| jlk | well that's interesting | 00:15 |
|---|---|---|
| jlk | one of the "variants" is a job itself, the other is a list of the job itself | 00:15 |
| jlk | I've been messed up all this time because my docker deploy is using my github-v3 branch, which hasn't been rebased in a while. But the feature branch has some changes on it that maybe fix this problem. | 01:06 |
| *** harlowja has quit IRC | 01:08 | |
| mordred | jlk: well - hopefully the days of you being on a long-term branch are coming to a close | 01:08 |
| jlk | I've seriously messed up my mac, it thinks I'm user 501, which doesn't actually match my user name | 01:10 |
| jlk | time for a reboot! | 01:10 |
| jlk | it's been 30 days, I guess that's okay | 01:10 |
| jlk | 10~ split terminal windows with 10+ vim processes in each. I'm sure I'll remember what I was doing | 01:11 |
| mordred | jlk: oh totaly. no context will be lost at all! | 01:14 |
| *** dkranz_ has quit IRC | 01:39 | |
| jlk | Well, the good news is that after fully rebasing, the problem persists. | 01:41 |
| leifmadsen | good news everyone! | 01:41 |
| jlk | I need to like tag a data object and get an pdb point every time that data object has something added to it | 02:46 |
| leifmadsen | mordred: ugh, my laptop locked up, and my znc didn't have enough scroll back (or the right kind of scroll back / history) for all the links you posted earlier. If you still have that blob of data, would you mind pastebin'ing it? (or anyone else) | 03:03 |
| jlk | Ooooh I think I found it | 03:32 |
| *** auggy has left #zuul | 03:40 | |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Extend job list when inheriting jobs https://review.openstack.org/465393 | 04:04 |
| jlk | mordred: jeblair not sure if you want a test for ^^ | 04:05 |
| *** harlowja has joined #zuul | 04:33 | |
| jlk | Does nodepool v3 use mysql any more? | 04:42 |
| jlk | looks like no | 04:43 |
| tobiash | jlk: mysql is gone with v3 :) | 05:01 |
| jlk | huzahh | 05:01 |
| SpamapS | (except when you want the sql reporter ;) | 05:02 |
| jlk | yeah | 05:02 |
| tobiash | SpamapS: isn't that part of zuul? | 05:02 |
| jlk | don't need that for what I'm tossing together | 05:02 |
| SpamapS | it is, just pointing out mysql moved from nodepool to zuul :) | 05:03 |
| jlk | which is funny, because in our v2.5 deployment, we were running mysql on the zuul node, instead of on the nodepool node | 05:04 |
| tobiash | so things are moving towards the right place | 05:05 |
| jlk | but for v3 we may not use the sql reporter | 05:06 |
| SpamapS | jlk: I could see it being useful for building a webapp kind of like the openstack health dashboard. | 05:07 |
| SpamapS | but otherwise probably not | 05:07 |
| jlk | fair | 05:07 |
| jlk | https://github.com/j2sol/z8s zuul v3 mostly running in containers. I say mostly because it's the nodepool part that needs more work. | 05:49 |
| jlk | I could use some help getting this to work without nodepool, if possible. | 05:49 |
| SpamapS | NEAT | 05:50 |
| * SpamapS was just reading it | 05:50 | |
| SpamapS | jlk: I'm guessing dib + docker == fail? | 05:50 |
| SpamapS | though not sure how it would fail | 05:51 |
| SpamapS | just guessing | 05:51 |
| jlk | haven't really got it to the point where it's trying to dib anything. I hvaen't defined any images yet | 05:52 |
| jlk | but the zuul side will respond to webhooks shoved at it, will clone the repo in question, and try to fire off my noop job | 05:53 |
| *** harlowja has quit IRC | 05:54 | |
| * jlk out | 05:55 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Add SSH Agent Primitives and usage https://review.openstack.org/462712 | 06:08 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Add support for bwrap https://review.openstack.org/453851 | 06:08 |
| tobiash | jlk: what are the problems with nodepool? | 06:18 |
| tobiash | I have currently everything running in docker | 06:18 |
| tobiash | (including nodepool-builder with dib) | 06:20 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Fix passing labels to Gerrit when they are not defined in All-Projects https://review.openstack.org/465420 | 06:35 |
| *** DangerousDaren has joined #zuul | 06:36 | |
| *** mgagne has quit IRC | 06:53 | |
| *** mgagne has joined #zuul | 06:56 | |
| *** mgagne is now known as Guest24103 | 06:56 | |
| *** bstinson has quit IRC | 06:58 | |
| *** bstinson has joined #zuul | 06:58 | |
| *** bhavik1 has joined #zuul | 07:08 | |
| *** bhavik1 has quit IRC | 07:52 | |
| *** isaacb has joined #zuul | 08:22 | |
| *** isaacb has quit IRC | 09:16 | |
| *** isaacb has joined #zuul | 09:25 | |
| *** Cibo_ has joined #zuul | 10:01 | |
| *** Cibo_ has quit IRC | 10:18 | |
| *** Cibo_ has joined #zuul | 10:27 | |
| *** jkilpatr has quit IRC | 10:38 | |
| *** Cibo_ has quit IRC | 10:48 | |
| *** jkilpatr has joined #zuul | 11:00 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: sql-reporter: add support for Ref change https://review.openstack.org/465539 | 12:38 |
| tristanC | SpamapS: Greeting sir, I'm actually trying to get a Jenkins-like dashboard for zuul jobs and found an issue with the sql reporter when used in the post pipeline... would you mind checking https://review.openstack.org/465539 ? | 12:48 |
| mordred | tristanC: +2 from me | 13:07 |
| mordred | Shrews: I +2d the stack at https://review.openstack.org/#/q/topic:py3-nodepool - which should also pass tests now with your test fixes | 13:07 |
| Shrews | mordred: ok. still battling some sort of illness, but will try to find some energy to look soon | 13:33 |
| mordred | Shrews: no worries/rush | 13:35 |
| mordred | Shrews: go be un-sick | 13:35 |
| *** dkranz_ has joined #zuul | 13:36 | |
| leifmadsen | woah, a Jenkins like dashboard for Zuul? | 13:46 |
| leifmadsen | tell me more! | 13:46 |
| *** Cibo_ has joined #zuul | 13:59 | |
| Shrews | pabelanger: mordred: the nodepool py3 changes lgtm except for https://review.openstack.org/463594 | 14:18 |
| *** rcarrillocruz has quit IRC | 14:20 | |
| Shrews | pabelanger: i think https://pythonhosted.org/six/#six.reraise will work there | 14:20 |
| *** Cibo_ has quit IRC | 14:26 | |
| *** Guest24103 is now known as mgagne | 14:32 | |
| *** mgagne has quit IRC | 14:32 | |
| *** mgagne has joined #zuul | 14:32 | |
| jlk | tobiash: no problems yet per se, I haven't tried to get it to build images yet. I was hoping I could have zuul execute noop job without needing a node, thus nodepool, but that wasn't working right. | 14:39 |
| jlk | tobiash: do you have your docker files and docker-compose files anywhere public? | 14:40 |
| clarkb | jlk the zuul 'noop' job should be a special case that just returns a successful status without executing remotely | 14:43 |
| jlk | clarkb: I totally agree with that statement! | 14:45 |
| jlk | however my zuul install does not. | 14:45 |
| clarkb | I mean that is how it is implemented. at least for v2.5 | 14:45 |
| jlk | let me bring up the stack and stuff an event into it | 14:46 |
| jlk | (this is really useful to have it all dockerized | 14:46 |
| jlk | ) | 14:46 |
| jlk | zuul-scheduler | DEBUG:zuul.IndependentPipelineManager:Adding node request <NodeRequest 100-0000000000 <NodeSet OrderedDict()>> for job noop to item <QueueItem 0x7f70044f8cd0 for <Change 0x7f7004487a10 1,bb71d444ce36ea561bf98d83d05d2ccefe38525a> in check> | 14:47 |
| jlk | and then it sits there, because it doesn't get a node | 14:47 |
| pabelanger | Shrews: I'll fix shortly | 14:48 |
| *** isaacb has quit IRC | 16:00 | |
| *** DangerousDaren has quit IRC | 16:07 | |
| jeblair | SpamapS, jlk: (from yesterday scrollback): at this point you can theoretically not run a zuul-merger (because zuul-executor implements the same functions), and i was going to remove it, however, due to the large amount of git repo operations related to configuration, i think it's best to keep it around because we may end up wanting to scale it separately from executors. | 16:09 |
| jlk | jeblair: I did notice that there is one function I don't think the executor does, which is the merger:update function. | 16:11 |
| jlk | jeblair: if we have a merger, it's doing information gathering things, right, it's not preparing trees for use in other places? It doesn't need any ports exposed or anything? | 16:12 |
| jeblair | mordred, jlk: there's some WIP documentation on job variants in https://review.openstack.org/463328 | 16:13 |
| jeblair | jlk: yeah, merger:update is no longer needed (it means make sure the git repo for this project is up to date on this merger so that a post job can clone from this merger and get the right stuff. that doesn't happen in v3 any more.) | 16:17 |
| jeblair | jlk: we should be able to drop that | 16:17 |
| SpamapS | jeblair: so you're thinking that with the merger we might have them just out there fetching trees and looking for config updates, communicating them to the scheduler? | 16:17 |
| jeblair | jlk: correct, a merger in v3 can be entirely on the internal network. | 16:17 |
| jeblair | SpamapS: exactly | 16:17 |
| SpamapS | I do like that, especially until we get further down the road and have scale-out schedulers | 16:18 |
| jeblair | SpamapS: and since we've realized we still need to do the speculative merge before sending things to the executor, that's work that mergers can do while executors do other things | 16:18 |
| *** rcarrillocruz has joined #zuul | 16:18 | |
| SpamapS | tristanC: Oh cool! I didn't know you were working on a dashboard. Is there one up somewhere that I could poke at? | 16:18 |
| jlk | jeblair: cool. I'll toss up a review to get rid of the update functionality. | 16:19 |
| *** isaacb has joined #zuul | 16:20 | |
| jlk | jeblair: I tracked down my variants issue to a really small logic bug though, so there's that :) | 16:20 |
| jeblair | jlk: re nodepool/noop -- while there is special-casing of noop around the executor (so nothing is actually run), i don't think we special cased it in the nodepool interface. actually, what should be special cased there is not noop per-se, but rather, a job with an empty nodeset. | 16:20 |
| jeblair | jlk: cool, that was my next question -- the list/str thing was the only variants problem? | 16:21 |
| jlk | well, it was the only thing causing a traceback | 16:21 |
| jeblair | jlk: therefore it must be the only error, right? :) | 16:22 |
| jlk | of course! Tests passed! | 16:22 |
| SpamapS | "the test suite passed" is the new "it compiles" | 16:22 |
| SpamapS | ship it! | 16:22 |
| jlk | I probably should have added some sort of test that caused the traceback, but... I didn't. | 16:23 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: WIP: add configuration documentation https://review.openstack.org/463328 | 16:23 |
| jeblair | more docs from 36,000ft ^ | 16:24 |
| SpamapS | ugh. this ssh agent stuff is proving more complex than I had expected.. seems I've created some kind of race and it's manifesting as locked up test suite | 16:27 |
| jeblair | i'm declaring a timeout on simple changes with +2 from before summit and single-core approving them | 16:29 |
| jeblair | (my goal this week is to get my zuul gertty dashboard back to a managable size) | 16:30 |
| jlk | +1 | 16:32 |
| *** harlowja has joined #zuul | 16:33 | |
| jeblair | pabelanger: i'm +3ing https://review.openstack.org/451470 with a comment | 16:33 |
| jlk | jeblair: so I don't know how the test suite gets things to work without a nodepool. Is it not possible at this point outside of the unit test engine to convince zuul to execute a job and consider it successful? For the sake of live testing zuul driver interface code without bothering to spin up a whole functional nodepool. | 16:33 |
| jeblair | pabelanger: however, it looks like 451470 is based on an abandoned patch? you want to rebase that? | 16:34 |
| jeblair | jlk: the test suite has a fake nodepool which satisfies those actual requests for zero-node nodesets | 16:34 |
| jeblair | jlk: so yeah, we need the short-circuit node request logic added to do what you want | 16:34 |
| jlk | ahhh | 16:34 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Move test_test_config to a cmd-focused test suite https://review.openstack.org/461493 | 16:35 |
| SpamapS | jeblair: does that include gear changes? ;-) | 16:35 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Re-enable test_live_reconfiguration_merge_conflict https://review.openstack.org/461502 | 16:36 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Log return code on failed dib build https://review.openstack.org/450782 | 16:36 |
| jeblair | SpamapS: i wouldn't consider those trivial... i'd like mordred or clarkb to take a look, though if they decline, i'm okay with you and i being happy with them. :) | 16:37 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Re-enable test_live_reconfiguration_failed_root https://review.openstack.org/461508 | 16:37 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Re-enable test_live_reconfiguration_failed_job https://review.openstack.org/461512 | 16:37 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Re-enable test_live_reconfiguration_shared_queue https://review.openstack.org/461522 | 16:37 |
| SpamapS | jeblair: have we ever identified any other significant users of gear? | 16:41 |
| jeblair | SpamapS: not that i'm aware of (which isn't to say they don't exist, they may just be quiet)... clarkb, fungi: ^? | 16:41 |
| clarkb | I know HPE was looking at using it for some cloud things back when they clouded things | 16:42 |
| jlk | I seem to recall it being a thing that Datadog can interact with, so apparently there are SOME users out there | 16:42 |
| SpamapS | jeblair: I was just hoping we might have somebody out there who has used gear that isn't us, and would like to try porting to the new API. I did use it for gearstore/gearhorn, which were experiments in making oslo.messaging gear drivers.. but that's still me, using my API change. ;-) | 16:43 |
| jlk | http://docs.datadoghq.com/integrations/gearman/ | 16:43 |
| jeblair | Shrews: this wants a once-over from you: https://review.openstack.org/458073 | 16:43 |
| fungi | i'm not at all aware of other consumers, but it comes as no great surprise to me if there are any | 16:43 |
| jlk | I'm assuming here that "gear" == "gearman" but I could be very wrong | 16:43 |
| SpamapS | Also there's a group that wrote yet another gearmand port, in go, that wants to look at translating the gearman protocol into gRPC | 16:43 |
| SpamapS | I _really_ like that idea. :) | 16:44 |
| jeblair | jlk: i think SpamapS was specifically focused on the 'gear' library and the slight api modifications we are considering | 16:44 |
| SpamapS | (because we'd finally get usable auth for free) | 16:44 |
| SpamapS | jlk: yeah, gearman itself is actually all over the place, quietly, but the python gear library .. we don't know if people depend on it. | 16:44 |
| SpamapS | We should assume they might | 16:44 |
| SpamapS | though it is 0.x¸ so we only need to be cordial, not fanatic, about breaking things too badly. | 16:45 |
| jlk | oops, sorry | 16:45 |
| SpamapS | I believe if we had resurrected gearman earlier, it might be in use more.. but unfortunately, it fell victim to a few life challenges in its maintainer(krow)'s life for a few years. | 16:46 |
| jeblair | i'm still pretty comfortable with "the py2 api isn't substantially changing, you only notice with py3, and py3 was so bad earlier i kind of hope no one was using it". but i agree, that's an entirely hypothetical construct it would be nice to realize or falsify. :) | 16:46 |
| SpamapS | jeblair: I would suggest that we make the change to .name (which is a little tiny bit 2.x intrusive, for like, users of non-utf8 queue names...) release as 0.8 | 16:47 |
| clarkb | I haven't had time to look at the changes, but we aren't adding utf8 support in a set of new classes that would be backward compat? | 16:47 |
| clarkb | (I thought that was the plan but may have misremembered) | 16:47 |
| SpamapS | clarkb: that's exactly what we did. | 16:47 |
| clarkb | oh then nothing should change for existing users right? (excluding unexpected bugs) | 16:47 |
| SpamapS | clarkb: but we also just made .name utf8 always, and added a binary class for those unfortunate enough to have done that evil thing. | 16:47 |
| SpamapS | clarkb: now that you have it top of mind.... | 16:48 |
| clarkb | gotcha so .name is the big change | 16:48 |
| jeblair | right, so it's technically an api break for the one person who registered utf8 pile-of-poo as a function name | 16:48 |
| SpamapS | https://review.openstack.org/461172 | 16:48 |
| SpamapS | jeblair: actually I think it won't break that person in py2 | 16:49 |
| jeblair | SpamapS: even better :) | 16:49 |
| SpamapS | it would break the chinese user who used GB2312 | 16:49 |
| SpamapS | which would also break the python interpreter in literals... | 16:50 |
| SpamapS | so they'd have to have _REALLY_ tried hard to make this easy for us to break. | 16:50 |
| * SpamapS may also be paying way too much attention to the corner cases here ;) | 16:50 | |
| SpamapS | actually GB2312 is a bad example... since it is a set of code points like unicode.. and both well known encodings are ascii compatible (EUC-CN and HZ) | 16:51 |
| SpamapS | so yeah, corner sufficiently shrunk :) | 16:52 |
| jeblair | Shrews: https://review.openstack.org/463338 | 16:54 |
| mordred | yah - I just +A'd the first patch ... it seems the only people who would be broken would be py3 users. py3 users are so unlikely that I'm comfortable with the corner cases | 16:54 |
| mordred | jeblair: you've +2'd https://review.openstack.org/#/c/461172 before but SpamapS revised it | 16:54 |
| jeblair | mordred: thx, diffing | 16:55 |
| jeblair | +3 | 16:55 |
| mordred | jeblair: while you're at it - https://review.openstack.org/#/c/400472 too - since we're adding py3 support and those calls raise a warning in py3 - seems like it'll be annoying if we don't merge that | 16:56 |
| jeblair | done | 16:57 |
| mordred | wow. that's everything that's not SpamapS thundering-herd patches | 16:57 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Python 3.x incompatible use of print operator https://review.openstack.org/463586 | 16:58 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove FakeFile from fakeprovider.py https://review.openstack.org/463587 | 16:58 |
| SpamapS | the thundering-herd patches can definitely wait | 17:00 |
| *** Cibo_ has joined #zuul | 17:00 | |
| SpamapS | in fact I think since writing them I discovered that libgearman failed to implement the client server selection hash algorithm correctly too | 17:00 |
| SpamapS | funny thing about gearman... I am pretty sure if they hadn't written a PECL PHP extension for libgearman and gotten it included early on in php's docs, it would have died a long, long time ago. PHP users keep showing up with insane gearman uses that would be entirely fine done as MySQL queries. :-P | 17:01 |
| jeblair | jamielennox: can you refresh https://review.openstack.org/446308 and https://review.openstack.org/447734 when you have a chance? | 17:10 |
| *** isaacb has quit IRC | 17:12 | |
| clarkb | SpamapS: ok got through https://review.openstack.org/#/c/461172/10 general shape of it seems fine. But some small belts and suspenders things noted inline | 17:14 |
| jeblair | pabelanger: https://review.openstack.org/452991 is out of date now i think? | 17:21 |
| jeblair | pabelanger: and https://review.openstack.org/454248 | 17:21 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Add some lookup plugin tests https://review.openstack.org/454396 | 17:23 |
| jeblair | i just dropped the WIP commit message from that. hopefully we'll find time to add more of those tests in the future, but for now, let's just land that ^ | 17:23 |
| jeblair | jamielennox: can you take a look at https://review.openstack.org/455508 ? | 17:24 |
| *** harlowja has quit IRC | 17:26 | |
| jlk | jeblair: oops, I reviewed your docs before you pushed patch set 2. Reviewing the new bits, but looks like my comments on version 1 stand. | 17:26 |
| jeblair | jlk: yes, airplanes. ps2 is almost entirely appending new text. i will certainly check both before doing a real revision. | 17:31 |
| jlk | so. much. text. | 17:38 |
| *** Cibo_ has quit IRC | 17:45 | |
| jlk | oh haha. So, looking at the merger:update function, looksl ike that gets called by the cat function, to ensure the repo is updated. | 17:55 |
| jlk | or, not exactly | 17:56 |
| *** jamielennox is now known as jamielennox|away | 17:59 | |
| jlk | Surprise benefit of docker-compose, once you've done an up, you can modify the container, docker-compose stop, docker-compose start and it'll run that modified container. You can modify zuul code within, restart the things and test changes! | 18:02 |
| jlk | don't have to figure out how to re-build the image with the altered zuul code fetched from some staging place | 18:02 |
| SpamapS | clarkb: thanks! | 18:03 |
| mordred | jlk: I was looking atyour code file and thinking it was surprisingly well-suited for being driven by a zuul as well | 18:04 |
| jlk | driven by a zuul? | 18:04 |
| jlk | as in the post-commit deploy stuff? | 18:04 |
| mordred | no - for testing ... on esec | 18:04 |
| jlk | oh, right. Spin up a set of containers, shove an event set through, check the results. | 18:05 |
| mordred | part of the big challenge in thinking about k8s workloads is "how do you get from the source repos that zuul has prepared to those running as images in containers" | 18:05 |
| *** jamielennox|away is now known as jamielennox | 18:05 | |
| mordred | the compose file seems to understand that sometimes you might want to work from some source repos and produce containers from those | 18:05 |
| mordred | so it won't actually solve the whole thing for us - but it might be a nice starting point from which to think | 18:06 |
| mordred | jlk: the most vexing part as I think forward - but that I'm not worrying about TOO much because we havne't actually dug in - is that k8s manifests and helm charts tend to be expressed in terms of container names. but a source repo that builds containers doesn't tend to express the name of the container it's going to build - so given a chart that needs 10 images and 2 source code repos that would provide the | 18:09 |
| mordred | containers for 2 of those - there's no real good 'automatic' way to infer the right thing to do to test the set of changes you have ... | 18:09 |
| mordred | jlk: now - it's VERY possible I'm _way_ overthinking this and that the answer is"the job is just going to have to have a thing that deals with conatiner builds as step one" | 18:09 |
| *** dkranz_ has quit IRC | 18:10 | |
| *** dkranz has joined #zuul | 18:12 | |
| jlk | yeah, I think you're a few turns farther down the road than I am at this point | 18:22 |
| SpamapS | I personally like the idea of build as step 1 | 18:24 |
| SpamapS | 5+ years of "test the bits that you run in production" are embedded deep in my consciousness. | 18:25 |
| SpamapS | so if I can get from build -> [matrix of tests with same built containers ] -> publish said containers | 18:25 |
| SpamapS | I'd be happy as a clam | 18:25 |
| jlk | right | 18:26 |
| jlk | particularly if it's done as gate, where zuul is testing the merge AS IT WILL HAPPEN | 18:26 |
| jlk | so you're essentially building containers of merged code, even if it hasn't yet "merged" in the canonical repo. | 18:26 |
| jlk | the hashes won't change, in theory | 18:26 |
| clarkb | they will because of timestamps unless we get to having zuul push to gerrit and only doing a single merge for all jobs for a gated change | 18:27 |
| jlk | timestamp causes the merge hash to change? | 18:28 |
| clarkb | yes | 18:28 |
| jlk | TIL.... | 18:28 |
| clarkb | this is why I dislike the cherrypick feature in git review | 18:28 |
| clarkb | leads to much confusion | 18:29 |
| jlk | So I can't cryptographically validate that the merge of code Zuul did to perform a test is the exact same as the merge of code that zuul caused to have happen when the tests completed? | 18:29 |
| mordred | nope. although having zuul push the merge rather than click the merge button is a thing we've talked about many times and would allow that | 18:30 |
| jeblair | not unless zuul pushes the work that it did upstream (which is a fairly easy enhancement to make and we have favorably considered doing it). | 18:30 |
| mordred | what jeblair said | 18:30 |
| jeblair | and mordred | 18:30 |
| mordred | I think at the VERY least zuul should grow that option | 18:30 |
| jlk | yeah that would make a lot of folks more comfortable maybe? | 18:31 |
| jlk | dunno | 18:31 |
| mordred | whether or not openstack decides to enable it or continue to have zuul click the button in gerrit is a different question I think | 18:31 |
| clarkb | there are other ways of verifying too fwiw | 18:31 |
| clarkb | hashes of patch diffs for example | 18:31 |
| clarkb | (did someone say darcs) | 18:31 |
| jeblair | jlk, SpamapS: we tested this with gerrit, btw -- if zuul pushes a change to a branch, gerrit will automatically close the change... it would be interesting to know if github would auto-close a PR if someone pushed the commits in that pr to the branch it was targeting. | 18:31 |
| mordred | I believe it does, yes | 18:31 |
| jlk | It is supposed to yes | 18:32 |
| jeblair | cool | 18:32 |
| jlk | however... | 18:32 |
| jlk | I don't think the github integration API allows us to push changes. | 18:32 |
| jeblair | jlk: that's okay, it's just a git op | 18:32 |
| jlk | since it's a bot instead of ostensibly a human. | 18:32 |
| jlk | it's a git op, but it's more of a credentials problem on the other side | 18:32 |
| jeblair | jlk: oh, you're saying projects would have to allow zuul as a team member or something | 18:32 |
| mordred | jeblair: so - an "integratoin" doesn't have a github user account associated with it | 18:32 |
| jlk | the repo owner on github would have to grant zuul the rights to push | 18:33 |
| jeblair | fun problem | 18:33 |
| jeblair | (incidentally, this is useful on its own, but is also essentially a pre-req for bidirectional cross-repo-dependencies (which are a two-phase commit problem)) | 18:33 |
| mordred | yah. althugh the other bit is the "publish the containers I built from the commit in gate - don't re-build them in post to publish them" | 18:34 |
| mordred | so there are a couple of places for fun mind thinking here | 18:34 |
| jeblair | mordred: yeah, another thing which is easier with stable shas, but likely possible to work around | 18:34 |
| jeblair | i need to go get a fried chicken sandwhich before all the hipsters show up | 18:35 |
| mordred | jeblair: you can never beat the hipsters | 18:35 |
| jlk | not with that attitude you won't. | 18:37 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Remove unused merger:update task https://review.openstack.org/465701 | 18:38 |
| jlk | re earlier convo ^^ | 18:38 |
| jlk | is there a gate_recheck in our gerrit/zuul? | 18:45 |
| *** jkilpatr has quit IRC | 18:50 | |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Add cachecontrol to requests to github https://review.openstack.org/461587 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Add support for requiring github pr head status https://review.openstack.org/449390 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Adds github triggering from status updates https://review.openstack.org/453844 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Implement trigger require/reject-approvals https://review.openstack.org/463387 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Implement pipeline requirement on github reviews https://review.openstack.org/453845 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Implement github trigger requirement status https://review.openstack.org/463386 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Actually filter status triggers https://review.openstack.org/463385 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Create github client each time it's called. https://review.openstack.org/463421 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Use integration_id with github https://review.openstack.org/463420 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Refactor integration key loading to always load https://review.openstack.org/463423 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Represent github change ID in status page by PR number https://review.openstack.org/460716 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Decode JSON body once for requests https://review.openstack.org/463422 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Include tenant name in github context https://review.openstack.org/463871 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Save installation ids to a cache and fetch them per project https://review.openstack.org/463424 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Catch integration key file not found in zuul https://review.openstack.org/463425 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Include exc_info in reporter failure https://review.openstack.org/460765 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Add trigger capability on github pr review https://review.openstack.org/449365 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Ensure PRs arent rejected for stale negative reviews https://review.openstack.org/460700 | 18:51 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Comment on PRs if a remote call to merge a change failed https://review.openstack.org/460762 | 18:51 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Extract ansible library file copying to a function https://review.openstack.org/465705 | 18:53 |
| mordred | SpamapS: ^^ noticed that while reviewing the first patch in your py3k stack | 18:54 |
| mordred | jlk, jeblair: re: recheck - it struck me yesterday (I needed to recheck a stack of changes that had all failed because of the sphinx issue) | 18:54 |
| *** jkilpatr has joined #zuul | 18:54 | |
| mordred | that a "recheck stack" would be nice - basically if you left it on the latest change of a stack it would recheck that change and all it depended on | 18:55 |
| jlk | we've got "recheck" which should cause a new +1 vote to come in, right? which would then potentially trigger gate, if other things are in place | 18:55 |
| mordred | it could also be a super easy DDOS - so may be a terrible idea | 18:55 |
| mordred | jlk: yes | 18:55 |
| jlk | in Bonny, we did a gate_recheck so that you could skip the check pipline | 18:56 |
| clarkb | but only if you dont already have a +1 or your gerrit is newer than ours | 18:56 |
| jlk | oh :( | 18:56 |
| clarkb | because gerrit wont emit the +1 if its already there on pur version | 18:56 |
| jlk | yeah I think github is the same | 18:56 |
| jlk | although I'd have to verify that | 18:56 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Copy dirs to handle __pycache__ in py3 https://review.openstack.org/463890 | 19:00 |
| *** mordred has quit IRC | 19:05 | |
| *** mordred has joined #zuul | 19:06 | |
| pabelanger | just catching up on back scroll. Internet outage for last 4 hours | 19:13 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Rename nodepool.py to launcher.py https://review.openstack.org/463807 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Ensure zookeeper_servers is a list https://review.openstack.org/463880 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Use six.moves.urllib for python3 compat https://review.openstack.org/463595 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Update exception message handling for python3 https://review.openstack.org/464034 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Sort flavors with operator.itemgetter('ram') https://review.openstack.org/463998 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Replace dict.iteritems() with dict.items() for python3 https://review.openstack.org/464026 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Switch to next(generator) for python3 https://review.openstack.org/464040 | 19:20 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool feature/zuulv3: Fix imports for python3 https://review.openstack.org/463808 | 19:20 |
| pabelanger | Shrews: jeblair: okay, I have rebased without 463594 for now | 19:21 |
| pabelanger | so I can read up more about it | 19:21 |
| SpamapS | mordred: \o/ yes I kept twitching patching those bits. Thanks! | 19:27 |
| * mordred is doing funny little cleanup patches today while reviewing other things | 19:33 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Convert some leftover changeish mentions to ref https://review.openstack.org/465718 | 19:33 |
| mordred | like that | 19:33 |
| mordred | (noticed while reviewing one of the github stack) | 19:33 |
| * clarkb is fighting the post travel plague and trying to get useful things like operating system updates and network infrastructure changes in place before needing all that hardware later in the week | 19:33 | |
| *** harlowja has joined #zuul | 19:39 | |
| *** harlowja has quit IRC | 19:53 | |
| jlk | weeeird. Gertty isn't showing me mordred's latest change | 19:55 |
| *** harlowja has joined #zuul | 19:57 | |
| jeblair | jlk: it has the github-v3 topic, in case you're looking at a dashboard that filters those | 19:58 |
| *** harlowja has quit IRC | 20:00 | |
| jlk | I have it showing all changes in openstack-infra/zuul | 20:03 |
| jlk | the most recent number it'll show me is 465701 | 20:03 |
| jeblair | jlk: and you're subscribed to zuul? | 20:03 |
| *** jkilpatr has quit IRC | 20:04 | |
| jlk | yeah | 20:05 |
| jlk | I'm going to run with logging | 20:05 |
| jeblair | jlk: ++ | 20:05 |
| jlk | oh now 465705 showed up | 20:05 |
| jeblair | maybe it was offline, or something was stuck... | 20:05 |
| jlk | 2017-05-17 13:07:25,747 GET: https://review.openstack.org/a/changes/?n=500&q=project:openstack-infra/zuul -age:19s | 20:07 |
| jlk | 2017-05-17 13:07:26,206 HTTP status code: 200 | 20:07 |
| *** jkilpatr has joined #zuul | 20:08 | |
| *** dkranz has quit IRC | 20:09 | |
| jlk | ¯\_(ツ)_/¯ | 20:09 |
| SpamapS | jlk: I find that ctrl-o is good for those situations. :) | 20:09 |
| SpamapS | as in, type in the number.. and gertty goes and finds it | 20:10 |
| jlk | oh I think there is something weird on OSX where ctrl o doesn't work right | 20:11 |
| SpamapS | maybe ask siri to press ctrl-o? | 20:12 |
| jeblair | jlk: yeah, you may want to remap that (there's an example for exactly that command in the sample config file) | 20:12 |
| jlk | nod that's where I remember seeing it | 20:13 |
| jlk | Uh, is gate-zuul-python27-ubuntu-xenial being flappy? EG https://review.openstack.org/#/c/445644/ it passed the test in check at 11:54, and then failed it in gate at 12:00 | 20:19 |
| jeblair | jlk: hrm, looks like it was still running when it hit the timeout | 20:21 |
| jeblair | maybe we just need to bump it | 20:23 |
| clarkb | File "tests/unit/test_github_driver.py", line 401, in test_failed_changes <- timeout waiting for zuul to settle | 20:23 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Better merge message for GitHub pull reqeusts https://review.openstack.org/445644 | 21:04 |
| *** dougbtv_ has joined #zuul | 21:19 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: GitHub file matching support https://review.openstack.org/446113 | 21:19 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Log GitHub API rate limit https://review.openstack.org/446150 | 21:22 |
| *** dougbtv has quit IRC | 21:22 | |
| *** harlowja has joined #zuul | 22:02 | |
| *** jkilpatr has quit IRC | 22:07 | |
| *** dougbtv__ has joined #zuul | 22:15 | |
| *** dougbtv_ has quit IRC | 22:17 | |
| *** dougbtv_ has joined #zuul | 22:18 | |
| *** harlowja has quit IRC | 22:18 | |
| *** dougbtv__ has quit IRC | 22:20 | |
| *** jkilpatr has joined #zuul | 22:24 | |
| *** dougbtv__ has joined #zuul | 22:26 | |
| *** dougbtv_ has quit IRC | 22:28 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Set filter according to PR/Change in URL https://review.openstack.org/446782 | 22:59 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Test gerrit and github drivers in same tenant https://review.openstack.org/448257 | 23:00 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add support for github enterprise https://review.openstack.org/449258 | 23:00 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add trigger capability on github pr review https://review.openstack.org/449365 | 23:06 |
| *** dougbtv_ has joined #zuul | 23:25 | |
| tristanC | SpamapS: (re dashboard) the goal is to inspect the last few post/periodic job to ease debug, the first attempt is https://review.openstack.org/#/c/464524/ | 23:26 |
| *** dougbtv__ has quit IRC | 23:27 | |
| jeblair | tristanC: cool -- we probably won't land any feature additions to the zuul v2.5 launcher -- maybe we can map that into zuul v3? | 23:31 |
| tristanC | SpamapS: well a better one based on the sql-reporter needs to be groomed first, I'll probably go for a basic pecan server with angular client, but perhaps that can be part of the webapp | 23:31 |
| tristanC | jeblair: indeed hopefully this could be ported to zuul v3, but unfortunaly the plan is to migrate from jenkins to zuul-launcher first | 23:33 |
| jeblair | tristanC: ++ using sql reporter. mtreinish was also looking at using the sqlreporter inside of openstack health. | 23:34 |
| jeblair | i don't think he's started on that | 23:35 |
| tristanC | jeblair: excellent, I'll get in touch. thanks! | 23:35 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!