| *** harlowja has joined #zuul | 00:49 | |
| *** harlowja has quit IRC | 01:42 | |
| *** harlowja has joined #zuul | 01:56 | |
| *** harlowja has quit IRC | 02:20 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: launcher: store console log in workspace https://review.openstack.org/464521 | 04:27 |
|---|---|---|
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: launcher: add simple email publisher https://review.openstack.org/464923 | 04:32 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: launcher: add results_dir option https://review.openstack.org/464524 | 04:38 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: launcher: terminate console server after job ends https://review.openstack.org/464520 | 04:40 |
| *** jamielennox is now known as jamielennox|away | 05:10 | |
| *** jamielennox|away is now known as jamielennox | 05:30 | |
| *** harlowja has joined #zuul | 05:47 | |
| *** jlk has quit IRC | 06:09 | |
| *** jlk has joined #zuul | 06:09 | |
| *** jlk has quit IRC | 06:09 | |
| *** jlk has joined #zuul | 06:09 | |
| *** isaacb has joined #zuul | 06:40 | |
| *** isaacb has quit IRC | 06:41 | |
| *** isaacb has joined #zuul | 06:41 | |
| *** harlowja has quit IRC | 06:50 | |
| *** DangerousDaren has joined #zuul | 06:54 | |
| *** isaacb_ has joined #zuul | 07:00 | |
| *** isaacb has quit IRC | 07:00 | |
| *** jamielennox is now known as jamielennox|away | 07:07 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool feature/zuulv3: Add max-ready-age to label config https://review.openstack.org/463338 | 07:24 |
| *** jamielennox|away is now known as jamielennox | 07:34 | |
| *** DangerousDaren has quit IRC | 08:19 | |
| *** DangerousDaren has joined #zuul | 08:19 | |
| *** isaacb_ has quit IRC | 08:59 | |
| *** isaacb has joined #zuul | 09:06 | |
| *** isaacb has quit IRC | 09:09 | |
| *** Cibo_ has joined #zuul | 09:25 | |
| *** Cibo_ has quit IRC | 09:57 | |
| *** jkilpatr has quit IRC | 10:41 | |
| *** jkilpatr has joined #zuul | 10:59 | |
| *** Cibo_ has joined #zuul | 11:06 | |
| *** Cibo_ has quit IRC | 11:10 | |
| *** yolanda has quit IRC | 11:27 | |
| *** jhesketh has quit IRC | 11:39 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Don't report start on unmanaged projects https://review.openstack.org/455711 | 11:41 |
| *** jhesketh has joined #zuul | 11:41 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Whitelist pydevd debug threads https://review.openstack.org/465046 | 11:41 |
| *** isaacb has joined #zuul | 11:43 | |
| *** yolanda has joined #zuul | 11:59 | |
| *** morgan_ is now known as morgan | 12:49 | |
| dmsimard | pabelanger: did you get the chance to chat about nodepool boot from volume at summit ? | 13:15 |
| *** dkranz_ has joined #zuul | 13:20 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Fix inventory vars containing spaces https://review.openstack.org/465072 | 13:55 |
| tobiash | hi | 14:08 |
| tobiash | Is something broken with the docs jobs? I'm not aware of changing something docs related in ^^ | 14:08 |
| pabelanger | dmsimard: yes: https://review.openstack.org/464283 | 14:09 |
| pabelanger | dmsimard: just need reviews / maybe gate job | 14:09 |
| pabelanger | dmsimard: and new shade release | 14:09 |
| dmsimard | pabelanger: nice, I'll have a look, thanks ! | 14:10 |
| pabelanger | tobiash: have a log? | 14:11 |
| dmsimard | pabelanger: lgtm at first glance | 14:11 |
| pabelanger | tobiash: Hmm, looks like PBR. Wonder if there is a new release of something | 14:12 |
| tobiash | pabelanger: http://paste.openstack.org/show/609669/ | 14:12 |
| tobiash | at least what I could find in the logs | 14:12 |
| tobiash | http://logs.openstack.org/72/465072/1/check/gate-zuul-docs-ubuntu-xenial/65c83d3/console.html | 14:12 |
| tobiash | pabelanger: will check tomorrow again, it's eod for me | 14:15 |
| pabelanger | tobiash: thanks, I'll look into it now | 14:20 |
| pabelanger | ya, sphinx 1.6.1 today | 14:22 |
| mordred | oh goodies | 15:37 |
| *** isaacb has quit IRC | 15:57 | |
| openstackgerrit | Leif Madsen proposed openstack-infra/zuul master: Cleanup quickstart documentation https://review.openstack.org/465108 | 16:12 |
| *** leifmadsen has joined #zuul | 16:12 | |
| *** dmsimard has quit IRC | 16:14 | |
| *** cmurphy has quit IRC | 16:14 | |
| *** dmsimard has joined #zuul | 16:17 | |
| *** isaacb has joined #zuul | 16:19 | |
| mordred | leifmadsen: ^^ thanks - comments on the review | 16:19 |
| leifmadsen | rock on | 16:19 |
| leifmadsen | looking | 16:19 |
| leifmadsen | ahhh, didn't look at remote branches | 16:22 |
| leifmadsen | so I guess master is v2.5? | 16:22 |
| SpamapS | yep | 16:22 |
| leifmadsen | not sure how the branching is currently setup | 16:22 |
| leifmadsen | ok thanks | 16:22 |
| SpamapS | feature/zuulv3 is where most dev is happening | 16:22 |
| jlk | o/ | 16:22 |
| leifmadsen | SpamapS: cool thanks | 16:22 |
| leifmadsen | playing / learning zuul today | 16:23 |
| leifmadsen | should only take an hour or two to become and expert right? | 16:23 |
| jlk | probably enough. | 16:24 |
| mordred | leifmadsen: you'll be in perfect shape in about 10 minutes ;) | 16:36 |
| leifmadsen | what will I do with the other 7 mins? | 16:36 |
| leifmadsen | mordred: so to your comments, I was wondering about the bindep.txt thing, but then realized the quickstart talks about using pip, which means there is no bindep.txt locally (that would imply you've cloned the repo) | 16:38 |
| leifmadsen | I could perhaps add a command to curl the bindep.txt file from the github repo I suppose | 16:38 |
| Shrews | Zuulers: FYI, some sort of summer ick has me totally out of commission today. Going back to bed. Hopefully more useful tomorrow. | 16:48 |
| leifmadsen | also, I get the impression that bindep doesn't work on CentOS? | 16:56 |
| leifmadsen | getting issues with running it during the profiles list part | 16:56 |
| pabelanger | leifmadsen: centos works, we use it in the gate today. What is the error you are getting? | 17:05 |
| leifmadsen | https://paste.fedoraproject.org/paste/OkZylTQkX0oM4A2o6jVLfF5M1UNdIGYhyRLivL9gydE= | 17:05 |
| jlk | wonder if it's looking for something in redhat-lsb | 17:10 |
| jlk | I really hate that ERRno 2 never shows the file it was looking for | 17:10 |
| jlk | like, the MOST IMPORTANT PART OF THE ERROR... | 17:10 |
| leifmadsen | jlk: welp that's exactly it :) | 17:10 |
| *** unrahul has joined #zuul | 17:11 | |
| leifmadsen | requires a `yum install redhat-lsb-core` before `bindep` works | 17:11 |
| jlk | Cool, you've just found a doc contribution you can make :) | 17:11 |
| *** isaacb has quit IRC | 17:11 | |
| leifmadsen | the worst^Wbest part of working on a new project is finding all the edge cases | 17:13 |
| clarkb | pretty sure that is dpcumebted somewhere, maybe just needs to be more prominent | 17:13 |
| leifmadsen | well I'm fixing up the quickstart guide since it's missing stuff from quickstarting :) | 17:14 |
| jlk | Most of the v3 docs are going to be old/wrong | 17:14 |
| leifmadsen | I'm on 2.5 | 17:14 |
| jlk | all the v3 stuff was being done without updating the docs. | 17:14 |
| jlk | ah okay | 17:14 |
| leifmadsen | well, that's where I started | 17:14 |
| leifmadsen | jlk: just like all great open source projects | 17:14 |
| jlk | If you're at all interested, we (BonnyCI) have a pile of Ansible for standing up a 2.5 (and a 3) environment. It uses ubuntu tho. | 17:15 |
| leifmadsen | moving on :) | 17:17 |
| jlk | Wasn't my choice, but I didn't want to disrupt inertia. | 17:18 |
| leifmadsen | all good, pabelanger has some playbooks as well | 17:18 |
| leifmadsen | I'm mostly just doing documentation review and writing | 17:18 |
| leifmadsen | not necessarily interested in spinning up infra yet | 17:18 |
| leifmadsen | huh... sphinx-build --version returns 1.6.1 but still getting error: error in setup.cfg: command 'build_sphinx' has no such option 'warning_is_error' | 17:19 |
| * mordred missed all sorts of good leifmadsen scrollback! | 17:22 | |
| leifmadsen | "good" | 17:22 |
| mordred | leifmadsen: so - there's a bug that just hit today with openstack/pbr and sphinx 1.6.1 | 17:22 |
| leifmadsen | lol | 17:22 |
| leifmadsen | of course | 17:22 |
| mordred | we're pinning back to 1.5 elsewhere in openstack while we figure it out | 17:22 |
| mordred | right? | 17:22 |
| leifmadsen | that's pretty much how it works with me :) | 17:22 |
| mordred | yay for sphinx release today! | 17:22 |
| leifmadsen | hrm | 17:23 |
| leifmadsen | pinned to 1.5.1 locally, still same error | 17:23 |
| leifmadsen | I'm no longer convinced that the version in the venv is what is getting used | 17:24 |
| jlk | Is there an updated division of duties doc on what scheduler vs merger vs executor does in v3, and how they all talk to each other? | 17:28 |
| * jlk is curious why on start up it seems both merger and executor are cloning down projects and scanning branches. | 17:31 | |
| *** unrahul has quit IRC | 17:38 | |
| *** Cibo_ has joined #zuul | 17:39 | |
| clarkb | to load config I think | 17:40 |
| jlk | They seem to be doing the exact same actions | 17:42 |
| *** mattclay has quit IRC | 17:42 | |
| *** mattclay has joined #zuul | 17:45 | |
| jlk | oh cool, caused a traceback by hitting the base /status url instead of a /<tenant>/status | 17:48 |
| *** harlowja has joined #zuul | 17:50 | |
| SpamapS | jlk: I believe there is a possible v3 future where merger doesn't do that, but I seem to recall we found a corner case where it was still needed. | 17:54 |
| SpamapS | or where merger isn't even a thing anymore. | 17:54 |
| * SpamapS has been too deep in implementation to think theory lately | 17:55 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Add SSH Agent Primitives and usage https://review.openstack.org/462712 | 17:57 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Add support for bwrap https://review.openstack.org/453851 | 17:57 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Add bubblewrap to bindep / test-setup.sh https://review.openstack.org/461849 | 17:57 |
| SpamapS | pabelanger: ^ FYI, I rebased the bindep change | 17:57 |
| SpamapS | pabelanger: also, base job.. it's removed from zuul, so is openstack-zuul-jobs the only place I can find it? | 17:58 |
| leifmadsen | since I'm new to infra, where are the definitions for the jobs that test zuul itself? | 18:07 |
| dmsimard | Would it be a reasonable RFE for Zuul to provide the reason why a job is queued ? i.e, 'waiting for node: <nodepool-label>' or something like that. We just troubleshooted a broken job in #openstack-infra and had to crawl through "compiled" jjb to figure out the node allocation was wrong. | 18:15 |
| dmsimard | In the patch json (i.e, http://zuul.openstack.org/status/change/465040,1 ) there could be an additional field | 18:16 |
| *** jkilpatr_ has joined #zuul | 18:17 | |
| *** openstackgerrit has quit IRC | 18:17 | |
| *** jkilpatr has quit IRC | 18:20 | |
| jlk | hrm, interesting. zuul-executor reads from the [merger] config block | 18:56 |
| jlk | So in v3, executor pushes the merged repo content into the node, yes? there does not need to be anything connecting _to_ the executor? It'll get jobs from gearman, reach out to nodepool nodes, etc..? | 18:59 |
| jlk | oh what??? | 19:02 |
| jlk | Both the executor/server.py and merger/server.py register the "merger:merge" and "merger:cat" functions. | 19:02 |
| jlk | jeblair: what's the deal with ^^ ? | 19:02 |
| *** DangerousDaren has quit IRC | 19:05 | |
| *** harlowja has quit IRC | 19:17 | |
| SpamapS | jlk: correct, executor merges and pushes via ssh/ansible | 19:29 |
| SpamapS | jlk: I don't recall why we still need merger | 19:30 |
| jlk | it has a task that executor doesn't do, update | 19:30 |
| SpamapS | Right that's for zuul refs. If those are needed we might still need that. I think. | 19:32 |
| jlk | it's curious that the executor is registering the same functions as merger | 19:33 |
| SpamapS | Agreed | 19:33 |
| *** jkilpatr_ has quit IRC | 19:40 | |
| *** Cibo_ has quit IRC | 19:42 | |
| jlk | Is there a type of job I can have zuul do that won't involve nodepool? I tried a noop job, but that seems to be waiting on nodepool for some reason. | 19:45 |
| *** pabelanger has quit IRC | 19:51 | |
| mordred | leifmadsen: so - there are two answers to your question ... because we currently have a zuul 2.5 running tests on zuul AND a zuul v3 running tests on zuul | 19:54 |
| mordred | leifmadsen: the v3 config starts here: http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/main.yaml | 19:55 |
| mordred | leifmadsen: which will get you to here: http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul.yaml | 19:55 |
| *** pabelanger has joined #zuul | 19:55 | |
| leifmadsen | mordred: cool thanks, I was just running into, what I figured were some pretty basic pip install issues, so wanted to see how that was being tested to see if 1) I was being dumb 2) I was testing a base scenario that wasn't actually built into a job yet | 19:55 |
| mordred | leifmadsen: and then this: http://git.openstack.org/cgit/openstack-infra/zuul/tree/.zuul.yaml?h=feature/zuulv3 | 19:56 |
| leifmadsen | +1 awesome will look through those | 19:56 |
| mordred | leifmadsen: ah - ok - so in that case you might want to look at: | 19:56 |
| mordred | leifmadsen: http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs and http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/layout.yaml - which are the v2 config- although that might be a bit hard to sift through | 19:57 |
| mordred | leifmadsen: http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/ and http://git.openstack.org/cgit/openstack-infra/openstack-zuul-roles/tree might also be interesting | 19:58 |
| leifmadsen | SpamapS: ping? | 20:02 |
| *** dougbtv has joined #zuul | 20:03 | |
| jlk | He's at the Dr office I think. | 20:04 |
| *** harlowja has joined #zuul | 20:04 | |
| SpamapS | Oh right it's time for the kubernetes testing sig meeting | 20:06 |
| SpamapS | leifmadsen: hey can you let them know I'm out sick at the moment? Thanks | 20:06 |
| leifmadsen | SpamapS: sure thing | 20:06 |
| SpamapS | Don't have that slack on my phone | 20:06 |
| leifmadsen | np | 20:06 |
| leifmadsen | done | 20:08 |
| leifmadsen | hopefully you're feeling better and we can hear about this next week! | 20:08 |
| SpamapS | Indeed | 20:15 |
| jlk | Does zuulv3 require the zuulv3 branch of nodepool, or master? | 20:15 |
| SpamapS | zuulv3 I thought | 20:15 |
| SpamapS | Until the shim is done | 20:16 |
| SpamapS | mordred: right? | 20:16 |
| jlk | doesn't look like we differentiate in hoist | 20:16 |
| SpamapS | Maybe it merged.. I never followed it closely | 20:16 |
| jlk | oh wait | 20:17 |
| jlk | I'm wrong | 20:17 |
| *** jkilpatr has joined #zuul | 20:53 | |
| *** eggshell has left #zuul | 21:03 | |
| *** openstackgerrit has joined #zuul | 21:16 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Block sphinx 1.6.1 https://review.openstack.org/465176 | 21:16 |
| mordred | jlk, SpamapS, clarkb, pabelanger: ^^ that should generall workaround the sphinx issue leifmadsen ran in to earlier today | 21:16 |
| *** jkilpatr has quit IRC | 21:19 | |
| jlk | reviewed | 21:20 |
| mordred | jlk: also - I'm working through the GH stack - a bunch of it would be merged now - except for that sphinx silliness :) | 21:21 |
| mordred | jlk: did you ever make progress with our friends at gh on the whole review stickiness issue? | 21:22 |
| jlk | YES | 21:22 |
| jlk | they implemented what we wanted mostly. | 21:22 |
| mordred | oh that's awesome news!!! | 21:22 |
| * mordred hands jlk a gold cookie | 21:22 | |
| mordred | wait. that's a terrible combinatoin. don't eat that | 21:23 |
| jlk | There is now an option under branch protection where you can have positive reviews automatically dismissed when a PR code is updated. | 21:23 |
| jlk | but first you have to turn branch protection on, _and_ require passing reviews to merge | 21:23 |
| jlk | "Dismiss stale pull request approvals when new commits are pushed | 21:25 |
| jlk | New reviewable commits pushed to a branch will dismiss pull request review approvals." | 21:25 |
| mordred | I think that's a huge improvement | 21:28 |
| pabelanger | +2 | 21:30 |
| mordred | jlk: also - so far the gh patches have been very understandable -thank you for your work on them! | 21:32 |
| jlk | yaaaay. At this point I have no idea how far they've drifted from original author's versions. | 21:33 |
| mordred | well- there's definitely jlk work in there I know - because some of it is based on discussions we've had in here | 21:34 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool feature/zuulv3: Block sphinx 1.6 https://review.openstack.org/465179 | 21:36 |
| mordred | pabelanger: next verse, same as the first ^^ | 21:37 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Block sphinx 1.6.1 https://review.openstack.org/465176 | 21:38 |
| pabelanger | mordred: +2 also | 21:47 |
| SpamapS | something very wrong with my ssh agent stuff | 21:47 |
| SpamapS | test suite locks up :-P | 21:47 |
| mordred | SpamapS: that made me sing "ssh aaaaagent man" in my head | 21:49 |
| jlk | and now we're all singing that | 21:50 |
| mordred | you're welcome | 21:51 |
| * SpamapS salutes the song with a fit of hacking cough | 21:52 | |
| jlk | AttributeError: 'list' object has no attribute 'hold_following_changes' | 21:57 |
| jlk | that's... weird. | 21:57 |
| jlk | out of manager/__init__.py | 21:57 |
| SpamapS | jlk: likely a yaml indentation fail | 21:58 |
| SpamapS | Sometimes I think yaml might have been a mistake. | 21:59 |
| SpamapS | The structure seems too loose to validate fully. | 21:59 |
| jlk | I don't see anything odd with http://paste.openstack.org/show/609724/ | 22:00 |
| mordred | jlk: in https://review.openstack.org/#/c/449365 I have left a -1 - but I also could be reading it wrong, so thought I'd ping you here | 22:00 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Support GitHub PR webhooks https://review.openstack.org/439834 | 22:00 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add basic Github Zuul Reporter. https://review.openstack.org/443323 | 22:00 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add 'push' and 'tag' github webhook events. https://review.openstack.org/443947 | 22:00 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add 'comment' github pull_request event action https://review.openstack.org/443959 | 22:00 |
| SpamapS | niiiiiiiiiiiiiiice | 22:01 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Configurable SSH access to GitHub https://review.openstack.org/444034 | 22:01 |
| mordred | SpamapS: it's happening | 22:01 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Support for github commit status https://review.openstack.org/444060 | 22:01 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Merge pull requests from github reporter https://review.openstack.org/444463 | 22:01 |
| mordred | jlk: also - no rush, because that's pretty much at the beginning of the part of the stack without a jeblair review | 22:02 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: support github pull request labels https://review.openstack.org/444511 | 22:02 |
| SpamapS | I can't believe I'm saying this.. but I wish ssh-agent implemented the systemd startup readiness thing. | 22:02 |
| SpamapS | letting it daemonize is making me twitch | 22:02 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Encapsulate determining the event purpose https://review.openstack.org/445242 | 22:02 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Support for dependent pipelines with github https://review.openstack.org/445292 | 22:02 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Allow github trigger to match on branches/refs https://review.openstack.org/445625 | 22:03 |
| *** jkilpatr has joined #zuul | 22:03 | |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Block sphinx 1.6 https://review.openstack.org/465179 | 22:08 |
| jlk | so... what are job variants? | 22:11 |
| mordred | jlk: I'm totally going to explain this wrong ... | 22:14 |
| mordred | jlk: but job variants are the thing that let you say stuff like "if the branch is stable/kilo, please run this on a trusty node instead of a xenial node" | 22:15 |
| jlk | hrm. | 22:16 |
| mordred | jlk: there's a section: http://specs.openstack.org/openstack-infra/infra-specs/specs/zuulv3.html#projects | 22:16 |
| mordred | that talks about it a little | 22:17 |
| mordred | although it occurs to me that the zuul docs probably want a terminology/concept primer | 22:17 |
| jlk | yeah, okay. Wondering if because noop is a job built in, that something is missing and it's choking on the variant stuff | 22:19 |
| mordred | oh- interesting question | 22:21 |
| jlk | but I would think it would choke in CI as a lot of the tests use noop | 22:22 |
| pabelanger | oh neat, github patches merge | 22:25 |
| pabelanger | time to setup a 3rd party CI test with github | 22:25 |
| jlk | some! | 22:26 |
| jlk | not all | 22:27 |
| mordred | pabelanger: yah - we have 2 really good candidates for that | 22:27 |
| mordred | pabelanger: we should lkely set up an "openstack-zuul" user - since the current user we use for things on gh is "openstack-gerrit" - which would be a weird user to test results to other communities with | 22:30 |
| mordred | actually - I say that ... | 22:30 |
| mordred | jlk: is that the "right" thing to do - or is there a different thing one wants to do for CI integrations? | 22:30 |
| pabelanger | mordred: ya, openstack-zuul seems sane to me | 22:31 |
| mordred | pabelanger: it seems that the fine folks at bonny have set up an "integration": https://github.com/integration/bonnyci | 22:33 |
| pabelanger | Oh, what is that? | 22:36 |
| pabelanger | something needed for github access? | 22:36 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add waitForImage to test_lost_request https://review.openstack.org/464075 | 22:37 |
| mordred | pabelanger: it's the way you register a service with github - and then projects that want to start interacting with your service just add it to their project | 22:38 |
| jlk | reading | 22:39 |
| jlk | what is it you want to do exactly? | 22:39 |
| jlk | There's two ways Zuul can auth to interact with github API. One is with a user's API key, zuul then impersonates said user. | 22:40 |
| jlk | the other way is as an integration bot, in which the integration service (zuul) will get short-lived access tokens that are scoped to specific projects (repositories) and are allowed to perform specific actions. | 22:40 |
| pabelanger | ya, API is what I was thinking. Just want to read a PR, run local zuul, and see what happens | 22:41 |
| pabelanger | I doubt I'd even comment at this point | 22:41 |
| jlk | You might be able to get away without even an API key | 22:42 |
| jlk | but I haven't tested that | 22:42 |
| pabelanger | Ya, I would assume that too | 22:42 |
| jlk | anonymously you can query the API and clone the repos | 22:42 |
| mordred | pabelanger: nod - for that absolutely, api seems the right way | 22:42 |
| pabelanger | but ya, testing | 22:42 |
| jlk | So what you need, is to stand up a zuul-scheduler that has a public exposed webapp (8001) | 22:43 |
| mordred | pabelanger: I think as we start poking at having Infra run tests triggered by PRs in associated projects - we'll want an actual integration bot | 22:43 |
| mordred | but starting with API sounds great | 22:43 |
| pabelanger | mordred: Ya, openstack-infra make sense | 22:43 |
| jlk | and configure a github repo's webhooks to send all events to that url, http://<whatever>:8001/connection/<name_of_github_connection>/payload | 22:43 |
| jlk | (send type json) | 22:43 |
| jlk | oooooh! | 22:57 |
| jlk | I think I just found a rather interesting bug! | 22:57 |
| jlk | can somebody help me rubber ducky explain this? | 22:57 |
| mordred | jlk: I can try? | 22:59 |
| jlk | Okay, so I have a .zuul.yaml file in a untrusted repo | 23:00 |
| jlk | and in it there is a project definition, with a job | 23:00 |
| jlk | http://paste.openstack.org/show/609724/ | 23:00 |
| jlk | this is on master branch | 23:00 |
| jlk | To test a pull request action, I opened a pull request, within the same repo. So then there was a second branch, "ferb" that had this exact same file. | 23:01 |
| jlk | Zuul, when started, poked at things to find configs, found multiple branches of this repo, found zuul files in both branches, and eventually blew up | 23:02 |
| jlk | Remove the second branch, everything starts up fine | 23:02 |
| jlk | something in the code that's trying to mash together config from multiple branches (whyyyy) is creating havoc | 23:02 |
| jlk | and I don't know that we have a test for this scenario anywhere in zuul | 23:04 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Update devstack plugin for security groups https://review.openstack.org/464709 | 23:07 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add check for ZK node empty data https://review.openstack.org/464665 | 23:07 |
| jlk | mordred: does that sound sane? | 23:15 |
| * mordred reading/grokking | 23:15 | |
| mordred | jlk: it sounds sane - although not immediately why it broke. I believe the reason we look for config in branches is so that, for instance, nova's stable/newton branch can have one set of jobs and nova's master branch can have a different set of jobs - although typing that I feel stupid, so I might be wrong | 23:17 |
| mordred | I would have thought that things in the branch's .zuul.yaml would have an implied branch: ferb on them | 23:18 |
| mordred | but at the very least I think this points to a place where there's either an edge case, a need for better docs - or both | 23:18 |
| jlk | yeah I don't know what zuul does internally when presented with multiple branches of zuul files | 23:39 |
| mordred | jlk: zuul/configloader.py _getImpliedBranches | 23:44 |
| mordred | there's also a comment down in fromYaml | 23:45 |
| jlk | oh this is relatively new code I think | 23:48 |
| jlk | May 3rd | 23:48 |
| jlk | a5a12494dc34866388a772c75f79c40f4caa9edb | 23:48 |
| mordred | jlk: yah - so I think the intent is that what you did shoudl work and not confuse anything | 23:51 |
| mordred | jlk: so maybe something about the noop job is, in fact, the thing that's breaking here? | 23:52 |
| jlk | I think there may be a small bug in how the variant is initialized and something is getting a list object instead of a dict | 23:52 |
| mordred | oh - well that would also suck | 23:52 |
| jlk | I'm still reading the code | 23:52 |
| jlk | I need to figure out a better way to pdb around in this container, I'm not sure how to restart this execution | 23:52 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!