| openstackgerrit | Merged openstack-infra/zuul master: Move github webhook from webapp to zuul-web https://review.openstack.org/535711 | 00:16 |
|---|---|---|
| openstackgerrit | Merged openstack-infra/zuul master: Move status_url from webapp to web section https://review.openstack.org/536773 | 00:16 |
| openstackgerrit | Merged openstack-infra/zuul master: Remove webapp https://review.openstack.org/536780 | 00:16 |
| openstackgerrit | Merged openstack-infra/zuul master: Add facility for plugins to register web routes https://review.openstack.org/537014 | 00:16 |
| corvus | pow! | 00:17 |
| jlk | webapp buy buy? | 00:17 |
| jlk | uh | 00:17 |
| jlk | bye bye | 00:17 |
| mordred | corvus, jlk: I just added a depends-on on a github PR: https://github.com/ansible/ansible/pull/20974 | 00:22 |
| mordred | corvus, jlk: there's a bug in devel right now and a PR up to fix it - so I added it to the list | 00:22 |
| jlk | I still mentally think of that as a diaper. | 00:22 |
| mordred | that should be a good example of a thing that you otherwise just simply can't do | 00:22 |
| mordred | jlk: :) | 00:23 |
| jlk | so much bot noise | 00:23 |
| jlk | so Zuul is pulling in the dependent change when testing, and passing judgement with that in the speculative state, right? | 00:24 |
| jlk | how is the zuul reporting to github configured? status for "check" and what about gate? Or does it not have any gate config? | 00:25 |
| jlk | huh, the check status link for zuul has https, but firefox throws a security fit when clicking on it. | 00:26 |
| fungi | really? not for me, but i'll do some deep probing of the ssl config | 00:35 |
| fungi | or is it still linking to zuul01.o.o instead of zuul.o.o? | 00:36 |
| fungi | if so, there's a fix (maybe already merged?) for that | 00:36 |
| jlk | yeah https://zuul01.openstack.org/ | 00:37 |
| fungi | okay, maybe we don't have that fix merged, or the comment is from before the fix landed, or... | 00:37 |
| jlk | nod | 00:37 |
| jlk | but it's know, and assumed fixed. | 00:38 |
| fungi | regardless, i've also got a qualys ssl labs analysis of https://zuul.openstack.org/ underway too just to make sure we have the cert there set up thoroughly correctly and with sane configuration | 00:38 |
| fungi | it scores a- due to allowing some semi-weak cipher suites and no rfc 6844 "certification authority authorization (caa)" records in dns and having only partial forward secrecy support | 00:43 |
| fungi | reports are linked from https://www.ssllabs.com/ssltest/analyze.html?d=zuul.openstack.org | 00:45 |
| fungi | actually i don't think it's dinging the score for lack of caa records, just pointing out that we might benefit from them | 00:46 |
| fungi | jlk: fix was https://review.openstack.org/539357 and merged a few days ago, so i'm guessing it would have become effective when the scheduler was last restarted (Jan 29 13:59) | 00:51 |
| fungi | what's the timestamp on the comment you're seeing with the old/wrong link? | 00:51 |
| jlk | https://github.com/ansible/ansible/pull/20974 | 00:52 |
| jlk | it got a pending status 30~ minutes ago | 00:52 |
| fungi | yep, i see that. so i guess the change wasn't entirely sufficient | 00:53 |
| fungi | mordred: ^ any clue what else it's missing? or whether there's more we have to do for that fix to take effect? | 00:53 |
| fungi | oh... that's applied to the executors, not the scheduler | 00:55 |
| fungi | should have have been set in the node entry for the scheduler instead? | 00:56 |
| fungi | or do we just need to do a rolling restart of executors? (last restart for ze01 was Jan 20) | 00:57 |
| pabelanger | mordred: that is pretty awesome, PR20974 | 01:08 |
| jlk | fungi: looking at the github driver, it fetches that info from the [webapp] config section, the status_url as a fallback, unless there is direct config for 'status-url' key in a layout file | 01:11 |
| jlk | hrm. | 01:11 |
| jlk | strike that, [web] not [webapp] | 01:12 |
| jlk | and that should go on the scheduler | 01:12 |
| fungi | yeah, so i guess https://review.openstack.org/539357 was wrong... fix inbound | 01:13 |
| fungi | jlk: mordred: https://review.openstack.org/540563 Set zuul status_url on scheduler not executors | 01:17 |
| jlk | how does that get into the [web] section? | 01:18 |
| fungi | mmm, checking the puppet plumbing for that now | 01:18 |
| fungi | here's where we're passing it from the global site manifest: http://git.openstack.org/cgit/openstack-infra/puppet-zuul/tree/manifests/init.pp#n36 | 01:19 |
| fungi | and here's where it ends up in the [webapp] (not [web]?) section of the config template: http://git.openstack.org/cgit/openstack-infra/puppet-zuul/tree/templates/zuulv3.conf.erb#n33 | 01:20 |
| fungi | do we maybe need it added to [web] there as well? | 01:20 |
| fungi | i vaguely remember a discussion from a few days ago suggesting that's work in progress... seeing if i can find it again | 01:22 |
| jlk | yeah this just changed I think from webapp to web | 01:22 |
| jlk | with something that merged [16:16:48] <openstackgerrit>Merged openstack-infra/zuul master: Move status_url from webapp to web section https://review.openstack.org/536773 | 01:23 |
| jlk | just over an hour ago :D | 01:23 |
| fungi | hah | 01:23 |
| fungi | so there's probably a corresponding puppet change out there already | 01:24 |
| fungi | mmm, nope! patch on the way for that too | 01:25 |
| fungi | jlk: mordred: tobiash: https://review.openstack.org/540564 Also include status_url in [web] | 01:30 |
| fungi | i left it in [webapp] for now as well since i wasn't sure whether there may be a transitional period where we need it in both places | 01:31 |
| jlk | I think at some point webapp goes completely away | 01:31 |
| jlk | where "some point" may be nowish. | 01:31 |
| fungi | yeah, i expect so, i just don't know whether now is that some point, right ;) | 01:32 |
| fungi | the only references i see to "webapp" in the current master branch tip are in tests | 01:34 |
| fungi | so quite possibly now, yes | 01:34 |
| fungi | as long as the server is properly upgrading its from-source deployment | 01:35 |
| fungi | mildly worried about puppet changing zuul.conf to drop the [webapp] section while services are still running from an older point in the branch | 01:37 |
| fungi | but i can't remember whether it only rereads that on restart, in which case it's probably entirely safe | 01:38 |
| *** harlowja has quit IRC | 02:12 | |
| *** harlowja has joined #zuul | 03:23 | |
| *** harlowja has quit IRC | 03:25 | |
| tobiash | fungi, jlk: the webapp is gone as soon as the scheduler is restarted | 05:36 |
| tobiash | corvus, mordred: the proxy config probably needs to be adapted otherwise you won't get gh events after a zuul restart | 05:37 |
| *** kmalloc has joined #zuul | 06:31 | |
| *** Wei_Liu has quit IRC | 07:43 | |
| *** Wei_Liu1 has joined #zuul | 07:43 | |
| *** Wei_Liu1 is now known as Wei_Liu | 07:46 | |
| *** bhavik1 has joined #zuul | 08:18 | |
| *** bhavik1 has quit IRC | 08:25 | |
| *** Wei_Liu has quit IRC | 11:35 | |
| *** Wei_Liu has joined #zuul | 11:38 | |
| *** JasonCL has quit IRC | 14:46 | |
| *** JasonCL has joined #zuul | 15:06 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/nodepool master: Convert nodepool-zuul-functional job https://review.openstack.org/540595 | 15:39 |
| fungi | tobiash: corvus: thanks for confirming! i've also pushed 540597 to clean up the unused [webapp] section from the v3 config template in puppet-zuul | 15:47 |
| corvus | i'll push up the puppet change for the github route | 15:48 |
| fungi | not quite sure what adjustment the proxy config requires (is zuul-web served from a different port?) | 15:48 |
| * SpamapS has reached a point where he's a bit afraid to sync in the latest zuul-jobs | 15:49 | |
| fungi | SpamapS: you could i suppose roll back that update pretty trivially if you know what commit you're on | 15:49 |
| corvus | SpamapS: and i don't think anything should be broken | 15:50 |
| SpamapS | fungi: how'd I know what broke? :-P | 15:50 |
| fungi | hopefully we've maintained backward compatibility | 15:50 |
| SpamapS | I hope too :) | 15:50 |
| corvus | we have a third-party ci checking us ;) | 15:50 |
| fungi | SpamapS: you don't have any way to run some jobs to confirm? | 15:50 |
| fungi | oh, also, maybe you could become a third-party ci system for zuul-jobs now | 15:51 |
| fungi | that would presumably give you greater confidence in its stability | 15:52 |
| fungi | as well as upstream | 15:52 |
| SpamapS | this is where it gets weird with 3rd-parties and zuul configs in-tree | 15:53 |
| SpamapS | well hm | 15:53 |
| SpamapS | what I'd like is the ability to propose a merge from zuul-jobs and have zuul run all my jobs with that merge. | 15:54 |
| corvus | SpamapS: that should be possible | 15:54 |
| SpamapS | thinking how I might be a 3rd-party ci for zuul-jobs | 15:55 |
| corvus | it would be harder to be a third-party ci while not running CD | 15:56 |
| corvus | if you want to continue running on fixed commits with manual updates though, you can just perform the merge locally and push it to your zuul, on which you can configure some jobs to run from a config-project | 15:57 |
| SpamapS | I'm running on a fixed commit right now because of the slack reporter patch. | 15:59 |
| SpamapS | I rebase it every now and then to pick up new zuul features. | 15:59 |
| SpamapS | For zuul-jobs I'm only on a fixed commit because it kept breaking 2 months ago. | 16:00 |
| SpamapS | But now I'd expect it to be a bit more stable. | 16:00 |
| SpamapS | And if I could report on it upstream and prevent it from breaking, that'd be great | 16:00 |
| SpamapS | ideally I wouldn't have a local mirror of zuul-jobs | 16:00 |
| SpamapS | Also, I have a new job for zuul-jobs ... markdownlint. | 16:01 |
| SpamapS | Quite handy for those pesky markdown users. ;) | 16:02 |
| corvus | SpamapS: we'd welcome third-party reporting on zuul-jobs, and that new job :) | 16:06 |
| corvus | SpamapS, mordred: if you could review https://review.openstack.org/540162 it'd be swell; i'd love to slip that in before our next restart | 16:07 |
| *** dmsimard|afk is now known as dmsimard | 16:10 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Add /info and /{tenant}/info route to zuul-web https://review.openstack.org/537011 | 16:10 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Add support for configuring graphite_url https://review.openstack.org/537012 | 16:10 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Move WebInfo config processing into zuul.model https://review.openstack.org/537013 | 16:10 |
| mordred | corvus: looking | 16:11 |
| mordred | corvus: fwiw, I think you don't need to try/except protect 'from unittest import mock' - it was added in 3.3 | 16:13 |
| mordred | corvus: but I just learned that this is a thing, so neat | 16:13 |
| corvus | mordred: oh, ok. i copy pastad that from the gerrit test, which is oldish | 16:19 |
| tobiash | corvus: approved 540162 with comment | 19:59 |
| corvus | tobiash: yes, thanks. | 20:01 |
| *** corvus has quit IRC | 20:01 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/nodepool master: Convert nodepool-zuul-functional job https://review.openstack.org/540595 | 20:09 |
| openstackgerrit | Merged openstack-infra/zuul master: Executor: Don't start too many jobs at once https://review.openstack.org/540162 | 20:15 |
| openstackgerrit | Merged openstack-infra/nodepool master: Provider wedge fix https://review.openstack.org/539316 | 20:24 |
| *** jeblair has joined #zuul | 21:23 | |
| *** jeblair is now known as corvus | 21:23 | |
| *** corvus has quit IRC | 21:56 | |
| *** corvus has joined #zuul | 22:04 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!