| jlk | So, when tests start up, a bunch of stuff gets created in /tmp/<TMP>/ like upstream dirs and org dirs and whatnot | 00:22 |
|---|---|---|
| jlk | Where is the code that creates all of that? I'm hitting stuff missing from there, and I'm not sure why | 00:22 |
| jeblair | jlk: yes! it should be in tests/base.py | 00:23 |
| jlk | ah, in the ZuulTestCase class | 00:23 |
| jeblair | jlk: setup_config and copyDirToRepo specifically | 00:24 |
| jeblair | jlk: (all of which are invoked in setUp) | 00:24 |
| jlk | oh interesting, hard coded for org/project3->6 | 00:24 |
| jeblair | jlk: oh, yeah, i want to remove those | 00:25 |
| jeblair | i just only got around to removing up to project2 | 00:25 |
| jeblair | jlk: but the "new way" is to make a "fixtures/config/git/" directory for your test class, and put whatever git repos you need in there (as standard directories -- copyDirToRepo makes it into a git repo for you) | 00:26 |
| jeblair | jhesketh: responded to both | 00:26 |
| jhesketh | ta | 00:28 |
| jlk | bam, thanks jeblair | 00:31 |
| jeblair | we should write prose for that... it belongs in https://docs.openstack.org/infra/zuul/feature/zuulv3/testing.html | 00:33 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove job trees from pipelines https://review.openstack.org/435944 | 00:41 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Handle merge mode in repeated project configs https://review.openstack.org/435945 | 00:41 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove inscrutable TODO https://review.openstack.org/435946 | 00:45 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add note about pipelines and dynamic layouts https://review.openstack.org/435947 | 00:45 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Use hostname in Nodepool requests https://review.openstack.org/435949 | 00:45 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove unused mergeChangeQueue method https://review.openstack.org/435950 | 00:45 |
| jlk | ever seen something like: with open(src, 'rb') as fsrc: | 00:46 |
| jlk | IOError: [Errno 21] Is a directory: '/home/ubuntu/zuul/zuul/ansible/action/__pycache__' | 00:46 |
| *** saneax is now known as saneax-_-|AFK | 00:53 | |
| jeblair | jlk: no -- but some quick googling suggests that's a python3-ism? | 00:58 |
| jeblair | jlk: we may want to exclude that when we copy the ansible plugins into place | 00:59 |
| jlk | Something really wrong went into my testing environment | 00:59 |
| jeblair | jlk: (and maybe something else previously made the pycache directory in your source repo even if you're using py2 for tests) | 01:00 |
| jlk | I think I was launching it with python3 on accident | 01:00 |
| jlk | I think I need to grok how tenant level separation is meant to happen | 01:22 |
| jlk | and what that means for drivers | 01:22 |
| jlk | jeblair: I think you mentioned that it might make sense for a connection to be able to register a webapp path that's outside of the tenant/ scope. such as /connection/<name>, or should it still be like /<tenant>/connection/<name> ? | 01:33 |
| jlk | each connection configured, is that per-tenant? | 01:33 |
| *** unrahul has joined #zuul | 01:41 | |
| *** persia has quit IRC | 02:09 | |
| *** persia has joined #zuul | 02:12 | |
| *** abregman has joined #zuul | 06:24 | |
| *** saneax-_-|AFK is now known as saneax | 07:31 | |
| *** isaacb has joined #zuul | 07:52 | |
| *** hashar has joined #zuul | 08:14 | |
| *** saneax is now known as saneax-_-|AFK | 10:07 | |
| *** openstackgerrit has quit IRC | 10:33 | |
| *** openstackgerrit has joined #zuul | 10:34 | |
| openstackgerrit | Javier Peña proposed openstack-infra/zuul master: Find fallback branch in zuul-cloner when indicated branch is missing https://review.openstack.org/442370 | 10:34 |
| openstackgerrit | Javier Peña proposed openstack-infra/zuul master: Find fallback branch in zuul-cloner https://review.openstack.org/442370 | 10:35 |
| *** abregman is now known as abregman|afk | 10:47 | |
| *** isaacb has quit IRC | 10:50 | |
| *** saneax-_-|AFK is now known as saneax | 11:10 | |
| *** isaacb has joined #zuul | 11:34 | |
| *** abregman|afk is now known as abregman | 11:38 | |
| *** isaacb has quit IRC | 12:20 | |
| *** lennyb has joined #zuul | 12:28 | |
| lennyb | Hi, We are using zuul/gerrit in another project and zuul/gerrit should get refs from https url. Looks like it's not supported in zuul, but maybe I am missing something. in the #link https://github.com/openstack-infra/zuul/blob/master/zuul/connection/gerrit.py#L394 I see that I cant provide credentials | 12:28 |
| lennyb | Is it possible to configure zuul, so it will use https user password connection? | 12:29 |
| *** isaacb has joined #zuul | 12:50 | |
| *** abregman has quit IRC | 12:57 | |
| *** abregman has joined #zuul | 13:00 | |
| *** isaacb has quit IRC | 13:14 | |
| rcarrillocruz | Have you thought about putting a reverse proxy in front, passing HTTPS to Zuul HTTP? | 13:15 |
| *** isaacb has joined #zuul | 13:19 | |
| rcarrillocruz | Ah, to connect to Gerrit | 13:21 |
| *** abregman is now known as abregman|mtg | 13:23 | |
| Shrews | i think only ssh auth is supported. jeblair is the best person to answer that question, though | 13:28 |
| dmsimard | jpena had a question about a pep8 gate failure in this review: https://review.openstack.org/#/c/442370/ seems related to pbr/gear mismatches | 13:55 |
| pabelanger | dmsimard: yes, we likely need to bump hacking on master | 14:18 |
| *** abregman|mtg has quit IRC | 14:22 | |
| *** abregman has joined #zuul | 14:23 | |
| *** dmellado has joined #zuul | 14:34 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Update hacking dependency https://review.openstack.org/442533 | 14:42 |
| pabelanger | dmsimard: ^should help | 14:43 |
| jeblair | jlk: connections are global, so you can omit the tenant name | 15:07 |
| openstackgerrit | Javier Peña proposed openstack-infra/zuul master: Find fallback branch in zuul-cloner https://review.openstack.org/442370 | 15:08 |
| openstackgerrit | Javier Peña proposed openstack-infra/zuul master: Find fallback branch in zuul-cloner https://review.openstack.org/442370 | 15:12 |
| *** isaacb has quit IRC | 15:26 | |
| *** saneax is now known as saneax-_-|AFK | 15:35 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Change mutex to counting semaphore https://review.openstack.org/442563 | 15:37 |
| *** hashar has quit IRC | 16:10 | |
| *** abregman has quit IRC | 16:20 | |
| mordred | jeblair: that ^^ is clearly a you patch | 16:30 |
| SpamapS | whoa weird, I git pulled and the number of @skips in unit test dir went up from 25 to 30 since yesterday | 16:45 |
| jeblair | SpamapS: we merged from master; i think we added a lot of tests (yay?) | 16:48 |
| jeblair | SpamapS: 441114 takes care of most of them | 16:49 |
| SpamapS | jeblair: oh cool! | 16:51 |
| SpamapS | that's a good reason then. :-D | 16:51 |
| *** isaacb has joined #zuul | 16:54 | |
| SpamapS | eggshell: did you see the discussion about the low-hanging-fruit tag yesterday? | 16:56 |
| *** hashar has joined #zuul | 16:56 | |
| SpamapS | rattboi: also you? :) | 16:56 |
| eggshell | not I. | 16:56 |
| SpamapS | eggshell: ok, so there's a tag you can search for in openstack-infra/zuul on storyboard that is 'low-hanging-fruit' | 17:00 |
| eggshell | found the list, thanks! | 17:00 |
| SpamapS | let me see if storyboard has a url we can share | 17:00 |
| SpamapS | https://storyboard.openstack.org/#!/search?project_id=679&tags=low-hanging-fruit | 17:01 |
| SpamapS | https://storyboard.openstack.org/#!/search?tags=low-hanging-fruit&project_id=668 for nodepool | 17:01 |
| pabelanger | those URLs didn't return anything for me | 17:03 |
| SotK | I suspect you may need to replace `search` in those urls with `story/list` for them to work currently | 17:03 |
| eggshell | yeah, those just go to the search page. | 17:03 |
| eggshell | no worries on my end though. good to know that tag is there. | 17:04 |
| SpamapS | SotK: weird | 17:05 |
| SpamapS | SotK: that seems like a common failing of storyboard-webclient's urls.. not being sharable | 17:06 |
| * SpamapS moves that to #storyboard | 17:06 | |
| jeblair | https://storyboard.openstack.org/#!/story/list?tags=low-hanging-fruit&tags=zuulv3 | 17:07 |
| jeblair | SpamapS: using the trick SotK just mentioned, we *can* make a sharable link to a search for two tags! | 17:07 |
| SpamapS | jeblair: look what you did there! sweet | 17:09 |
| SpamapS | I tried _so_ hard to read the API code | 17:09 |
| SpamapS | but I got lost in ORM hell and just gave up. | 17:09 |
| SpamapS | especially when I stumbled on use of FULLTEXT | 17:09 |
| SotK | jeblair: hm, the webclient should be fixed to let you do that without handcrafting a url too | 17:15 |
| rattboi | SpamapS: thanks for the link. I'll take a look. | 17:16 |
| *** isaacb has quit IRC | 17:25 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Zuul v3 readme updates https://review.openstack.org/442638 | 17:29 |
| jeblair | SpamapS, mordred, Shrews, pabelanger: ^ | 17:30 |
| SpamapS | hooray for docs | 17:32 |
| Shrews | i'd be curious to know how you figured out the multi-tag search thing | 17:54 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: WIP Add Directed Acylic Graph for usage as a JobGraph https://review.openstack.org/439663 | 18:00 |
| jeblair | Shrews: well, SotK said "replace search with story/list", and i recall from previous work on the api server that it generally honored multiple values for a key in a request, so i just gave that a shot. | 18:00 |
| Shrews | jeblair: oh, that info is in the scrollback I missed during lunch, too. | 18:02 |
| Shrews | but i like having you type things at me anyway | 18:03 |
| jeblair | Shrews: htnoeutnho.uhtneuhteuhte | 18:04 |
| Shrews | aww, how sweet | 18:05 |
| *** hashar is now known as hasharDinner | 18:06 | |
| mordred | Shrews: it seems he said it to you in dvoark too | 18:07 |
| Shrews | luckily I'm bi-keyboardal | 18:08 |
| jlk | jeblair: I see. Will have to account for that in the webapp patch then I think. | 18:11 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove layoutvalidator https://review.openstack.org/442684 | 18:59 |
| jlk | jeblair: is there a deeper doc on tenant stuff? The spec and existing docs on v3 branch don't really cover much. If a connection is global, and a hook comes in, is zuul supposed to search for the project across all tenants? I guess I just need to understand where boundaries are set for multi-tenant stuff | 19:08 |
| jeblair | jlk: nothing deeper than that. the events that a driver submits are actually not tenant-bound either; the scheduler should automatically apply a given event to each tenant. | 19:53 |
| jeblair | jlk: (Scheduler.process_event_queue) | 19:54 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Allow using webapp from connections https://review.openstack.org/439831 | 20:36 |
| *** hasharDinner is now known as hashar | 20:38 | |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Support GitHub PR webhooks https://review.openstack.org/439834 | 20:54 |
| jlk | alright I think I accounted for top level hooks | 20:54 |
| jlk | I think tests will pass on those. | 21:04 |
| jlk | jeblair: would be good to get some early input on those above reviews, the first salvos on a github driver in v3. Please stop me from continuing on with something overly stupid :) | 21:05 |
| jlk | hah, uh, minus the pep8 failures | 21:07 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Support GitHub PR webhooks https://review.openstack.org/439834 | 21:09 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove internal Swift support https://review.openstack.org/442789 | 21:10 |
| mordred | jeblair: ^^ zuulv3 caught a pep8 error (very very quickly) | 21:14 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove internal Swift support https://review.openstack.org/442789 | 21:15 |
| jeblair | mordred: nice :) | 21:15 |
| jlk | that darn zv3. SO FAST | 21:15 |
| jhesketh | Morning | 21:17 |
| jlk | jeblair: from that patch to remove swift, it means any zuul config with a [swift] section is going to error? If so, I'll have to rebase github patches :)D | 21:19 |
| jeblair | jlk: well, i don't think we fail on extra sections, but it would certainly be good to remove them if there are swift sections in new zuul.conf files. i doubt there will be structural changes affecting the github series, but some code is likely to be in proximity and require a rebase just for merge conflict resolution. | 21:22 |
| jeblair | jlk: there will certainly be more changes to zuul.conf before we're done too | 21:22 |
| jlk | sure | 21:22 |
| Shrews | jeblair: pabelanger: what is the "manager_name" recorded by nodepool into statsd? | 21:26 |
| jeblair | i'm working on triaging remaining skipped tests here: https://etherpad.openstack.org/p/zuulv3skips | 21:26 |
| jeblair | Shrews: zuul v2.5 launcher host -- replaces jenkins master. there is no v3 analog except for the requestor, which in our case will pretty much always be 'zuul'. | 21:27 |
| Shrews | jeblair: ok. do we just want to hardcode zuul? or not send that info at all? | 21:28 |
| jeblair | Shrews: let me look at its use real quick | 21:28 |
| Shrews | i'm removing "target" data since that seems irrelevant now | 21:29 |
| jeblair | Shrews: i think it may be useful mapping the requestor into some of the places the manager/target (those were also roughly equivalent) shows up. mostly so that if we end up in a position with nodepool servicing multiple clients (two different zuuls, or even if we start to develop more tools that use nodepool), we can categorize the stats. | 21:30 |
| jeblair | pabelanger: btw, did you see comments on https://review.openstack.org/429850 ? | 21:32 |
| Shrews | jeblair: so you're saying i'm going to add the Node.requestor support that you added in https://review.openstack.org/435949 ? :) | 21:32 |
| Shrews | going to have to add | 21:32 |
| jeblair | Shrews: sorry i thought you were asking about nodepool statsd | 21:33 |
| pabelanger | jeblair: I did, I've been meaning to loop back to it | 21:33 |
| Shrews | jeblair: i am. if you want to map th requestor, i need that requestor info | 21:33 |
| Shrews | my NodeRequest object (not Node, sorry...) doesn't have that | 21:34 |
| jeblair | Shrews: ah, yes. i should have added "which zuul is now sending back". sorry :) | 21:34 |
| Shrews | coolio | 21:34 |
| Shrews | jeblair: so use NodeRequest.requestor for manager_name, is what I came away with. Still don't see where we need target data | 21:35 |
| jeblair | Shrews: generally, yeah. manager ~= target. | 21:35 |
| Shrews | fun with words | 21:36 |
| jeblair | Shrews: you'll notice they are mutually exclusive. basically if you use nodepool 0 with jenkins, targets are interesting. if you use it with zuul v2.5, managers are interesting. | 21:36 |
| Shrews | jeblair: you're interesting | 21:36 |
| Shrews | in a good way :) | 21:37 |
| * jeblair blends in with the terminal background | 21:37 | |
| jeblair | also, that does pretty much sound like a southern curse. "Bless his heart; he's interesting isn't he?!" | 21:38 |
| Shrews | lol | 21:38 |
| jeblair | SpamapS: https://etherpad.openstack.org/p/zuulv3skips | 21:59 |
| jeblair | SpamapS: when i wrote "straightforward" i was thinking "assuming we haven't created new bugs, this test shouldn't require any work other than somewhat mechanically changing the test itself". | 22:01 |
| SpamapS | jeblair: that's a great roadmap | 22:07 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Zuul v3 readme updates https://review.openstack.org/442638 | 22:23 |
| *** hashar has quit IRC | 23:18 | |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Support GitHub PR webhooks https://review.openstack.org/439834 | 23:24 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove layoutvalidator https://review.openstack.org/442684 | 23:28 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove internal Swift support https://review.openstack.org/442789 | 23:28 |
| *** saneax-_-|AFK is now known as saneax | 23:29 | |
| jlk | okay NOW my tests pass | 23:46 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!