash | label | 15:05 |
| pabelanger | tobiash: okay, we've dropped code review for now due to single user on repo issue | 15:05 |
| tobiash | pabelanger: but unlabel on gate start sounds good | 15:06 |
| pabelanger | I want to add it back, but can't get past the need 2 github accounts | 15:06 |
| tobiash | yes, github doesn't support self-review | 15:06 |
| tobiash | when I asked about this the answer was: 'self review doesn't make sense so we don't support it' | 15:07 |
| pabelanger | yah | 15:07 |
| jlk | pabelanger: I think we didn't use labels to control that | 15:07 |
| jlk | but yeah you're running into the self review thing :/ | 15:07 |
| tobiash | jlk: any chance that self-review will be possible/allowable somewhere in the future? | 15:08 |
| jlk | honestly I don't know. | 15:09 |
| jlk | Nothing is going to get changed there right now, but after Universe I might be able to bend some ears on this | 15:09 |
| jlk | although the answer just may be "use labels" | 15:10 |
| jlk | pabelanger: we should probably walk through the label strategy and see where it's inadequate. | 15:12 |
| pabelanger | jlk: agree, would like that also | 15:17 |
| *** pcaruana has quit IRC | 15:30 | |
| fungi | another zuul case study just got posted on superuser: http://superuser.openstack.org/articles/zuul-case-study-packet-host/ | 15:46 |
| dmsimard | sweet! | 15:48 |
| SpamapS | pabelanger: yes I do use it | 15:52 |
| SpamapS | pabelanger: I have a specific pipeline just for unlabeling things that are 'approved' | 15:53 |
| corvus | when nicole sends blurbs for the pending user stories, i'll update the website page to add them | 15:53 |
| SpamapS | http://paste.openstack.org/show/731181/ | 15:54 |
| SpamapS | tobiash: ^^ | 15:54 |
| *** bhavikdbavishi has joined #zuul | 15:55 | |
| SpamapS | I have that templated into all my projects now | 15:55 |
| pabelanger | SpamapS: thank you! I'll test that this morning, also for your check pipeline, you still trigger on pull_request, event opened / changed / reopened + recheck? | 15:56 |
| SpamapS | pabelanger: correct | 16:02 |
| SpamapS | On the self-review thing.. honestly.. the labels workflow is pretty easy to grok and you can use my little hack to tack labels on when somebody approves something via review. | 16:03 |
| SpamapS | One thing I haven't managed to circle back to is triggering things to happen on negative/dismissed reviews. | 16:04 |
| pabelanger | SpamapS: yah, I also was looking into branch protection things, which seem to also be able to setup number of approvals for a project. | 16:05 |
| Shrews | fungi: neat. it'd be nice if the article's nodepool doc link pointed to actual nodepool docs though | 16:16 |
| fungi | Shrews: i'll admit, i only saw someone mention the article, i'm still behind enough this morning i haven't read it yet | 16:17 |
| fungi | oh, lookie there, it's not even morning here now :/ | 16:17 |
| Shrews | when you're behind, it can still be morning :) | 16:17 |
| fungi | i expect it's going to be morning all day. maybe all week | 16:18 |
| SpamapS | pabelanger: note that if you have zuul as a required check status in branch protection then dependent pipelines won't pass the "is this mergable" check until that status is set to OK | 16:20 |
| corvus | Shrews: good catch, i'll ask nicole to update it | 16:21 |
| tobiash | SpamapS: so you'r running a noop job in unlabel-on-push? | 16:22 |
| *** samccann is now known as away | 16:24 | |
| SpamapS | tobiash: yep, on every project | 16:24 |
| *** away is now known as Guest62683 | 16:24 | |
| SpamapS | I actually would love to have a way to basically hide these | 16:24 |
| *** Guest62683 is now known as samccann | 16:25 | |
| tobiash | SpamapS: so your mergers don't have to be afraid of being jobless ;) | 16:25 |
| SpamapS | I have no dedicated mergers | 16:26 |
| SpamapS | 3 executors, they seem to get the job done. | 16:26 |
| corvus | SpamapS: by hide -- do you mean don't include the job result message when reporting to gh? | 16:27 |
| tobiash | SpamapS: I guess you mean to hide the noop job from the pipeline of every job? | 16:29 |
| tobiash | SpamapS: if yes, there is a way: regex project matching in a trusted repo | 16:29 |
| corvus | hrm, 'comment' should do what i said: https://zuul-ci.org/docs/zuul/admin/drivers/github.html#attr-pipeline.%3Creporter%3E.%3Cgithub%20source%3E.comment | 16:29 |
| SpamapS | corvus: don't show in the UI, and don't even require a noop job to be specified in the template. It's not really a pipeline. It's just a little zuul magic. | 16:31 |
| SpamapS | I do have a regex that attaches it. | 16:31 |
| SpamapS | I just find that a bit silly | 16:31 |
| pabelanger | SpamapS: okay, thanks! I was able to get unlabel-on-push pipeline to work. I guess you have a global project-template where you add unlabel-on-push: jobs: -noop to all projects, right? | 16:51 |
| SpamapS | pabelanger: yeah | 16:51 |
| pabelanger | SpamapS: great, thanks again! | 16:52 |
| *** pcaruana has joined #zuul | 16:56 | |
| *** jpena is now known as jpena|off | 17:11 | |
| *** nilashishc has quit IRC | 17:16 | |
| *** electrofelix has quit IRC | 17:16 | |
| *** nilashishc has joined #zuul | 17:55 | |
| *** nilashishc is now known as nilashish | 17:55 | |
| *** nilashish has quit IRC | 18:06 | |
| *** jimi|ansible has joined #zuul | 18:08 | |
| *** nilashishc has joined #zuul | 18:10 | |
| *** bhavikdbavishi has quit IRC | 18:14 | |
| *** gundalow has quit IRC | 18:14 | |
| *** gundalow has joined #zuul | 18:14 | |
| Shrews | mordred: the install-docker role uses an 'ansible_user' variable (assumes this is defined). but when i use that role, i do not have it defined (and doesn't appear to be a fact var). | 18:24 |
| Shrews | mordred: do we need to mention this as a required var in the README.rst? | 18:24 |
| corvus | pabelanger: can you look at http://logs.openstack.org/34/589634/5/check/openstack-zuul-jobs-linters/a6b5435/job-output.txt.gz ? i can't figure out why the skip tag isn't working | 18:25 |
| Shrews | or is it *suppose* to reference a fact var | 18:25 |
| Shrews | mordred: i'm also assuming you wrote that role. maybe i should check... | 18:26 |
| Shrews | yep | 18:26 |
| pabelanger | corvus: sure | 18:31 |
| pabelanger | corvus: looking at project-config, it looks like we are not adding zuul_return into the ANSIBLE_LIBRARY path: http://git.openstack.org/cgit/openstack-infra/project-config/tree/tox.ini#n14 which means, we need to also add zuul as a dependency. http://git.openstack.org/cgit/openstack-infra/project-config/tree/tox.ini#n30 | 18:47 |
| pabelanger | doing that locally made syntax-check work as expected | 18:47 |
| pabelanger | your code for tags is correct | 18:48 |
| corvus | pabelanger: i shouldn't skip_ansible_list have worked anyway? | 18:49 |
| pabelanger | corvus: in this case no, this is ansible-playbook that is raising the error, not ansible-lint | 18:50 |
| corvus | pabelanger: http://git.openstack.org/cgit/openstack-infra/zuul-jobs/tree/roles/upload-logs/tasks/main.yaml#n53 is a zuul_returnn | 18:51 |
| pabelanger | corvus: yah, we must not be running a playbook now that invokes upload-logs in zuul-jobs repo. I think in this case, it is the playbooks/tox/ playbooks that is getting loaded with --syntax-check | 18:52 |
| pabelanger | http://paste.openstack.org/show/731201/ was the playbooks it found for me locally (I've modified tox.ini to strip flake / ansible-lint) | 18:53 |
| pabelanger | this must be the first time we are trying to load zuul_return with it | 18:54 |
| corvus | pabelanger: oh, so it's running the syntax check on all playbooks in the repo, and since there's a job+playbook in repo that uses the role, it's hitting that | 18:55 |
| corvus | pabelanger: okay, then, yeah, i think we want to do as you suggest | 18:56 |
| pabelanger | yes, exactly right | 18:56 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Extract pep8 messages for inline comments https://review.openstack.org/589634 | 19:04 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add zuul to the linters tox env https://review.openstack.org/607026 | 19:04 |
| corvus | pabelanger: think that'll do it? ^ | 19:07 |
| pabelanger | corvus: +2 | 19:07 |
| *** nilashishc has quit IRC | 19:37 | |
| *** nilashishc has joined #zuul | 19:39 | |
| *** nilashishc has quit IRC | 19:48 | |
| *** nilashishc has joined #zuul | 19:49 | |
| *** nilashishc has quit IRC | 19:56 | |
| Shrews | mordred: oh, nm. i see where that comes from now | 19:56 |
| dmsimard | There is a Zuul session at the AnsibleFest contributor summit bound to be starting any minute now -- it's hosted in #ansible-network and there should also be a bluejeans | 20:02 |
| Shrews | dmsimard: hrm, bluejeans link? | 20:06 |
| Shrews | dmsimard: nm, found it | 20:07 |
| *** jimi|ansible has quit IRC | 20:16 | |
| *** pcaruana has quit IRC | 20:43 | |
| *** samccann has quit IRC | 21:05 | |
| clarkb | what is/was subject of conversation? using zuul for ansible or making ansible better for zuul or? | 21:16 |
| clarkb | I just landed in austin | 21:16 |
| Shrews | corvus: tristanC: fwiw, I'm having problems getting minikube running with the --vm-driver=none option. And I'm not sure I understand completely why we can't use virtualization there. May need tristanC or someone more k8s experienced to help me out. | 21:23 |
| corvus | clarkb: bit of both; etherpad https://etherpad.openstack.org/p/ansible-summit-october-2018-zuul there will be irc and bluejeans recordings | 21:24 |
| *** caphrim007 has joined #zuul | 21:35 | |
| *** caphrim007 has quit IRC | 22:00 | |
| corvus | https://meetbot.fedoraproject.org/ansible-network/2018-10-01/ansiblefest_austin:_contributors_summit_-_zuul.2018-10-01-20.05.log.html | 22:01 |
| *** ianw is now known as ianw_pto | 22:17 | |
| *** threestrands has joined #zuul | 22:42 | |
| gouthamr | hi, i'm having an issue getting the "logs" directory copied over as part of the Zuul post-run - http://logs.openstack.org/29/604929/3/check/manila-tempest-dsvm-mysql-generic/9328255/ (https://review.openstack.org/#/c/604929/) | 23:27 |
| gouthamr | I'm changing the nodeset that the jobs use, and per clarkb's suggestion, i added a definition for the nodeset and set the groups to tempest | 23:27 |
| gouthamr | nodeset definition: ^ https://review.openstack.org/#/c/605810/ | 23:28 |
| gouthamr | !! wow :P i named the node "controller" and post.yaml on my jobs is expected to copy "logs" from a non-existent "primary" | 23:44 |
| openstack | gouthamr: Error: "!" is not a valid command. | 23:44 |
| gouthamr | openstack: it should be :) | 23:44 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!