| *** armstrongs has joined #zuul | 00:28 | |
| *** tosky has quit IRC | 00:38 | |
| *** armstrongs has quit IRC | 00:38 | |
| *** rf0lc0 has quit IRC | 00:50 | |
| *** wxy-xiyuan has joined #zuul | 00:54 | |
| *** jamesmcarthur has joined #zuul | 01:12 | |
| *** rlandy has quit IRC | 01:17 | |
| *** jamesmcarthur has quit IRC | 01:23 | |
| *** sgw has quit IRC | 01:26 | |
| *** jamesmcarthur has joined #zuul | 01:37 | |
| *** jamesmcarthur has quit IRC | 02:38 | |
| *** jamesmcarthur has joined #zuul | 02:53 | |
| *** bhavikdbavishi has joined #zuul | 03:13 | |
| *** bhavikdbavishi1 has joined #zuul | 03:16 | |
| *** bhavikdbavishi has quit IRC | 03:17 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:17 | |
| *** sgw has joined #zuul | 03:21 | |
| *** jamesmcarthur has quit IRC | 03:55 | |
| *** evrardjp has quit IRC | 05:34 | |
| *** evrardjp has joined #zuul | 05:34 | |
| *** jamesmcarthur has joined #zuul | 05:57 | |
| *** jamesmcarthur has quit IRC | 06:01 | |
| *** sanjayu_ has joined #zuul | 06:17 | |
| *** bhavikdbavishi has quit IRC | 07:08 | |
| *** yoctozepto has quit IRC | 07:28 | |
| *** raukadah is now known as chkumar|rover | 07:28 | |
| *** yoctozepto has joined #zuul | 07:29 | |
| *** yoctozepto has quit IRC | 07:33 | |
| *** bolg has joined #zuul | 07:45 | |
| *** ianychoi has joined #zuul | 07:46 | |
| *** Defolos has joined #zuul | 07:51 | |
| *** yoctozepto has joined #zuul | 07:59 | |
| *** tosky has joined #zuul | 08:22 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Support pausing merge jobs https://review.opendev.org/707192 | 08:28 |
|---|---|---|
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: ensure-tox: include tox-venv plugin https://review.opendev.org/707237 | 08:29 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: WIP: ensure-tox to verify module calling https://review.opendev.org/707335 | 08:42 |
| *** felixedel has joined #zuul | 08:42 | |
| tobiash | corvus: when you have time, it would be awesome to have a new gear release now the mac support patches are merged | 08:44 |
| *** jpena|off is now known as jpena | 08:50 | |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: Authorization rules: add templating https://review.opendev.org/705193 | 08:52 |
| *** felixedel has quit IRC | 10:13 | |
| *** bhavikdbavishi has joined #zuul | 10:29 | |
| *** bhavikdbavishi1 has joined #zuul | 10:32 | |
| *** bhavikdbavishi has quit IRC | 10:34 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 10:34 | |
| *** sshnaidm_ has joined #zuul | 10:49 | |
| *** sshnaidm has quit IRC | 10:49 | |
| *** sshnaidm_ is now known as sshnaidm | 10:49 | |
| *** zxiiro has quit IRC | 11:13 | |
| *** avass has joined #zuul | 11:19 | |
| *** wxy-xiyuan has quit IRC | 11:34 | |
| *** mhu has quit IRC | 11:54 | |
| *** rfolco has joined #zuul | 12:10 | |
| *** ysastri has joined #zuul | 12:14 | |
| *** sshnaidm has quit IRC | 12:20 | |
| *** ianychoi has quit IRC | 12:25 | |
| *** rfolco has quit IRC | 12:26 | |
| *** rfolco has joined #zuul | 12:26 | |
| *** sshnaidm has joined #zuul | 12:27 | |
| *** sshnaidm has quit IRC | 12:31 | |
| *** bhavikdbavishi has quit IRC | 12:32 | |
| *** sshnaidm has joined #zuul | 12:41 | |
| *** jpena is now known as jpena|lunch | 12:49 | |
| *** rlandy has joined #zuul | 12:57 | |
| *** sshnaidm has quit IRC | 13:02 | |
| *** bolg has quit IRC | 13:06 | |
| *** bolg_ has joined #zuul | 13:06 | |
| *** bolg_ has quit IRC | 13:09 | |
| *** jamesmcarthur has joined #zuul | 13:09 | |
| *** sshnaidm has joined #zuul | 13:17 | |
| *** sshnaidm has quit IRC | 13:18 | |
| *** sshnaidm has joined #zuul | 13:18 | |
| *** sshnaidm has quit IRC | 13:28 | |
| *** sshnaidm has joined #zuul | 13:29 | |
| *** jpena|lunch is now known as jpena | 13:33 | |
| *** jamesmcarthur has quit IRC | 13:35 | |
| *** bhavikdbavishi has joined #zuul | 13:39 | |
| *** jamesmcarthur has joined #zuul | 13:45 | |
| *** jamesmcarthur has quit IRC | 14:11 | |
| *** jamesmcarthur has joined #zuul | 14:12 | |
| *** bhavikdbavishi has quit IRC | 14:12 | |
| *** jamesmcarthur has quit IRC | 14:17 | |
| *** sgw has quit IRC | 14:34 | |
| *** jamesmcarthur has joined #zuul | 14:42 | |
| *** jamesmcarthur has quit IRC | 14:47 | |
| *** swest has quit IRC | 14:50 | |
| *** swest has joined #zuul | 14:51 | |
| *** swest has quit IRC | 14:51 | |
| *** jamesmcarthur has joined #zuul | 15:00 | |
| *** chkumar|rover is now known as raukadah | 15:02 | |
| *** sgw has joined #zuul | 15:04 | |
| *** jamesmcarthur has quit IRC | 15:16 | |
| *** jamesmcarthur has joined #zuul | 15:31 | |
| *** vivobg has joined #zuul | 15:33 | |
| *** ysastri has quit IRC | 15:33 | |
| *** jamesmcarthur_ has joined #zuul | 16:03 | |
| *** jamesmcarthur has quit IRC | 16:06 | |
| vivobg | Hi all. I wanted to ask how is the Bitbucket and Gitlab integration support coming along? I found the following https://review.opendev.org/#/c/657837/ and https://review.opendev.org/#/q/status:open+project:zuul/zuul+branch:master+topic:gitlab | 16:12 |
| *** zxiiro has joined #zuul | 16:24 | |
| *** fbo has joined #zuul | 16:29 | |
| fungi | vivobg: i gather the gitlab one is basically functional? a few folks have mentioned testing it out themselves, we probably just need to finish reviewing that but fbo can probably confirm | 16:32 |
| fungi | vivobg: as for the bitbucket driver, last i spoke to osfos (at fosdem last week) he was trying to nail down some random api disconnects he's seeing between the scheduler and the bitbucket service | 16:33 |
| fungi | sounds like it's basically working but he was hesitant to call it ready without a better grip on (and hopefully fix for) the connectivity issue he's wrestling with | 16:34 |
| clarkb | fungi: vivobg I expect both drivers would be happy to have more testers though if you are willing to give it a spin | 16:35 |
| fungi | er, s/osfos/ofosos/ | 16:35 |
| *** sgw has quit IRC | 16:36 | |
| fbo | fungi: yes the gitlab driver is functional to be used in a check style pipeline. for gating and approval events it will need further developement | 16:36 |
| vivobg | Thank you for the update. Is there somewhere else we can track the progress, other than the links above ? | 16:37 |
| fungi | vivobg: when those changes merge the zuul release notes for the next release will mention them, as will the service documentation | 16:37 |
| fungi | you can watch the zuul-announce mailing list for release announcements | 16:38 |
| vivobg | awesome, thank you | 16:38 |
| fungi | but generally, detailed discussion about source code changes like these happens in our code review system, so the links you already referenced | 16:38 |
| *** Defolos has quit IRC | 16:39 | |
| fungi | fbo: so if the gitlab driver lacks sufficient feature parity with the other connection drivers to support gating, do you think we should mark it as "experimental" or wait to merge it until it gains that support? | 16:40 |
| fbo | vivobg: I don't have the channel history but here https://storyboard.openstack.org/#!/story/2006632 you'll the get accurate state about the gitlab driver effort | 16:44 |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Simplify virtualenv install and execution https://review.opendev.org/707428 | 16:45 |
| clarkb | zuulians ^ that is a cleanup based on virtualenv updates from yesterday. Not urgent but helps us avoid using non default behavior | 16:45 |
| vivobg | fbo: thanks | 16:46 |
| fbo | fungi: well I don't know both way are valid I think but it depends how zuul cores want to handle this kind of changes. I mean if we merge waht we have today people could start to use the driver for simple use case. The documentation is accurate about what is supported in the driver atm. But adding an experimental flag could be nice yes | 16:47 |
| mordred | fbo: honestly, I think landing it if it's solid for check will allow people to start interacting with it which could then help provide justificaiton for the followup work | 16:50 |
| mordred | but that's just me - other folks may have different views | 16:51 |
| mordred | I do agree that marking it clearly as not being at feature parity woudl be good - but that might be a nice documentation matrix to have anyway. for instance, gerrit currently does inline comments where github doesn't | 16:52 |
| clarkb | it might be good to write up a general todo list or feature matrix type of thing | 16:52 |
| mordred | clarkb: feature matrix jinx :) | 16:52 |
| clarkb | nodepool drivers could use similar as well | 16:52 |
| mordred | yah | 16:52 |
| clarkb | then people aren't surprised when some combo doesn't do everything they want. Or they can look and decide they want to add that functionality | 16:52 |
| mordred | ++ | 16:53 |
| *** sgw has joined #zuul | 16:53 | |
| *** maxamillion[m] has joined #zuul | 16:55 | |
| tristanC | mordred: could we please apply the same strategy for the zuul-operator, it would be quite nice to have a zuul-operator image ready to use | 16:56 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add an ensure-tox test job https://review.opendev.org/706371 | 17:02 |
| corvus | i think we can land the operator changes -- i think we just wanted to leave them out there a bit for everyone who is interested to review them | 17:07 |
| fungi | i do like the feature parity matrix idea both for zuul connection drivers and nodepool provider drivers, though i don't know if i have the bandwidth to put those together | 17:08 |
| corvus | fbo, mordred: what's blocking gating in gitlab? | 17:09 |
| vivobg | clarkb: We are quite keen on both. How do we get started on testing gitlab and bitbucket integration? Are there particular branches we should use, docs etc ? | 17:11 |
| clarkb | vivobg: the changes you've linked represent the "branches". The way gerrit review works is you propose things against HEAD and they land when ready. But you can pull them as is and run them. (rather than pulling a specific branch) | 17:12 |
| clarkb | vivobg: in the top right should be a download dropdown and from that you'll get git commands for pulling the refs | 17:13 |
| clarkb | (top right of the gerrit change UI I mean) | 17:13 |
| fbo | corvus: implementing gating with gitlab seems technically feasible according to the quick study i did before starting to write the driver. Basically it's just time for the moment, but I'll be able to resume my work on it soon. | 17:14 |
| clarkb | vivobg: documentation for the drivers themselves will hopefully be in the chagnes. if not that would be good feedback. If looking for documentation on running from source, that might be lacking | 17:16 |
| *** igordc has joined #zuul | 17:17 | |
| vivobg | Thanks. I am quite excited to try them | 17:18 |
| clarkb | for the deploying from source bit, I think we'd be happy to help and maybe we can produce a bit of documentation out of your experience? | 17:19 |
| clarkb | I think the zuul from scratch bits will cover everything but pulling the gerrit changes | 17:20 |
| * clarkb looks | 17:20 | |
| clarkb | vivobg: yup I think https://zuul-ci.org/docs/zuul/howtos/zuul-from-scratch.html might be a good place to start. Then after you git clone zuul you'd run the gerrit supplied command to apply the proposed changes to the repo | 17:20 |
| clarkb | then the install steps will install those changes you've pulled down | 17:21 |
| vivobg | I will give that a go, thanks | 17:21 |
| *** Tahvok has quit IRC | 17:26 | |
| *** Tahvok has joined #zuul | 17:27 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - bootstrap the driver structure + Webhook support https://review.opendev.org/685682 | 17:33 |
| *** evrardjp has quit IRC | 17:34 | |
| *** evrardjp has joined #zuul | 17:34 | |
| openstackgerrit | James E. Blair proposed zuul/gcp-authdaemon master: Initial commit https://review.opendev.org/707438 | 17:41 |
| corvus | clarkb, tobiash: does this look good for a release? commit 9360224e0f3e29251d4bc77ffece6cf1c20a3a27 (HEAD -> master, tag: 3.16.1, origin/master, origin/HEAD) | 17:42 |
| clarkb | looking | 17:43 |
| clarkb | corvus: that commit looks correct to me. Not sure if you want to get https://review.opendev.org/707428 in too as that will allow old virtualenv to work too. Actually I should update that change to make this mroe explicitly | 17:44 |
| clarkb | *explicit | 17:44 |
| corvus | the tox remote jobs have been failing more recently, i don't know why | 17:45 |
| corvus | clarkb: you're going to change the req to a "!=" ? that sounds like a good idea | 17:46 |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Simplify virtualenv install and execution https://review.opendev.org/707428 | 17:47 |
| clarkb | corvus: yup ^ | 17:47 |
| *** jamesmcarthur_ has quit IRC | 17:47 | |
| clarkb | now old and new virtualenv will work | 17:47 |
| clarkb | just not the known bad versions | 17:47 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: WIP: Make ensure-tox pass cross-platform https://review.opendev.org/707439 | 17:48 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add tox env for update-test-platforms https://review.opendev.org/706404 | 17:55 |
| corvus | mordred, tristanC: want to go ahead and +3 (or not) https://review.opendev.org/707428 ? | 17:59 |
| tristanC | corvus: that seems correct, i +2, but haven't had time to check it works as expected to +3 | 18:01 |
| tristanC | i mean, i haven't looked what happen with virtualenv and why it stopped working | 18:02 |
| mordred | tristanC: it's a giant mess and you probably don't want to know more about it :) | 18:05 |
| corvus | tristanC: thanks -- that it doesn't look like it will mess up SF.io is a helpful review :) i'll leave it to mordred to +3 | 18:06 |
| mordred | tristanC: tl;dr - they introduced a new behavior where they cached some important basic virtualenv contents like pip and pkg_resources, put them in a folder in ~/.local and then symlinked those into the virtualenvs instead of copuying. this made things like root-creates-venv-but-someone-else-uses (and a host of other things) break ... they understood the breakages and rolled out 20.0.2 as an update which still | 18:07 |
| mordred | puts things in ~/.local but copies them on venv creation | 18:07 |
| *** igordc has quit IRC | 18:08 | |
| tristanC | oh well, at least they used a major version number bump to introduce the feature^Wbug . thank you for the summary | 18:09 |
| *** jpena is now known as jpena|off | 18:11 | |
| *** vivobg has quit IRC | 18:13 | |
| webknjaz | Also you can fix the dependency resolver bug by upgrading six | 18:16 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul master: Flesh out the glossary significantly https://review.opendev.org/704391 | 18:16 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: ensure-tox: include tox-venv plugin https://review.opendev.org/707237 | 18:17 |
| *** jamesmcarthur has joined #zuul | 18:18 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: bindep: use venv when possible https://review.opendev.org/707078 | 18:23 |
| *** igordc has joined #zuul | 18:23 | |
| *** jamesmcarthur has quit IRC | 18:31 | |
| *** jamesmcarthur has joined #zuul | 18:31 | |
| tobiash | Oh did that issue get resolved upstream? | 18:41 |
| tobiash | (re 707428) | 18:41 |
| *** jamesmcarthur has quit IRC | 18:42 | |
| *** AJaeger_ is now known as AJaeger | 18:42 | |
| *** jamesmcarthur has joined #zuul | 18:43 | |
| clarkb | tobiash: yes 20.0.2 stopped symlinking by default | 18:45 |
| tobiash | :) | 18:45 |
| clarkb | tobiash: they still copy the files from ~/.local but they are copied so running in bwrap or whatever is fine | 18:45 |
| tobiash | So we're waiting for 707428 before releasing? | 18:45 |
| clarkb | I think it might be a nice thing to do so that old virtualenv users work again | 18:46 |
| clarkb | but its probably not strictly necessary if the github fix is super urgent | 18:46 |
| tobiash | ++ | 18:46 |
| *** tosky has quit IRC | 18:47 | |
| *** jamesmcarthur has quit IRC | 18:48 | |
| tobiash | with latest gear master and a fake bwrap I got most test cases running on MacOS natively now :) | 18:48 |
| tobiash | that'll make developing for me much easier | 18:49 |
| *** jamesmcarthur has joined #zuul | 18:54 | |
| corvus | mordred: I hit the submit button on https://gerrit-review.googlesource.com/c/zuul/ops/+/254715 and then started an email to the repo-discuss list, and before i was even done, look what showed up in the labels tab: https://gerrit-zuul.inaugust.com/t/gerrit/labels | 18:57 |
| corvus | we're going to be giddy when we get this going for opendev :) | 18:57 |
| mordred | SWEET | 18:57 |
| *** jamesmcarthur has quit IRC | 18:57 | |
| mordred | corvus: yah. very much | 18:57 |
| *** jamesmcarthur has joined #zuul | 18:58 | |
| corvus | mordred: i'm going to start some bazel stuff in gerrit's zuul, but with the idea that maybe we move it into zuul-jobs when it's working | 19:01 |
| corvus | i'm thinking "install-bazelisk" is maybe a generally useful role | 19:01 |
| tobiash | corvus: is there already something bazel related in review? | 19:03 |
| tobiash | I thought I remember something in that direction | 19:04 |
| clarkb | tobiash: ya there is | 19:04 |
| corvus | https://review.opendev.org/693513 | 19:04 |
| tobiash | ah cool | 19:05 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - bootstrap the driver structure + Webhook support https://review.opendev.org/685682 | 19:05 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - Basic handling of merge_requests event https://review.opendev.org/685990 | 19:05 |
| corvus | i think i'm going to be starting with bazelisk, which may be orthogonal to that... | 19:06 |
| fungi | just don't look it in the eyes | 19:06 |
| corvus | it seems like the gerrit folks like using bazelisk to handle running bazel (probably because they chase the bleeding edge and often have very specific version requirements) | 19:06 |
| tobiash | I think bazelisk is a good choice from what I've heard from collegues | 19:07 |
| clarkb | is it weird to anyone else that versions are so specific for build tools like that? | 19:08 |
| clarkb | I mean I can take my makefiles from 15 years ago and pretty sure they would work | 19:08 |
| *** jamesmcarthur has quit IRC | 19:08 | |
| clarkb | "back in my da" | 19:08 |
| tobiash | bazel is still under heavy development | 19:08 |
| corvus | yeah. it's an open question in my mind whether bazel will stabilize before the ecosystem decides that bazelisk is just always the way you should run bazel, and that becomes permanent :) | 19:09 |
| *** jamesmcarthur has joined #zuul | 19:09 | |
| tobiash | It might get more stable now that the 1.0 is out | 19:09 |
| openstackgerrit | Merged zuul/zuul master: Simplify virtualenv install and execution https://review.opendev.org/707428 | 19:09 |
| pabelanger | just catching up, so 20.0.2 should be fine now for virtualenv? | 19:10 |
| corvus | pabelanger: yes, and we'll release a zuul .1 soon with that update | 19:11 |
| clarkb | pabelanger: the answer is that it largely still depends. However the chagne above allows you to use old virtuaelnv if you still need it because the comamnd line doesn't need to change | 19:11 |
| clarkb | 20.0.2 doesn't break bwrap by default at least | 19:11 |
| clarkb | it does change its version command output and conflicts with system packaged six on centos 7 and 8 | 19:11 |
| pabelanger | k, let me see if tox works again without cap on centos-8 | 19:11 |
| pabelanger | Ah | 19:11 |
| clarkb | and the change above means you can use old virtualenv if those other problems are a problem for you | 19:12 |
| *** jamesmcarthur has quit IRC | 19:13 | |
| *** gothicmindfood has joined #zuul | 19:22 | |
| *** jamesmcarthur has joined #zuul | 19:30 | |
| *** jamesmcarthur_ has joined #zuul | 19:31 | |
| *** Defolos has joined #zuul | 19:32 | |
| *** jamesmcarthur has quit IRC | 19:35 | |
| *** jamesmcarthur_ has quit IRC | 19:54 | |
| *** igordc has quit IRC | 19:58 | |
| *** sanjayu_ has quit IRC | 20:15 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - bootstrap the driver structure + Webhook support https://review.opendev.org/685682 | 20:16 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - Basic handling of merge_requests event https://review.opendev.org/685990 | 20:16 |
| *** avass has quit IRC | 20:34 | |
| *** ChanServ has quit IRC | 20:50 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - bootstrap the driver structure + Webhook support https://review.opendev.org/685682 | 21:10 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - Basic handling of merge_requests event https://review.opendev.org/685990 | 21:10 |
| *** ChanServ has joined #zuul | 21:13 | |
| *** orwell.freenode.net sets mode: +o ChanServ | 21:13 | |
| corvus | clarkb, tobiash, mordred, fungi: commit 7c6503aba07f359f8ba023f2cceb7c06bb0829d0 (HEAD -> master, tag: 3.16.1, origin/master, origin/HEAD, refs/changes/28/707428/2) | 21:14 |
| corvus | that look right for a release? | 21:14 |
| tobiash | ++ | 21:15 |
| fungi | lookin' | 21:16 |
| clarkb | corvus: lgtm | 21:17 |
| jlk | Hey all, not sure if you've seen it already or not, but this looks like a possibly way to get off of using the search API in GitHub https://developer.github.com/changes/2019-04-11-pulls-branches-for-commit/ | 21:17 |
| *** rfolco has quit IRC | 21:18 | |
| fungi | corvus: does the github app auth revert need it to be 3.17.0? | 21:18 |
| fungi | just skimming `git log --no-merges 3.16.0..origin/master` | 21:18 |
| tobiash | fungi: no, it's just a bugfix that fixes a regression introduced short before | 21:19 |
| fungi | okay, then yeah 3.16.1 lgtm | 21:19 |
| tobiash | jlk: interesting, that blog post is almost a year old | 21:21 |
| jlk | yeah, I'm sad I missed it | 21:21 |
| jlk | but we just talked about it internally at Github | 21:21 |
| tobiash | jlk: do you know if that landed already in ghe? | 21:21 |
| jlk | we were discussing the desire to easily go from commit hash to a PR it came from | 21:21 |
| jlk | tobiash: I do not. | 21:21 |
| jlk | tobiash: and I'm pretty sure it's not in github3.py either | 21:21 |
| tobiash | That would be no problem, we already have quite a few raw calls | 21:22 |
| corvus | 3.16.1 pushed | 21:23 |
| jlk | also it sounds like this only works for _merged_ PRs | 21:24 |
| jlk | so blah. | 21:24 |
| tobiash | well, the docs sounds promising: Lists all pull requests containing the provided commit SHA, which can be from any point in the commit history. The results will include open and closed pull requests | 21:27 |
| tobiash | we'de have to filter for the head commit, but sounds usable | 21:27 |
| jlk | I'll report any further discussion. | 21:27 |
| jlk | It sounds like the data store used to look it up isn't ideal | 21:27 |
| tobiash | and it's already included in ghe 2.17 | 21:27 |
| tobiash | (which is the oldest currently supported release) | 21:28 |
| corvus | jlk: let me guess -- it's stored on the blockchain? | 21:33 |
| jlk | pffft | 21:33 |
| jlk | no, we're MSFT now, so it's data in a sharepoint somewhere | 21:33 |
| corvus | shoulda stuck with excel | 21:34 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Report retried builds in a build set via mqtt. https://review.opendev.org/632727 | 21:36 |
| tobiash | I'd have an easy change test case addition https://review.opendev.org/704983 (verify that retried builds don't trigger fail-fast) that needs reviews | 21:41 |
| corvus | tobiash: lgtm; a thought that occurs to me is that we could try to simplify some of the layouts used for tests like that (really only needs 2 jobs) | 21:42 |
| corvus | that may help with our test runtime | 21:42 |
| tobiash | good idea | 21:43 |
| corvus | i bet half of our tests are running with the extra-complex "standard" layout we inherited from the zuulv1/v2 days | 21:43 |
| tobiash | probably | 21:44 |
| corvus | mordred, mnaser: i wrote this role intending to add it to zuul-jobs, with everything i think we've learned about what an "ensure-" role should do: https://gerrit-review.googlesource.com/c/zuul/jobs/+/254772 | 21:44 |
| mnaser | corvus: it’s late here but I like that already. FYI the set fact sets the wrong fact name :) | 21:47 |
| corvus | mnaser: huzzah for code review :) | 21:47 |
| mnaser | :) | 21:48 |
| corvus | i think i just found a bug in gerrit checks support -- i think we report 'enqueue' events on non-live items | 21:48 |
| mordred | corvus: it looks great - I left some comments, only one of them is the reason for the -1 | 21:49 |
| tobiash | mordred: you were quicker with login ;) I was just about to report the same | 21:49 |
| mordred | tobiash: :) | 21:49 |
| mordred | tobiash: I cheated, I was already logged in | 21:49 |
| mordred | jlk: is that an azure hosted sharepoint aaS ? | 21:51 |
| mordred | jlk: ooh - you should reimplement sharepoint using github actions, then use that as the backing store for github ... wcpgw? | 21:51 |
| openstackgerrit | Merged zuul/zuul master: Don't run jobs if only their file matchers are updated https://review.opendev.org/706399 | 21:53 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Report retried builds in a build set via mqtt. https://review.opendev.org/632727 | 21:54 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Don't report enqueue of non-live item https://review.opendev.org/707479 | 22:01 |
| corvus | mordred, tobiash: ^ we're going to want that landed before we demonstrate more for the gerrit folks -- the current bug is counterproductive to my idea of building a nice long patch series. | 22:02 |
| clarkb | I feel like we've madethat change before but this is checks api specific | 22:03 |
| clarkb | so probably the one Im thinking of was in regular gerrit reporter | 22:04 |
| corvus | clarkb: probably something similar -- i think we added the 'enqueue' reporter for the checks api | 22:04 |
| openstackgerrit | Merged zuul/zuul master: Gerrit: add polling support for refs https://review.opendev.org/706138 | 22:06 |
| tobiash | corvus: that fixes the common pipeline manager, why does this only show up in gerrit checks? | 22:09 |
| tobiash | is that filtered in the other reporters? | 22:09 |
| corvus | tobiash: it would affect any reporter -- but gerrit checks is the only one where anyone is likely to configure an 'enqueue' reporter action. we added it specifically for gerrit checks (so that it could change the status to "scheduled") | 22:11 |
| corvus | i think it may have applicability to other systems in the future | 22:11 |
| tobiash | ah ok, I guess I confused it with the start reporting | 22:12 |
| corvus | yeah -- almost the same, but start is when the first job actually begins | 22:12 |
| corvus | tobiash, clarkb: https://gerrit.googlesource.com/zuul/config/+/refs/heads/master/zuul.d/pipelines.yaml#12 vs line 19 | 22:13 |
| tobiash | got it | 22:14 |
| *** sshnaidm has quit IRC | 22:17 | |
| *** sshnaidm has joined #zuul | 22:18 | |
| tobiash | corvus: btw, the first two optimization passes (offload setrefs and offlead reset) had a huge impact on our job preparation times. The 90th quantile dropped from >20min to <10min when the system is under load (~2500 job startups per hour) | 22:20 |
| *** jamesmcarthur has joined #zuul | 22:20 | |
| tobiash | (which is still quite a lot) | 22:24 |
| *** jamesmcarthur has quit IRC | 22:24 | |
| *** jamesmcarthur has joined #zuul | 22:25 | |
| tobiash | For further improvement I noticed that it sets all refs twice (reset and restore repo state) I'll also have a look if this can be avoided when we have a repo state | 22:26 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - Add zuul_event_id and set use get_annotated_logger https://review.opendev.org/692799 | 22:32 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - Minimal reporter ables to comment on MR https://review.opendev.org/694346 | 22:33 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Gitlab - Implement the note event and the comment trigger action https://review.opendev.org/698964 | 22:33 |
| *** sgw has quit IRC | 22:59 | |
| *** jamesmcarthur has quit IRC | 23:02 | |
| fungi | jlk: you couldn't even muster a microsoft bob joke there? | 23:22 |
| *** rlandy is now known as rlandy|bbl | 23:26 | |
| corvus | mordred: i'm looking at https://opendev.org/opendev/system-config/src/branch/master/playbooks/zuul/gerrit/repos.yaml#L23-L33 -- and i don't quite understand the difference between those 2 submodules and the ones above them | 23:26 |
| corvus | here's the gitmodules file: https://gerrit.googlesource.com/gerrit/+/refs/heads/master/.gitmodules | 23:26 |
| corvus | mordred: i don't see what "the actual ref" is in the case of download-commands | 23:27 |
| mordred | corvus: I also don't know how to know that | 23:29 |
| corvus | well, i mostly just don't understand, but i think i do now | 23:29 |
| corvus | you mean the sha of the submodule that was committed to the gerrit repo | 23:30 |
| corvus | (which isn't in the .gitmodules file, but is in the placeholder dir) | 23:30 |
| mordred | yeah - I believe that is what I mean | 23:31 |
| corvus | i think i was hung up on "ref" and thinking i was missing where that was specified. but the comment you left makes sense now if i think 'sha' instead | 23:31 |
| mordred | it's entirely possible that just overriding teh branch checkout for those to be master would DTRT | 23:31 |
| mordred | ++ | 23:31 |
| mordred | sha is what I should have said | 23:31 |
| corvus | mordred: oh, interesting -- so override-checkout: master on those, then do the mv instead of the update? | 23:32 |
| mordred | yeah - I think that ight actually be better from a zuul pov | 23:32 |
| mordred | since those are tracking submodules in gerrit anyway | 23:32 |
| mordred | so they're always going to be expecting to be whatever the state of their tracking branch is - in this case master since they don't have stable branches | 23:32 |
| mordred | *** I THINK *** | 23:33 |
| clarkb | why doesn't the submodule file say branch = master? | 23:33 |
| clarkb | is . short for HEAD maybe | 23:33 |
| jlk | fungi: I think Bob was let go by the previous leadership tim. | 23:33 |
| jlk | team | 23:33 |
| fungi | the entire team was named tim, got it | 23:34 |
| corvus | clarkb: yeah i think so: https://git-scm.com/docs/gitmodules#Documentation/gitmodules.txt-submoduleltnamegtbranch | 23:34 |
| corvus | i don't understand how that correlates with download-commands not having certain branches | 23:35 |
| *** sgw has joined #zuul | 23:36 | |
| corvus | clarkb, mordred: when i check out stable-3.0 of gerrit, i get a commit in download-commands which is a few commits back of master | 23:40 |
| corvus | so i'm guessing it sort of gets frozen once there isn't a matching branch? | 23:40 |
| corvus | (ie, gerrit's submodule tracking doesn't do anything anymore?) | 23:41 |
| *** sgw has quit IRC | 23:41 | |
| corvus | so would the most correct thing be to: if zuul.branch is a branch in the plugin repo, move the repo in place. otherwise, is there a change in the dependency chain for the plugin repo? if so, fail the job (the dependent change will not be updated in the submodule even when it lands). otherwise, do a 'submodule update --init' in the manner of mordred's playbook. | 23:44 |
| mordred | corvus: oh - yeah - maybe it's not magic trackigng and is a normal submodule | 23:45 |
| mordred | corvus: yes - I think so - although it's a little sad making in the fail-the-change in the case of dep in dep chain | 23:45 |
| corvus | mordred: i was thinking it was the gerrit tracking, but that gerrit tracking can't work if the branch specified doesn't exist | 23:45 |
| corvus | mordred: yeah, though we can fail with a message that says "make a stable-3.0 branch if you want to depend on a stable-3.0 change" | 23:46 |
| corvus | ok, this is my starting task list: http://paste.openstack.org/show/789488/ | 23:51 |
| mordred | corvus: ++ I like that message idea | 23:54 |
| mordred | and also - I like that it provides a vehicle to show a value of a stable branch always getting made | 23:55 |
| mordred | because the current "we don't always branch the plugins" approach is a bit messy and hard to deal with | 23:55 |
| corvus | mordred: i just went over the docs and the gerrit config, and i'm pretty sure they are using gerrit submodule subscriptions | 23:57 |
| corvus | i believe that https://gerrit.googlesource.com/Public-Plugins/+/3a681b5799af3decb4636af95bd108162d1492fc/project.config#30 along with a relative url in .gitmodules will activate it | 23:58 |
| mordred | yeah - I think the subscriptions work for all the things with matching branches | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!