| *** rlandy has quit IRC | 00:20 | |
| *** jamesmcarthur has joined #zuul | 00:32 | |
| *** jamesmcarthur has quit IRC | 00:35 | |
| *** jamesmcarthur has joined #zuul | 00:35 | |
| *** tosky has quit IRC | 00:42 | |
| *** jamesmcarthur has quit IRC | 00:42 | |
| *** jamesmcarthur has joined #zuul | 00:44 | |
| *** ricolin has quit IRC | 00:46 | |
| *** ricolin has joined #zuul | 00:50 | |
| *** jamesmcarthur has quit IRC | 00:52 | |
| *** jamesmcarthur has joined #zuul | 00:59 | |
| *** jamesmcarthur has quit IRC | 01:30 | |
| *** jamesmcarthur has joined #zuul | 02:16 | |
| *** jamesmcarthur has quit IRC | 02:19 | |
| *** jamesmcarthur has joined #zuul | 02:21 | |
| *** jamesmcarthur has quit IRC | 02:31 | |
| *** vishalmanchanda has joined #zuul | 03:00 | |
| *** jamesmcarthur has joined #zuul | 03:04 | |
| *** jamesmcarthur has quit IRC | 03:08 | |
| *** jamesmcarthur has joined #zuul | 03:08 | |
| *** ykarel has joined #zuul | 03:17 | |
| *** jamesmcarthur has quit IRC | 04:06 | |
| *** jamesmcarthur has joined #zuul | 04:14 | |
| *** jamesmcarthur has quit IRC | 04:20 | |
| *** jamesmcarthur has joined #zuul | 04:20 | |
| *** jamesmcarthur has quit IRC | 04:50 | |
| *** jamesmcarthur has joined #zuul | 04:51 | |
| *** ikhan has quit IRC | 04:54 | |
| *** jamesmcarthur has quit IRC | 04:56 | |
| *** jamesmcarthur has joined #zuul | 05:04 | |
| *** wuchunyang has joined #zuul | 05:19 | |
| *** wuchunyang has quit IRC | 05:23 | |
| *** ajitha has joined #zuul | 05:26 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #zuul | 05:33 | |
| *** jamesmcarthur has quit IRC | 05:44 | |
| *** jamesmcarthur has joined #zuul | 05:45 | |
| *** jfoufas1 has joined #zuul | 05:45 | |
| *** jamesmcarthur has quit IRC | 05:52 | |
| *** saneax has joined #zuul | 05:53 | |
| *** jamesmcarthur has joined #zuul | 06:15 | |
| *** ykarel_ has joined #zuul | 06:36 | |
| *** ykarel has quit IRC | 06:38 | |
| *** piotrowskim has joined #zuul | 06:51 | |
| *** sshnaidm has quit IRC | 07:41 | |
| *** jamesmcarthur has quit IRC | 07:54 | |
| *** saneax has quit IRC | 07:58 | |
| *** sshnaidm has joined #zuul | 08:01 | |
| *** ykarel_ is now known as ykarel | 08:04 | |
| *** saneax has joined #zuul | 08:05 | |
| *** jamesmcarthur has joined #zuul | 08:09 | |
| *** rpittau|afk is now known as rpittau | 08:23 | |
| *** jamesmcarthur has quit IRC | 08:30 | |
| *** hashar has joined #zuul | 08:32 | |
| *** tosky has joined #zuul | 08:34 | |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul-jobs master: Fixes all tasks should be named rule https://review.opendev.org/c/zuul/zuul-jobs/+/779177 | 08:41 |
|---|---|---|
| *** jamesmcarthur has joined #zuul | 08:44 | |
| *** jamesmcarthur has quit IRC | 08:49 | |
| tobiash | tristanC, clarkb: could you add https://review.opendev.org/c/zuul/zuul/+/767084 to your review list? This is the last missing part of the cyclic dependencies stack. | 08:57 |
| *** jpena|off is now known as jpena | 08:58 | |
| *** jamesmcarthur has joined #zuul | 09:01 | |
| zbr | avass: tobiash tristanC: if you can help with #779177, it will help reducing the size of the ansible-lint v5 one. | 09:05 |
| avass | zbr: why does ansible-lint 5 force you to name tasks? | 09:05 |
| avass | zbr: already commented :) | 09:06 |
| zbr | avass: read https://github.com/ansible-community/ansible-lint/pull/1413 | 09:06 |
| *** jamesmcarthur has quit IRC | 09:07 | |
| zbr | naming is very important on reporting part, unnamed task can easily make hard to identify the source. there were some discussions on irc, nobody argued against removing exceptions, it went in. | 09:09 |
| avass | Imo there's not really much use in naming an 'include..' since you can easily find where that include is happening or the tasks inside it will have names (if relevant) | 09:10 |
| *** holser has joined #zuul | 09:11 | |
| zbr | if you have 10 includes,... remember that you could include the same role multiple times with different arguments. | 09:11 |
| zbr | is fun as 5 years ago, when I was working for another company, I do remember one or two folks that were finding the naming in general utterly useless. | 09:12 |
| avass | I do agree that naming is useful but it shouldn't be enforced everywhere | 09:13 |
| avass | or at least it should have more descriptive names than 'Include role ..' :) | 09:14 |
| zbr | maybe you could watch proposed changes? (you could have blocked it). | 09:14 |
| zbr | i will try to find better descriptions, but usually the original writer is supposed to give better explanations. | 09:15 |
| avass | agree | 09:15 |
| zbr | in fact i was surprised to find so few unnamed task in this repo, based on the number of tasks we have. | 09:16 |
| avass | yeah I think I blocked the notifications since they were flooding my github notifications :) | 09:16 |
| *** nils has joined #zuul | 09:16 | |
| zbr | my github notifications look like my email, and i only clean them every... quarter. | 09:16 |
| zbr | there is only one custom rule remaining in zuul-jobs, I wonder if someone could explain me better its purpose (same-owner). | 09:17 |
| avass | cleaning your inbox? people do that? | 09:17 |
| zbr | Ctrl-A, dev | 09:18 |
| zbr | del | 09:18 |
| avass | zbr: when pulling/pushing paths between executor and test node we can't keep the owner since the owner might not exist | 09:18 |
| avass | same with archive/unarchive | 09:20 |
| *** jamesmcarthur has joined #zuul | 09:20 | |
| zbr | is there a way we could rebrand this rule in order to include it inside the linter? I am looking for a way to ease maintenance. | 09:20 |
| avass | yeah I'm sursprised it's on by default tbh since I'd expect the users to differ more often than they match | 09:20 |
| avass | zbr: https://zuul-ci.org/docs/zuul-jobs/policy.html#preservation-of-owner-between-executor-and-remote | 09:21 |
| zbr | I guess the problem is not unique to zuul, it does affect any file transfer between two machines, so we could make a generic rule, disabled by default. | 09:23 |
| zbr | like I did with role var prefix. is not enabled until you define a prefix. | 09:23 |
| avass | yup I agree | 09:24 |
| avass | I'll take another look at 779177 and propose some better names later. I think I should have plenty of time this week. | 09:24 |
| zbr | i will try to do this and in parallel i will improve the names on ^. | 09:24 |
| *** jamesmcarthur has quit IRC | 09:26 | |
| *** jamesmcarthur has joined #zuul | 09:38 | |
| *** saneax has quit IRC | 09:40 | |
| *** saneax has joined #zuul | 09:42 | |
| *** saneax has quit IRC | 09:43 | |
| *** jamesmcarthur has quit IRC | 09:44 | |
| *** saneax has joined #zuul | 09:48 | |
| *** saneax has quit IRC | 09:52 | |
| *** jamesmcarthur has joined #zuul | 09:58 | |
| *** saneax has joined #zuul | 10:01 | |
| *** jamesmcarthur has quit IRC | 10:04 | |
| *** mugsie__ is now known as mugsie | 10:11 | |
| *** jamesmcarthur has joined #zuul | 10:15 | |
| *** jamesmcarthur has quit IRC | 10:22 | |
| *** jamesmcarthur has joined #zuul | 10:33 | |
| *** jamesmcarthur has quit IRC | 10:40 | |
| *** jangutter has quit IRC | 10:41 | |
| *** jangutter has joined #zuul | 10:42 | |
| *** jangutter has quit IRC | 10:43 | |
| *** jangutter has joined #zuul | 10:44 | |
| *** saneax has quit IRC | 10:53 | |
| *** saneax has joined #zuul | 10:53 | |
| *** jamesmcarthur has joined #zuul | 10:53 | |
| *** jamesmcarthur has quit IRC | 10:59 | |
| *** jangutter has quit IRC | 11:07 | |
| *** jangutter has joined #zuul | 11:07 | |
| *** jamesmcarthur has joined #zuul | 11:12 | |
| *** jamesmcarthur has quit IRC | 11:21 | |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul-jobs master: Fixes all tasks should be named rule https://review.opendev.org/c/zuul/zuul-jobs/+/779177 | 11:24 |
| *** jangutter has quit IRC | 11:29 | |
| *** jangutter has joined #zuul | 11:30 | |
| *** jamesmcarthur has joined #zuul | 11:35 | |
| *** hashar has quit IRC | 11:37 | |
| *** jangutter has quit IRC | 11:42 | |
| *** jamesmcarthur has quit IRC | 11:42 | |
| *** jangutter has joined #zuul | 11:42 | |
| *** ykarel has quit IRC | 11:52 | |
| *** ykarel has joined #zuul | 11:53 | |
| avass | tobiash, corvus: can you take another look at https://review.opendev.org/c/zuul/zuul/+/775382 (enabling ansible_shell_type in zuul)? | 12:06 |
| *** jamesmcarthur has joined #zuul | 12:11 | |
| *** ykarel_ has joined #zuul | 12:17 | |
| *** ykarel has quit IRC | 12:20 | |
| *** jpena is now known as jpena|lunch | 12:31 | |
| *** jamesmcarthur has quit IRC | 12:45 | |
| *** iurygregory has quit IRC | 12:46 | |
| tristanC | tobiash: left a nit comment on 767084, happy to approve if that's ok with you | 12:59 |
| *** jamesmcarthur has joined #zuul | 13:00 | |
| tobiash | tristanC: good catch, however gitlab doesn't use that variable so I think I can fix this in a followup :) | 13:01 |
| tristanC | tobiash: works for me | 13:02 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Fix default parameter in GitlabSource https://review.opendev.org/c/zuul/zuul/+/779255 | 13:04 |
| tobiash | tristanC, corvus: the followup ^ | 13:04 |
| tobiash | tristanC: thanks! | 13:05 |
| *** jamesmcarthur has quit IRC | 13:06 | |
| tristanC | tobiash: you're welcome, thank you for pushing that complex feature :-) | 13:07 |
| zbr | avass: do you want me to switch to the names suggested by you or those that I put in 2nd patch are ok? | 13:13 |
| *** hashar has joined #zuul | 13:13 | |
| *** jcapitao has joined #zuul | 13:20 | |
| *** ykarel_ is now known as ykarel | 13:23 | |
| *** ikhan has joined #zuul | 13:23 | |
| *** jpena|lunch is now known as jpena | 13:24 | |
| avass | zbr: I'm fine with that, however I usually try to not template in the 'name:' section since that usually makes the task harder to grep :) | 13:25 |
| zbr | avass: we think alike on that, maybe I should add a new rule for that and alienate few users that love templating name | 13:27 |
| zbr | especially as templating does not work in some cases, like loops. | 13:27 |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul-jobs master: Fixes all tasks should be named rule https://review.opendev.org/c/zuul/zuul-jobs/+/779177 | 13:30 |
| avass | zbr: I'll +3 that since it just changes the names | 13:32 |
| zbr | sure | 13:32 |
| *** sduthil has joined #zuul | 13:32 | |
| *** jangutter has quit IRC | 13:34 | |
| *** jangutter has joined #zuul | 13:34 | |
| *** jangutter has quit IRC | 13:44 | |
| *** jangutter has joined #zuul | 13:44 | |
| fungi | i'm starting to get the impression the "tasks must have names" rule is really a proxy for wanting tasks to have unique identifiers | 13:58 |
| fungi | can we add uuids to tasks? that would make them easy to find in a grep or codesearch | 13:58 |
| openstackgerrit | Merged zuul/zuul master: Allow refreshing volatile data in canMerge check https://review.opendev.org/c/zuul/zuul/+/767084 | 13:59 |
| *** mnaser has quit IRC | 14:03 | |
| *** mnaser has joined #zuul | 14:03 | |
| *** mnaser has quit IRC | 14:03 | |
| *** mnaser has joined #zuul | 14:03 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Fixes all tasks should be named rule https://review.opendev.org/c/zuul/zuul-jobs/+/779177 | 14:08 |
| openstackgerrit | Merged zuul/zuul master: Check cycle items are mergeable before reporting https://review.opendev.org/c/zuul/zuul/+/743450 | 14:13 |
| *** jangutter has quit IRC | 14:16 | |
| *** jangutter has joined #zuul | 14:17 | |
| avass | fungi: uuid would make that harder to grep in a repo and less 'guessable' | 14:25 |
| pabelanger | fungi: corvus: if you have time this morning: https://review.opendev.org/c/zuul/zuul-jobs/+/773474 is updated to revert changes to vars folder. using include_role multiple times with the different vars worked as expected | 14:29 |
| *** saneax has quit IRC | 14:29 | |
| avass | fungi: err I take that back, less guessable but not harder to grep. also it would be annoying if you were required to add a uuid to every task :) | 14:29 |
| avass | I suppose you could add uuids as tags on tasks though | 14:32 |
| corvus | swest, felixedel, tobiash: i've been looking at the "hashtag:sos" patches; i left some comments on them and proposed an alternate stack. it still needs a little polishing; i can finish that up today. but maybe you could take a look at those? | 14:35 |
| corvus | swest, felixedel, tobiash: my intent was to repropose the changes with different change ids without altering the original stack, though i messed that up on the last two. i did put them back the way they were though. | 14:36 |
| tobiash | corvus: regarding 777489 (zk tls in tests). I see the test-setup-docker creates a ca directory on demand. Should we add that to gitignore to avoid a dirty working copy? | 14:47 |
| corvus | tobiash: good diea | 14:47 |
| corvus | idea even | 14:47 |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul-jobs master: Bits to keep ansible-lint happy https://review.opendev.org/c/zuul/zuul-jobs/+/779276 | 14:47 |
| corvus | i also forgot to git add zoo.cfg, probably because it's right next to the dirty directory warning i was ignoring :) | 14:48 |
| tobiash | :) | 14:51 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Make reporting asynchronous https://review.opendev.org/c/zuul/zuul/+/691253 | 14:51 |
| corvus | tobiash: since you've been doing a lot of nodepool driver work, can you take a quick look at https://review.opendev.org/779071 and its child and give me your first impression on that? do you think that's going to lead to more efficiency? | 14:54 |
| corvus | tobiash: i guess i could probably use your benchmark script with a fake openstack setup, and then a fake state-machine setup? | 14:55 |
| tobiash | corvus: looking later today :) | 14:55 |
| corvus | kk :) | 14:55 |
| tobiash | I'll test it in int against a real cloud ;) | 14:55 |
| *** wuchunyang has joined #zuul | 14:57 | |
| corvus | tobiash: i think porting the openstack driver over will be complex and maybe the last part of this -- i'm thinking of starting with azure, then gce, then aws, then openstack. however, i do think azure is complex enough that it will exercise all the functionality we need for openstack. | 14:59 |
| tobiash | oh, ok, then not testing in int | 14:59 |
| *** wuchunyang has quit IRC | 15:02 | |
| *** jamesmcarthur has joined #zuul | 15:02 | |
| tobiash | corvus: so 779072 should already work as a fake state-machine setup? | 15:04 |
| corvus | tobiash: yep | 15:05 |
| tobiash | I think to be comparable the load testing would need to add ~30s latencies to wait-for-server and wait-for-delete calls | 15:05 |
| corvus | tobiash: should be doable, just memoize the state transition time in the state machine | 15:06 |
| corvus | tobiash: note that right now it suffers from the waterfall_sleep being 10 seconds, so can be a bit slow. i lowered that to 1 second in my testing. | 15:06 |
| tobiash | k | 15:06 |
| tobiash | corvus: one further optimization I could think about would be to distribute the key scanning to multiprocessing to offload that from the gil | 15:07 |
| tobiash | but I think the state machine would likely have more impact | 15:07 |
| corvus | tobiash: basically because it's the main provider request handling thread that now also drives the state machines. so i don't know if we want to just lower that sleep (maybe to 0, unless there are no state machines), or add a single new thread to drive the state machines. | 15:08 |
| *** jamesmcarthur has quit IRC | 15:08 | |
| *** hashar has quit IRC | 15:14 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Use ZooKeeper TLS in tests https://review.opendev.org/c/zuul/zuul/+/777489 | 15:17 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Move setupZK() helper function to BaseTestCase class https://review.opendev.org/c/zuul/zuul/+/779082 | 15:17 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Initialize ZooKeeper connection in server rather than in cmd classes https://review.opendev.org/c/zuul/zuul/+/779083 | 15:17 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Make ConnectionRegistry mandatory for Scheduler https://review.opendev.org/c/zuul/zuul/+/779086 | 15:17 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Instantiate executor client, merger, nodepool and app within Scheduler https://review.opendev.org/c/zuul/zuul/+/779087 | 15:17 |
| corvus | okay, i've resolved all known issues in that stack | 15:18 |
| dmsimard | btw tomorrow is ansible contributor summit: https://www.eventbrite.com/e/ansible-contributor-summit-202103-registration-141735886853 && https://hackmd.io/@ansible-community/contrib-summit-202103 | 15:26 |
| fungi | so brace for new announcements from red hat to the ansible contributors/community about what to expect for the coming year | 15:30 |
| dmsimard | fungi: I get the snark but there is a now a community team (that I am a part of) and we're trying in good faith to be better about that and switch things up | 15:43 |
| dmsimard | we have meetings every wednesdays in #ansible-community and everyone is invited to participate | 15:44 |
| fungi | dmsimard: oh, awesome. the only acs i've attended, the announcements were entirely about decisions made by red hat staff behind closed doors months earlier | 15:44 |
| * dmsimard nods | 15:45 | |
| fungi | good to know they're getting the community more involved in their decision-making process | 15:45 |
| *** jamesmcarthur has joined #zuul | 15:45 | |
| dmsimard | I remember learning about collections about the same time as you did back in austin (or atlanta?) | 15:45 |
| fungi | i think it was atlanta, yeah | 15:46 |
| fungi | anyway, i'm glad to hear that the sort of open design concept we've been promoting for ages is starting to take root in more otherwise open communities | 15:46 |
| dmsimard | I don't expect everything to be out in the open (yet?) but the motivation to do more in the open is there at least from a community perspective | 15:47 |
| fungi | it's certainly a challenge to balance that with product marketing teams who want things to be kept secret so they can make a big media splash about them, i get that, but i suppose it's a question of priorities | 15:48 |
| dmsimard | yeah, product gotta productize and sell I suppose :/ | 15:48 |
| dmsimard | we'll be standing up a steering community and some non red hat folks will have a seat | 15:49 |
| dmsimard | I've "grown up" with things like the openstack TC and board already set up with great and knowledgeable people running them | 15:50 |
| *** jangutter has quit IRC | 15:50 | |
| dmsimard | it's another matter to try and retrofit where it didn't exist before but again, the motivation is there | 15:51 |
| *** jangutter has joined #zuul | 15:51 | |
| fungi | that's an excellent start, for sure | 15:51 |
| *** gmann is now known as gmann_afk | 16:01 | |
| *** ajitha has quit IRC | 16:15 | |
| *** ajitha has joined #zuul | 16:41 | |
| *** jamesmcarthur has quit IRC | 16:50 | |
| *** jamesmcarthur has joined #zuul | 16:52 | |
| *** ikhan has quit IRC | 16:58 | |
| *** ikhan has joined #zuul | 16:59 | |
| *** ykarel has quit IRC | 17:02 | |
| zbr | there are few interesting bits on the summit agenda | 17:05 |
| *** jpena is now known as jpena|brb | 17:15 | |
| *** iurygregory has joined #zuul | 17:17 | |
| *** jcapitao has quit IRC | 17:17 | |
| corvus | tobiash: should we move 'validate-tenants' from the zuul client to zuul-scheduler? 1) we're looking at removing the "zuul" client in favor of zuul-client; 2) it's really a scheduler operation (it's implemented as a scheduler that does no work; it puts jobs out to mergers like a scheduler does, and otherwise has all of the system requirements of a scheduler) | 17:22 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Make ConnectionRegistry mandatory for Scheduler https://review.opendev.org/c/zuul/zuul/+/779086 | 17:23 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Instantiate executor client, merger, nodepool and app within Scheduler https://review.opendev.org/c/zuul/zuul/+/779087 | 17:23 |
| *** jangutter has quit IRC | 17:26 | |
| *** jangutter has joined #zuul | 17:26 | |
| *** hamalq has joined #zuul | 17:29 | |
| tobiash | corvus: ++ | 17:32 |
| *** jpena|brb is now known as jpena | 17:37 | |
| *** iurygregory has quit IRC | 17:49 | |
| *** rpittau is now known as rpittau|afk | 18:16 | |
| *** jfoufas1 has quit IRC | 18:26 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Instantiate executor client, merger, nodepool and app within Scheduler https://review.opendev.org/c/zuul/zuul/+/779087 | 18:29 |
| *** jpena is now known as jpena|off | 18:30 | |
| *** iurygregory has joined #zuul | 18:46 | |
| *** jamesmcarthur has quit IRC | 19:15 | |
| *** jamesmcarthur has joined #zuul | 19:15 | |
| *** jamesmcarthur has quit IRC | 19:20 | |
| *** jamesmcarthur has joined #zuul | 19:20 | |
| *** gmann_afk is now known as gmann | 19:32 | |
| *** jamesmcarthur has quit IRC | 19:48 | |
| *** jamesmcarthur has joined #zuul | 19:49 | |
| *** jamesmcarthur has quit IRC | 19:54 | |
| *** jamesmcarthur has joined #zuul | 19:55 | |
| *** hashar has joined #zuul | 20:03 | |
| Phoenikzz | It would really be appreciated if someone had a look at https://review.opendev.org/c/zuul/zuul/+/775382 | 20:10 |
| *** piotrowskim has quit IRC | 20:11 | |
| fungi | Phoenikzz: i was previously unfamiliar with ansible_shell_type, but if i'm following the various issue threads correctly it breaks become until the become plugins work in ansible 2.11.0b1 and later, is that going to be a problem? | 20:18 |
| fungi | avass: i suppose i should be asking you ^ | 20:19 |
| avass | fungi: I'm not sure I follow that question :) | 20:20 |
| clarkb | fungi: I suspect it is related to https://docs.ansible.com/ansible/latest/user_guide/windows_setup.html#configuring-ansible-for-ssh-on-windows | 20:20 |
| avass | anyway we only need to set that for windows machines currently | 20:20 |
| fungi | long-standing ansible issue since ~2015 about using ansible_shell_type and become together | 20:20 |
| avass | at the moment we're managing that for host-vars but it would be good to have that in the nodepool config | 20:20 |
| avass | ah | 20:21 |
| fungi | apparently become is broken unless you use bourne shell as your shell | 20:21 |
| avass | no, not a problem so long as you don't sett ansible_shell_type | 20:21 |
| avass | and we only set it to cmd to use ssh on windows | 20:21 |
| fungi | they've completely reworked become into a plugin architecture in ansible 2.11, which in theory solves some lingering problems with it | 20:21 |
| avass | the implementation shouldn't be changing zuul's defaults so it's really up to the user. | 20:22 |
| avass | clarkb: yep | 20:23 |
| fungi | avass: a more detailed commit message explaining why this is needed/what it fixes might help | 20:23 |
| fungi | i'm left trying to guess that in a bit of a vacuum | 20:23 |
| avass | Phoenikzz: ^ | 20:24 |
| fungi | is it "this allows us to not have to separately specify ansible_shell_type in the environment on all our windows nodes" maybe? | 20:24 |
| avass | wyeah it allows us to specify that in the connection configuration in nodepool instead of setting ssh connection in nodepool and needing to set 'ansible_shell_type: cmd' as a host var for each job using those nodes | 20:26 |
| avass | because if ansible_connection is set to ssh ansible defaults to formatting it's commands as if it was a bourne shell | 20:26 |
| fungi | makes sense to me then. if something like that can find its way into the commit message, it would probably help reviewers | 20:27 |
| openstackgerrit | Daniel Blixt proposed zuul/zuul master: Use shell-type config from nodepool https://review.opendev.org/c/zuul/zuul/+/775382 | 20:29 |
| fungi | thanks Phoenikzz! | 20:29 |
| Phoenikzz | Thanks for taking a look fungi! | 20:30 |
| fungi | my pleasure | 20:30 |
| avass | maybe we should add releasenotes for that and/or nodepool | 20:35 |
| fungi | or even documentation? | 20:37 |
| fungi | at the moment it's a bit of magic to infer that | 20:37 |
| corvus | pretty sure it's documented in nodepool | 20:38 |
| fungi | yeah, what's not documented is that zuul will now infer it from nodepool | 20:38 |
| corvus | i'm not sure that's necessary? it's a nodepool setting | 20:38 |
| fungi | though if we have a pattern of doing that with other configured values, maybe users will assume iot | 20:38 |
| fungi | it | 20:39 |
| corvus | yeah, it's not infering it so much as honoring it | 20:39 |
| avass | yeah there's documentation for it in nodepool. I'll add a releasenote for it there tomorrow if i remember :) | 20:39 |
| fungi | in that case, sounds like maybe a release note is still in order to let users who were previously working around it know they can clean up their workarounds | 20:39 |
| corvus | ++ | 20:40 |
| avass | so something like a releasenote in notepool with 'Hey this thing is now configurable' and one in zuul saying 'zuul will now use this value when it's supplied by nodepool' | 20:41 |
| fungi | yeah, if the nodepool piece was only added since the last release, then a release note in there could also help bring it to the attention of users as a new feature | 20:43 |
| clarkb | corvus: for https://review.opendev.org/c/zuul/zuul/+/777489 is the tools/docker-compose.sh not used in CI? The tox jobs should all use ensure-zookeeper and the functional jobs deploy zk themsevles too iirc | 20:44 |
| clarkb | corvus: I've +2'd that change but left a note you may want to check on (I also didn't approve it since I'm not sure if it needs coordination with other repos due to the job changes) | 20:45 |
| corvus | clarkb: they all use ensure-zookeeper now, so that's just for humans | 20:47 |
| *** vishalmanchanda has quit IRC | 20:47 | |
| corvus | clarkb: other job change is here and passing: https://review.opendev.org/779089 | 20:47 |
| corvus | clarkb: so i think it's gtg, but let's give tobiash/swest/felix a chance to chime in on that stack before +w | 20:48 |
| corvus | (i mean, i'm still pretty confident that change is desirable in isolation, but just in case we're missing some interaction they've thought of) | 20:49 |
| clarkb | ok | 20:49 |
| clarkb | corvus: also left a question on https://review.opendev.org/c/zuul/zuul/+/779082 | 20:51 |
| corvus | clarkb: i think it's just defined there, not invoked | 20:52 |
| clarkb | oh I see ya you are right | 20:53 |
| clarkb | you still need to call the setup function | 20:53 |
| corvus | ya | 20:53 |
| *** arxcruz is now known as arxcruz|off | 21:04 | |
| *** ikhan has quit IRC | 21:12 | |
| clarkb | corvus: I left a note/question on https://review.opendev.org/c/zuul/zuul/+/779086 and now I see zuul -1'd it. I wonder if those two things are related | 21:15 |
| openstackgerrit | Daniel Blixt proposed zuul/zuul master: Use shell-type config from nodepool https://review.opendev.org/c/zuul/zuul/+/775382 | 21:16 |
| clarkb | fungi: ^ that adds a release note to explain it should only be used for windows hosts | 21:22 |
| fungi | ahh, cool | 21:23 |
| clarkb | that change also conflicts with corvus' stack that I was reviewing | 21:23 |
| clarkb | may want to do some coordinated updates here to avoid unexpected fallout | 21:23 |
| *** ikhan has joined #zuul | 21:29 | |
| openstackgerrit | Daniel Blixt proposed zuul/nodepool master: Update docs for shell-type driver config https://review.opendev.org/c/zuul/nodepool/+/779370 | 21:39 |
| Phoenikzz | ^ In response to the discussion, I updated docs for shell-type on nodepool and add a release note there too | 21:41 |
| clarkb | Phoenikzz: thanks I saw the previous change. My only other note is it appears to conflictwith corvus' large stack so we may want to coordinate those | 21:41 |
| *** jamesmcarthur has quit IRC | 21:41 | |
| corvus | if shell is ready to go, let's do it first | 21:45 |
| corvus | should be easy to rebase the others on that | 21:45 |
| clarkb | I think the zuul side is ready I gave it a +2 | 21:46 |
| clarkb | and nodepool side is just docs updates | 21:46 |
| *** jamesmcarthur has joined #zuul | 21:48 | |
| fungi | yeah, i just now +2'd that | 21:57 |
| *** jamesmcarthur_ has joined #zuul | 22:08 | |
| *** jamesmcarthur has quit IRC | 22:12 | |
| *** hashar has quit IRC | 22:29 | |
| *** ajitha has quit IRC | 22:35 | |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Azure: switch to Azul https://review.opendev.org/c/zuul/nodepool/+/778814 | 22:37 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add a state machine driver framework https://review.opendev.org/c/zuul/nodepool/+/779071 | 22:37 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add an example state machine driver https://review.opendev.org/c/zuul/nodepool/+/779072 | 22:37 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Improve state machine keyscan handling https://review.opendev.org/c/zuul/nodepool/+/779379 | 22:37 |
| corvus | tobiash: ^ after your comments earlier, i realized key scanning in the state manager driver was serialized, so that uses your idea of an executor pool to parallelize it. | 22:38 |
| corvus | basically we want to serialize cloud api interaction (rate limits!) and paralellize everything else | 22:39 |
| fungi | "azul" is an awesome bit of spanish wordplay, and also the color of the zuul logo! | 22:41 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Update docs for shell-type driver config https://review.opendev.org/c/zuul/nodepool/+/779370 | 22:43 |
| corvus | fungi: thank you for noticing, i'm really proud of that one :) | 22:43 |
| fungi | and it's the name of a rather fun tile game. highly recommend | 22:45 |
| fungi | though now i can't stop thinking about the harvey birdman attorney at law episodes where blue falcon appears as harvey's arch rival, azul falconé | 22:54 |
| *** nils has quit IRC | 22:54 | |
| mordred | corvus: stray print in the azul patch | 22:55 |
| corvus | derp | 22:56 |
| fungi | ahh, yep, i scanned right past print(disk) | 22:58 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Azure: switch to Azul https://review.opendev.org/c/zuul/nodepool/+/778814 | 22:58 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add a state machine driver framework https://review.opendev.org/c/zuul/nodepool/+/779071 | 22:58 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add an example state machine driver https://review.opendev.org/c/zuul/nodepool/+/779072 | 22:58 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Improve state machine keyscan handling https://review.opendev.org/c/zuul/nodepool/+/779379 | 22:58 |
| corvus | clarkb: replied on 779086 | 23:04 |
| corvus | i do not understand why that test is failing though; it passes locally | 23:04 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: DNM: debug https://review.opendev.org/c/zuul/zuul/+/779387 | 23:18 |
| clarkb | corvus: can you check my response on that one. | 23:27 |
| corvus | clarkb: yep, gotcha | 23:30 |
| corvus | sorry i misunderstood | 23:30 |
| clarkb | no worries | 23:31 |
| clarkb | one thing that gerrit doesn't make easy is talking about code that is unchagned in a change within a change's context | 23:31 |
| openstackgerrit | Merged zuul/nodepool master: Update docs for shell-type driver config https://review.opendev.org/c/zuul/nodepool/+/779370 | 23:32 |
| *** jamesmcarthur_ has quit IRC | 23:36 | |
| clarkb | mugsie: ^ not sure if you've seen 778814, but that sort of feedback may be useful to yall at azure? | 23:53 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!