| *** diablo_rojo is now known as Guest2025 | 01:46 | |
| *** amoralej|off is now known as amoralej | 06:12 | |
| *** diablo_rojo is now known as Guest2093 | 08:04 | |
| *** marios is now known as marios|call | 08:47 | |
| *** marios|call is now known as marios | 09:04 | |
| opendevreview | Elod Illes proposed openstack/releases master: Zed final releases for cycle-with-rc projects https://review.opendev.org/c/openstack/releases/+/860011 | 10:22 |
|---|---|---|
| *** dviroel_ is now known as dviroel | 11:40 | |
| opendevreview | Elod Illes proposed openstack/releases master: Mark Zed as released https://review.opendev.org/c/openstack/releases/+/860029 | 11:46 |
| opendevreview | Alfredo Moralejo proposed openstack/releases master: New release of puppet-swift for Yoga https://review.opendev.org/c/openstack/releases/+/859937 | 13:39 |
| *** dviroel is now known as dviroel|lunch | 15:11 | |
| *** amoralej is now known as amoralej|off | 15:48 | |
| *** marios is now known as marios|out | 15:51 | |
| *** diablo_rojo is now known as Guest2153 | 16:08 | |
| *** dviroel|lunch is now known as dviroel | 16:38 | |
| *** Guest2153 is now known as diablo_rojo | 16:55 | |
| JayF | Hey folks o/ I'm curious about how difficult it'd be to push a patch release from Ironic's bugfix/[] branches. It was indicated to me this might be a manual process (and different from cutting patch releases from supported-integrated release branches) | 18:10 |
| clarkb | it doesn't look like ironic's gerrit acl delegates any project specific perms to push tags https://opendev.org/openstack/project-config/src/branch/master/gerrit/acls/openstack/ironic.config | 18:14 |
| clarkb | assuming that a manual process is necessary I believe that means a release manager will need to tag the repo and push the tag. Or the acl will need to be modified to allow ironic to do it | 18:14 |
| JayF | That sounds roughly like what I was told would be needed (a human creating a tag). | 18:15 |
| clarkb | the manual process isn't too bad. YOu git tag -s the proper commit (double check the commit is known by gerrit first). Then git push the tag to gerrit. Then assuming your jobs are configured properly zuul should run off and build the artifacts and publish them for that version | 18:16 |
| JayF | Aight; I'll follow process here https://releases.openstack.org/reference/using.html#requesting-a-release and request releases of all ironic stable branches with changes (including bugfix/*) branches. Please let me know if there's something I push that's not accurate; this is my first time doing this :D | 18:17 |
| clarkb | as far as why this would require a manual push I wonder if the problem is in release tooling verification? Otherwise I don't think this is different than any other release | 18:18 |
| JayF | I honestly don't know; I was told it was a manual process so I thought I'd ask before requesting to ensure I'm not creating onerous amounts of work for others :) | 18:19 |
| JayF | clarkb: docs are unclear; is it a problem if I push a large number of releases in a single gerrit review? | 19:24 |
| * JayF is releasing stable patch releases ironic/ironic-inspector/ironic-python-agent/metalsmith from wallaby thru yoga | 19:25 | |
| JayF | In lieu of additional information; I'm going to do one review per project | 19:28 |
| JayF | scratch that, going to just do one PR, it's simpler if it's okay | 19:29 |
| opendevreview | Jay Faulkner proposed openstack/releases master: Release stable branches from w-y for Ironic projects https://review.opendev.org/c/openstack/releases/+/860125 | 19:31 |
| *** dviroel is now known as dviroel|afk | 19:37 | |
| clarkb | JayF: I'm not an expert on the release tooling but I think it is ok | 19:45 |
| JayF | Awesome. | 19:45 |
| JayF | Also, afaict, there is no way for me to git-request the tags for the Ironic bugfix/[] branches | 19:45 |
| JayF | did I miss something, or should I do that via email to the releases team? | 19:45 |
| JayF | Hmm. I see it now | 19:47 |
| JayF | ah, they are listed but no real way to request a release | 19:57 |
| clarkb | Not sure this is the correct venue, but I wonder why do the bugfix releases need to live beyond the point in time where there is a subsequent stable release | 20:07 |
| clarkb | seems like the bugfix releases should get supplanted by the stable releases | 20:07 |
| clarkb | otherwise what you actually have is a second set of stable releases and maybe the best way to deal with those is to call them stable and treat them like every other stable release | 20:07 |
| JayF | clarkb: Ironic has a set of users who use Ironic without integration with large amounts of other OpenStack projects (such as metal3). The bugfix/ intermediate releases are cut for these projects, and we maintain them and backport almost everything that goes in stable/() to the bugfix/() branches as well | 20:09 |
| JayF | We have a lot of known consumers of downstream releases dervived from those bugfix branches; there was some question in Ironic community if releases are similar consumed. | 20:10 |
| clarkb | right I undersatnd that | 20:10 |
| clarkb | what I'm saying is why can't they use the stable releases once they are cut | 20:10 |
| JayF | So I took the step of looking at actual pypi download data: | 20:10 |
| clarkb | basically the bugfix release comes out before stable/foo | 20:11 |
| JayF | clarkb: because those introduce new features/deprecations whereas the bugfix/[] is a point in time cut | 20:11 |
| clarkb | when stable/foo exists delete the bugfix release and move to stable/foo | 20:11 |
| clarkb | JayF: right in that case you have extra stable releases lets call them that | 20:11 |
| clarkb | and then maybe we can make the tooling work to support the use case | 20:11 |
| JayF | I do not love the naming as it exists; but the naming is not something I chose or really was around to collab with. | 20:11 |
| clarkb | basically either these are stable releases and the feature set and careful curation of fixes is important or they are not and you can upgrade to a stable release when it exists | 20:11 |
| JayF | Your sentiment is *exactly* why I'm looking to cut these releases. | 20:12 |
| JayF | I was frustrated we treated them as stable releases, but then only really gave access to those bugfixes to deployers sophisticated enough to pull them from git | 20:12 |
| JayF | or customers of downstream packagers who integrated those fixes from git | 20:12 |
| clarkb | ah that makes sense | 20:12 |
| JayF | I'm trying to ensure this effort is available to all users of Ironic, so I'm trying to get actual-releases cut | 20:13 |
| JayF | I am somewhat expecting some number of "WTF" and "please help fix the tooling" responses, but requesting something be done in a nonideal way is often the best path to learning the ideal way :D | 20:13 |
| clarkb | I think if I were a user and I saw that ironic has bugfix and stable releases I would be wary of using bugfix releases if what I wanted was a stable release | 20:13 |
| JayF | I think the users of Ironic outside of OpenStack don't view it as stable / bugfix branches, they just see version numbers and pin | 20:14 |
| clarkb | the other upside to using consistent terminology is that tools built around the terminology will suddenly work (though I grant there may be confusion over what the stable branch names should be) | 20:16 |
| clarkb | in the past ironic has struggled with mapping bugfix branches onto other branches of other software. Zuul by default maps to your default branch if there isn't a 1:1 match across repos | 20:16 |
| clarkb | so bugfix/old that is roughly stable/foo gets tested with master of other stuff instead of stable/foo | 20:16 |
| JayF | We do have patches up that ... carefully curate that | 20:17 |
| JayF | but it's mostly humans moving CI configs around, not really automation | 20:17 |
| JayF | and I had to -1 one today b/c it was missing something | 20:17 |
| clarkb | right its totally doable to override it. but if you align things by default then in theory a lot of that manual curation goes away | 20:17 |
| clarkb | But I'm not sure how you align things if everyone else is not having a stable/foo.5 | 20:17 |
| clarkb | I'd have to think about that a bit more | 20:17 |
| JayF | yeah exactly, and I think that might be why it's named differently | 20:18 |
| JayF | I know how it is today, I don't know *why* it was done that way | 20:18 |
| JayF | for that I'd have to do some historical reading | 20:18 |
| opendevreview | Douglas Mendizábal proposed openstack/releases master: Update Barbican release liaison https://review.opendev.org/c/openstack/releases/+/860152 | 20:29 |
| JayF | clarkb: ... sometimes I wonder if you're clairvoyant | 21:27 |
| JayF | clarkb: https://zuul.opendev.org/t/openstack/config-errors | 21:27 |
| JayF | ironic is on this list, for disused bugfix branches | 21:27 |
| JayF | which have no real retirement process afaict | 21:28 |
| clarkb | ya I think that is something ironic may still need to sort out (the retirement process for those branches) | 21:34 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!