Thursday, 2025-08-28

TheJuliastevebaker[m]: ack00:38
opendevreviewIury Gregory Melo Ferreira proposed openstack/ironic-prometheus-exporter master: Simplify some logic in IPE  https://review.opendev.org/c/openstack/ironic-prometheus-exporter/+/95871403:26
abongalegood morning ironic o/08:37
queensly[m]Good morning o/08:48
opendevreviewHervé Beraud proposed openstack/ironic master: Add hacking check to ban eventlet imports  https://review.opendev.org/c/openstack/ironic/+/95873809:33
dtantsurcardoe: since metal3 already defaults to quay.io, it feels naturally (to me) to pick up IPA images from there too10:08
dtantsurta-dam!! https://www.cncf.io/blog/2025/08/27/metal3-io-becomes-a-cncf-incubating-project/10:19
iurygregoryUHULLLLLLLLLLLLL11:34
JayFI didn't know ironic was a CNCF project 😂12:54
dtantsurand an incubating one!12:54
dtantsurTheJulia: hey, do I get it right that instance_info[image_pull_secret] accepts the token itself, not the entire configuration?12:55
TheJuliaWell, being in the larger LF family enables things to get mentioned in press releases without issue/concern... so things become easier12:55
TheJuliadtantsur: that is correct, if you want to hand the process the config, you can use the override setting if it is not at the expected path12:56
TheJuliathat code matches by hostname if memory serves12:56
dtantsurTheJulia: we need to document how to use image_pull_secret then.. I'm not sure the users are familiar with the format12:56
* TheJulia tried to leave others *all the options*12:56
TheJuliaInteresting you'd say that because the mapping seemed quite direct to me, at least with OCP in mind12:57
dtantsurWe mention the config in https://docs.openstack.org/ironic/latest/admin/oci-container-registry.html#authentication but don't say "you need to get this stuff from the right section"12:57
TheJuliaYeah, often not drawn as a direct line12:58
TheJuliaI'll make a doc update after I have another cup of coffee in hand12:58
dtantsurAnd that means we need to extract the token in BMO. Should not be hard though.12:59
dtantsur(background: I start feeling that mounting the OCP's pull secret is a dead end - in addition to it being surprisingly hard)12:59
TheJuliaI thought you were going to use the path of just pulling the credentials from the cluster and presenting it to ironic in general?12:59
TheJulia(... oh. that stinks)12:59
dtantsuryeah, cooling down a bit about this idea12:59
TheJulia:(12:59
TheJuliaI guess it depends on what as well12:59
dtantsurEspecially since I've realized that per-host pull secrets are probably inevitable anyway (custom registries, private quay accounts)13:00
TheJuliaFair enough, I think there is some frameworks to do container level injections for specific use credentials as well, so going broader for image access could definitely be harder13:02
JayFTheJulia: I figured it was something like that. Not that I actually care, it was just funny to read13:05
dtantsurYeah, a trivial oversight. The blog post was composed really quickly, basically in one evening, by a few busy people.13:06
TheJuliaIt is what it is, I'm sooo worried ;) </sarcasm>13:07
TheJulialol13:08
TheJuliawe need to get john garbutt to join in the idrac fun :)13:16
dtantsurwhat did he do to you? :)13:16
TheJulia.... he posted two patches which screamed idrac ;)13:19
TheJuliaand now I'm curious :)13:19
TheJuliaand more, curious because I care and want things to work13:19
TheJulialess actual curiosity, but its the only word i have to describe the feeling13:19
opendevreviewJulia Kreger proposed openstack/ironic master: docs: trivial: clarify pull secrets for OCI image access  https://review.opendev.org/c/openstack/ironic/+/95876513:26
TheJuliadtantsur: lmk if that provides more clarity13:26
dtantsurTheJulia: much better, maybe mention explicitly extracting the value under the right host key? (and point to the example below?)13:28
TheJuliahmmm13:28
TheJuliayour looking at it very differently then, but I see your point13:28
TheJuliarevising13:28
opendevreviewJulia Kreger proposed openstack/ironic master: docs: trivial: clarify pull secrets for OCI image access  https://review.opendev.org/c/openstack/ironic/+/95876513:40
TheJuliaokay, a bit more verbose and allowing for direct lines to be drawn13:40
opendevreviewJulia Kreger proposed openstack/ironic master: docs: trivial: clarify pull secrets for OCI image access  https://review.opendev.org/c/openstack/ironic/+/95876513:48
TheJuliadtantsur: revised, thanks!13:49
JayFTheJulia: those two patches are not targeting at fixing a Dell for what it's worth13:59
JayFTheJulia: The ones from John.14:00
TheJuliaOh, oh :(14:00
JayFI think maybe there's just a pattern of new stuff is weirdly shaped14:00
TheJulia-a popular white box vendor?14:00
TheJuliaYeah, likely14:01
JayFI'm generally not permitted to say specifically what hardware we use in public.14:01
TheJuliaK14:01
opendevreviewMerged openstack/ironic master: Follow-up: Fix improper HTTP status code usage  https://review.opendev.org/c/openstack/ironic/+/95866314:34
clifis tenks a defunct/dead project? the link on the bug deputy page sends me to a 404 page on launchpad14:59
clifwell, it looks active-ish when I look at the repo15:00
clifcommits from 9 months ago at least15:01
clifhmm maybe the project tracking for tenks has moved here: https://storyboard.openstack.org/#!/project/openstack/tenks15:05
opendevreviewDmitry Tantsur proposed openstack/ironic master: Revert "Switch from local RPC to automated JSON RPC on localhost"  https://review.opendev.org/c/openstack/ironic/+/95877615:11
opendevreviewClif Houck proposed openstack/ironic master: Tenks bug tracker has moved  https://review.opendev.org/c/openstack/ironic/+/95877715:16
JayFclif: it should be on launchpad, not storyboard15:22
JayFclif: but also tenks is ... probably getting close to time for it to go out to pasture15:22
JayFor for our stackhpc friends to maintain it better15:23
clifshould be? ok, but right now the documentation for the project points to the link I posted15:32
clifif the project itself is linking to storyboard.openstack.org what should we link to?15:33
JayFI'm saying it's a project-governance-bug that there's no launchpad tracker for it15:34
JayFwe as a project are not supposed to be on storyboard anymore15:34
JayFthat being said; the fact it's wrong is emblematical of how dead that project is15:34
JayFso 1) from a bug deputy perspective I wouldn't worry about it and 2) we really need to retire it OR do the legwork to make it right15:34
* JayF votes to retire it15:34
clifit does have a commit from a couple of months ago15:37
clifare you just saying it shouldn't be part of openstack anymore?15:38
JayFI'm saying IMO it's not an ironic project, it's just a roommate for ironic in practice :)15:39
JayFRetiring things requires concern about them though, so you end up in this ... semi-rotten state where only the parts that directly impact folks get addressed15:39
JayFand there aren't many folks using it to be impacted15:39
clifshould I just drop it from the bug deputy list then?15:43
JayFI would +2 such a change personally15:44
cliffair enough15:45
JayFthe real evidence it's borderline dead from an openstack release perspective: last tagged release was Jan 2024 afaict https://github.com/openstack/tenks/tree/2.0.0 and it added... python 3.10 support :D 15:45
TheJuliaI like the analogy15:52
TheJulia(the, its a friend that lives in our house)15:52
JayFit'd be more work for us to retire tenks than anyone who works on other parts of openstack have spent on it in years combined15:53
TheJuliayeah15:53
JayFso we just leave the guy on couch alone and let 'em sleep ;) 15:53
TheJuliaheh15:53
TheJuliayeah15:53
opendevreviewMerged openstack/networking-baremetal master: tox: Remove ineffective ignore_basepython_conflict  https://review.opendev.org/c/openstack/networking-baremetal/+/95763116:27
opendevreviewMerged openstack/networking-generic-switch master: Ignore newly enforced pep8 rules  https://review.opendev.org/c/openstack/networking-generic-switch/+/95870816:29
tkajinamhmm I don't know why all of the flake8 plugins from neutron-lib are now enabled16:32
tkajinamI'm seeing the same problem started hitting heat suddenly, too...16:33
tkajinamI suspect any change in hacking/flake8 caused this but I didn't find anything clearly related...16:33
TheJuliatkajinam: fun :(16:36
TheJuliatkajinam: https://opendev.org/openstack/neutron-lib/commit/9b6a1aafc394df6e29cde3534167e53279f727f616:42
TheJuliaI guess if neutron-lib is installed, then the new rules get pulled in16:42
tkajinamohhh k16:42
tkajinamI'll revert it16:42
tkajinamthey can use local-plugins to enable these plugins afaik16:43
TheJuliaYeah, seems like it is crossing project boundaries 16:43
tkajinamor probably they should have default off16:43
tkajinammaybe default off is a better approach16:43
TheJuliamaybe, dunno, they likely need to be aware of it and revert (and release again)16:44
tkajinam( I was checking only hacking directory. that's a good fine16:44
tkajinamyeah16:44
TheJuliaI guess that is the other problem, if the release team will even let them release it at this point16:44
tkajinamyeah16:44
TheJulia(I was like, wait a second... and used codesearch for N530)16:44
tkajinamanyway I'll report a bug to kick the discussion16:45
TheJuliacool, thanks!16:45
opendevreviewMerged openstack/networking-generic-switch master: Minor doc generation fixes  https://review.opendev.org/c/openstack/networking-generic-switch/+/95849816:59
opendevreviewSteve Baker proposed openstack/networking-generic-switch master: Create service plugin to configure security groups  https://review.opendev.org/c/openstack/networking-generic-switch/+/94983417:19
opendevreviewClif Houck proposed openstack/ironic master: Remove tenks references  https://review.opendev.org/c/openstack/ironic/+/95879817:37
TheJuliaqotd: As a project, what kind of building is Ironic?!18:12
JayFWe're the sub basement with all the creepy creepy corners18:38
TheJuliaHas anyone considered AI driven code review?20:51
TheJuliato at least be a basic first pass?20:51
JayFIf we were github, I'd probably have tried it on ironic.20:53
JayFGenerally, I am ... suspicious it'd be useful20:53
JayFespecially given how much *I* have to rein in the tool even when I'm using it20:53
JayFthey really, really like to follow some average of all python patterns even when you give it lots of context about how openstack does things20:54
TheJuliaJust toying with claude and some specific changes, and it was interesting to see what it was coming back with20:54
JayFPlus apparently the copilot code review stuff on github can't be blocked at a project level... so I don't love that.20:54
JayFHow did you do it? just git review -d and point claude asking it to review HEAD20:54
TheJuliaand... interestingly enough, it was coming back with some feedback which mirrors some of the review feedback20:54
TheJuliayup20:54
TheJulia"review git commit $blah, and evaluate risk of regression, provide a summary in 80 words"20:55
TheJuliafun is added when  you ask it to provide a risk analysis table, or when you ask it to do so as if it was shakespear20:55
JayFI've done that pattern before where I have claude do a thing in window A, then ask claude in window B to review it20:55
JayFI think you can use /agent to do that in one window now20:55
JayFMy biggest claude pattern I think is useful: I have ~/dev/openstack/ with a CLAUDE.md, all my openstack projects checked out beneath. That allows you to use other openstack projects as templates.20:56
JayFFor the wsgi docs change, that's the approach I took; I had it use other wsgi examples from openstack 20:56
TheJuliaI've been using it with repos in isolation, but that is likely a better case for some cross-project related patterning20:59
JayFwell I started it because of asking questions about nova ironic driver and ironic behavior21:00
JayFjust makes it easier to reason about when you can say "do X similarly to how they did it in path/to/module"21:00
TheJuliayeah21:01
JayFbtw if you wanna see21:01
JayFhttps://www.irccloud.com/pastebin/VgwvIPWd/claude-md.txt21:02
opendevreviewVerification of a change to openstack/networking-generic-switch master failed: Create service plugin to configure security groups  https://review.opendev.org/c/openstack/networking-generic-switch/+/94983421:13
TheJuliagaaaah21:13
* cardoe stares at the wall.21:25
TheJuliaso, that job should have passed but nova failed21:26
TheJuliawhich is sort of the same21:26
* TheJulia stares at the wall.21:26
opendevreviewJay Faulkner proposed openstack/ironic master: Add node.instance_name  https://review.opendev.org/c/openstack/ironic/+/95279022:41
JayFhttps://review.opendev.org/c/openstack/ironic/+/957613/4#message-dcae860b0e59de395e9a75f3742a6c7a839960e3 if other folks have a different opinion or agree please say so22:43
opendevreviewMerged openstack/ironic master: Add hacking check to ban eventlet imports  https://review.opendev.org/c/openstack/ironic/+/95873823:00

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!