Wednesday, 2022-04-27

ianwclarkb: thanks for working through that.  some meta comments on 838863; i think wait for a few other eyes on it but lgtm06:36
rpittauclarkb, ianw, thanks for working on that, much appreciated07:02
rpittaulooks like also the issue we had with the upgrade job has gone, still need to retest with the correct stable branch, but it looks ok now07:02
clarkbfungi: if you get a chance can you look at the comments about the relaese notes in https://review.opendev.org/c/openstack/diskimage-builder/+/838863 and see if you've got suggestions on how to address that?15:18
fungiclarkb: to summarize, the revert is also reverting a release note which appeared in a prior release?15:20
clarkbfungi: yes beacuse it undoes what the previous release note documented.15:20
clarkbmaybe we keep the previous release note and then add a new release note that mentions the change and just track it that way?15:21
fungii would keep persistent-netifs-fallback-1b861e037dd8a27c.yaml and make sure dhcp-all-interfaces-debian-issue-2ec7f1a5186f0b16.yaml is clear that it's a reversal15:21
fungiremoving release notes is akin to rewriting history and hiding problems15:21
clarkbcool I'll update the revert to not revert the release note15:22
fungiclarkb: any idea what changed with reno? https://docs.openstack.org/releasenotes/diskimage-builder/15:39
fungithe current release of dib is 3.20.3 but for some reason the release notes are smashing all entries after 3.18.0 together and adding a -n suffix which seems to correspond to the number of additional tags15:41
opendevreviewClark Boylan proposed openstack/diskimage-builder master: Revert "Fallback to persistent netifs names with systemd"  https://review.opendev.org/c/openstack/diskimage-builder/+/83886315:41
clarkbfungi: no clue. I'm kinda surprised dib is in the openstack release notes though15:42
fungiwell, it's an openstack project15:42
clarkbya but it isn't managed by the release team15:42
clarkbintentionally so to avoid the chicken and egg the release team hit recently with other tools15:42
clarkbyou can't have the tools the release team depends on managed by the release tooling or you can get wedged15:43
fungiright, but the release notes jobs aren't dependent on the openstack release management tooling as far as i'm aware, so even independent release projects run them15:44
clarkbI wonder if the release notes updates rely on integration with the actual release process now and since dib doesn't do that the builds are confused about what the release is?15:44
clarkbfungi: independent release projects don't manually tag and push they still go through the release tooling15:44
fungihttps://releases.openstack.org/reference/release_models.html#independent "Release tags for deliverables using this tag are managed without oversight from the Release Management team."15:46
clarkbhuh that must've changed? I thought all independent meant was that they didn't release on a set schedule15:47
clarkblooks like there is a release notes tox env which I assume is used to produce the documents that are weird15:48
fungibut yeah, you may be right, https://opendev.org/openstack/releases/src/branch/master/deliverables/_independent has things which are being tagged through automation rather than directly15:48
fungimaybe "oversight" doesn't mean what i think it means15:49
clarkbif I run `tox -e releasenotes` it seems to do the right thing15:50
clarkbI get those 3.18.0-5 entries listed under 3.19.0 locally and my edit from the change above as 3.20.3-515:50
clarkbmaybe the job that updates stopped running before the 3.19.0 tag was made15:51
clarkbso we're seeing a snapshot in time that is stale15:51
clarkbfungi: ^ thats my best guess. The job that updates that page is just not runnign anymore for one reason or another16:00
fungioh! yes, i bet you're correct. the -5 was probably the number of commits and reno isn't listing that as "in development" like it does normallyu16:01
fungiclarkb: confirmed, promote-openstack-releasenotes last ran for dib on the 5th commit after 3.18.016:16
funginow to figure out why it stopped being run (it's not failing, just no longer in the pipeline)16:17
fungiunfortunately it stopped after 2022-02-21 so we no longer have logs indicating where it was coming from16:18
fungiluckily there were a bunch of changes merged that day, so i can say it stopped between 18:56:53 and 19:34:0016:20
fungioh, except that job probably doesn't get added if the files list doesn't include a release note16:24
fungiaha! so that was in fact the last release note merged. because release notes don't get updated in the release/tag pipeline, the published release notes end before the 3.19.0 tag was made16:26
opendevreviewMaksim Malchuk proposed openstack/diskimage-builder master: Revert "Fallback to persistent netifs names with systemd"  https://review.opendev.org/c/openstack/diskimage-builder/+/83886316:29
fungiclarkb: any idea why it calculates a dev version number there instead of showing "in development" like our release notes in opendev and zuul? https://zuul-ci.org/docs/zuul/latest/releasenotes.html16:32
fungii wonder if that's configurable16:32
fungianyway, other openstack projects seem to include publish-openstack-releasenotes-python3 in their tag pipeline to solve this16:32
fungii'll propose that addition16:33
clarkbI don't know16:35
fungiseems other openstack projects get it by adding the release-notes-jobs-python3 project-template directly in their .zuul.yaml16:36
fungidib missed the transition from release-notes-jobs to release-notes-jobs-python3 and the former dropped its tag pipeline in 2018 with https://review.openstack.org/62243016:39
opendevreviewJeremy Stanley proposed openstack/diskimage-builder master: Switch to release-notes-jobs-python3  https://review.opendev.org/c/openstack/diskimage-builder/+/83959916:43
fungiit would be very nice to include that ^ before the next release16:43
fungigit show16:43
fungihah, you're not my shell terminal16:43

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!