| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 902118: Freeze job dependencies with job graph https://review.opendev.org/c/zuul/zuul/+/902118 | 02:51 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 899650: dnm: Filter regex deprecation warnings in web API https://review.opendev.org/c/zuul/zuul/+/899650 | 12:32 | |
| @harbott.osism.tech:regio.chat | is there a plan when the next zuul release shall be made? I assume with the SQL migration it would be a major version bump? | 14:51 |
|---|---|---|
| @jim:acmegating.com | probably soon, and i expect it to be a minor version bump. we typically only do major version bumps with user-visible breaking changes and have not done so for a sql migration. | 14:54 |
| @jim:acmegating.com | * probably soon, and i expect it to be a minor version bump. we typically only do major version bumps with breaking changes and have not done so for a sql migration. | 14:55 |
| @fungicide:matrix.org | the sql migration should be safe to perform online during a rolling upgrade, right? | 15:16 |
| @jim:acmegating.com | fungi: no -- https://zuul-ci.org/docs/zuul/latest/releasenotes.html#upgrade-notes | 15:27 |
| @fungicide:matrix.org | aha. looks like the last schema migration we mention in changelogs was for 6.1.0 though it doesn't explicitly state the schedulers should be stopped, the migration in 5.1.0 does say that though. i agree we have precedent for schema migrations (even offline ones) in minor releases | 15:33 |
| @jim:acmegating.com | yeah, i think we have precedent here, but also, the rules are flexible. i'm just trying to be consistent, but the numbers don't cost us anything so if we want to do a major bump, that's fine (but if we do, i'm going to propose yanking a whole bunch of stuff out first -- i've been holding off on that :) | 15:35 |
| @fungicide:matrix.org | are we overdue for dropping another ansible version? | 15:36 |
| @fungicide:matrix.org | we've definitely done major version bumps coincident with that | 15:36 |
| @jim:acmegating.com | we could stand to drop 6. yes | 15:36 |
| @tony.breeds:matrix.org | RH has an internal gitlab and we need to use it as a config repo and support all there goodness of that. The issue I *think* I have is that we need to support muliple connections because we don't have a "global" zuul-ci user, so each project group has it's own access token. I think this is a problem as ultimately it's the same server so zuul will pick the first matching connection which may or may not have access to the repos we need. | 16:20 |
| @tony.breeds:matrix.org | 1) Have I more or less understood correctly ; 2) Is there a work around, other than getting a global user which can be added to the appropriate groups? | 16:20 |
| @clarkb:matrix.org | tony.breeds: I suspect you may know best simply beacuse you've got a gitlab and a zuul talking to each other. However, zuul connection have canonical names associated with them and those names are part of fully qualified canonical repo names. For example opendev's zuul talks to review.opendev.org but canonically these repos are named opendev.org/foo/bar and that prefix comes from the connection settings. I suspect that you can force specific connection to be used if you always use fully qualified names | 16:39 |
| @clarkb:matrix.org | but I'm not positive of that | 16:39 |
| @clarkb:matrix.org | it does seem weird to em that you would need different identities for the same system though. Zuul should be able to operate as the "rhzuul" in rhgitlab and then projects within that system could give appropriate connections to the account | 16:41 |
| @clarkb:matrix.org | I doubt you force your human users to authenticate as different users depending on the project they push to | 16:41 |
| @tony.breeds:matrix.org | Clark: Yes I agree that it's weird. ATM the gitlab operators are against making a rhzuul user inside rhgitlab. | 16:45 |
| @tony.breeds:matrix.org | Thanks for the help. | 16:46 |
| @mordred:inaugust.com | hujdttkhfrnt | 16:51 |
| @clarkb:matrix.org | tony.breeds: if you do test the canonical names thing and it works it might be good to check if we sufficiently cover the behavior in docs. If it doesn't work then that might be a bug? | 16:52 |
| @tony.breeds:matrix.org | Clark: good point. | 16:53 |
| @jim:acmegating.com | zuul-maint: how does this look for a zuul release? | 16:58 |
| commit 27a4beb6980ac914d191dbac119aff785e242b3a (HEAD -> master, tag: 9.3.0, origin/master, gerrit/master, refs/changes/07/902707/1) | ||
| @jim:acmegating.com | that's current head, one commit past what opendev is running; the one commit is the test fix for the ansible cve | 16:58 |
| @clarkb:matrix.org | tony.breeds: corvus I made some small edits to the zuul annual report doc to accomodate your feedback can you check if it looks better now? | 17:28 |
| @clarkb:matrix.org | also the number of releases is about to change. I'll make a note to update that value when we provide a "final" version to the foundation | 17:28 |
| @clarkb:matrix.org | also I'll double check the contribution numbers haven't changed drastically at that point in time | 17:29 |
| @tony.breeds:matrix.org | Clark: looks good to me | 17:30 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 902118: Freeze job dependencies with job graph https://review.opendev.org/c/zuul/zuul/+/902118 | 18:57 | |
| @clarkb:matrix.org | https://learn.stanford.edu/DonKnuthXmas-2023-Registration.html there appears to be some overlap in this year's Christmas talk and the speculative execution zuul performs | 21:02 |
| @jim:acmegating.com | i remember the dancing links talk | 21:08 |
| @clarkb:matrix.org | > <@jim:acmegating.com> i remember the dancing links talk | 21:12 |
| Do you go to those in person? | ||
| @jim:acmegating.com | so far yes; i may not make it over there this year and might just watch the stream | 21:13 |
| @clarkb:matrix.org | nice | 21:18 |
| @jim:acmegating.com | they're getting pretty crowded, so it's kinda hard to actually talk to him now. good if you want a free book or to get one signed. otherwise, the stream has better audio. :) | 21:20 |
| @jim:acmegating.com | (*note: free book may be in a random human language) | 21:22 |
| @fungicide:matrix.org | you had me at human language | 21:31 |
| @jim:acmegating.com | 9.3.0 released and email announcement sent | 21:39 |
| @clarkb:matrix.org | corvus: we didn't manage to run into each other doing that with the gerrit restart did we? | 21:41 |
| @clarkb:matrix.org | (I don't think we did) | 21:41 |
| @jim:acmegating.com | nope; the tag push happened a while ago | 21:43 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!