Thursday, 2025-09-04

TheJuliacardoe: comments welcome, I think the evpn way is the go-to way when I really think of it, because then we balance hot spotting versus a huge amount of complexity, but I would need to go udnerstand the evpn model and all that00:06
TheJuliaIt still might need to be a service, u dunno00:16
TheJuliaerr, I dunno00:16
cardoeTheJulia: first review comment… `pre-commit install` in your shell in the ironic-specs directory.00:31
TheJuliaYeah, its still a wip which I know won't pass linting rightnow00:33
*** bbezak_ is now known as bbezak06:22
abongalegood morning ironic o/07:49
opendevreviewGrzegorz Koper proposed openstack/networking-generic-switch master: Support 802.3ad port groups on Arista devices  https://review.opendev.org/c/openstack/networking-generic-switch/+/95911809:39
FreemanBoss[m]good morning09:39
opendevreviewGrzegorz Koper proposed openstack/networking-generic-switch master: Fix incorrect command when unplugging bond subports  https://review.opendev.org/c/openstack/networking-generic-switch/+/95942309:50
opendevreviewGrzegorz Koper proposed openstack/networking-generic-switch master: Support 802.3ad port groups on Arista devices  https://review.opendev.org/c/openstack/networking-generic-switch/+/95911810:48
opendevreviewGrzegorz Koper proposed openstack/networking-generic-switch master: Support 802.3ad port groups on Arista devices  https://review.opendev.org/c/openstack/networking-generic-switch/+/95911810:49
opendevreviewJakub Jelinek proposed openstack/ironic-python-agent master: Fix skip block devices for RAID arrays  https://review.opendev.org/c/openstack/ironic-python-agent/+/93734211:06
kubajjWe are sorry  for the size of the patch ^11:07
opendevreviewBartosz Bezak proposed openstack/networking-generic-switch master: Support 802.3ad port groups on Arista devices  https://review.opendev.org/c/openstack/networking-generic-switch/+/95911812:08
opendevreviewluocanhui proposed openstack/ironic master: Add UT for metrics modules  https://review.opendev.org/c/openstack/ironic/+/95802812:18
TheJuliafolks, could we get more eyes on https://review.opendev.org/c/openstack/ironic/+/959389 I'd like to get a final release out the door sooner rather than later13:29
opendevreviewTakashi Kajinami proposed openstack/ironic master: Replace unused oslo.service options  https://review.opendev.org/c/openstack/ironic/+/95951013:46
cardoe+214:12
TheJuliaOnly one more core for the reno!14:31
dtantsurTheJulia: we don't use the term "introspection" any more in Ironic, is it something you want to fix?14:32
TheJuliadtantsur: why don't you rev the text directly14:33
TheJuliaIn the interests of speed14:33
dtantsurThe question is whether you want to lose the +2 :)14:33
TheJuliaand efficiency14:33
opendevreviewDmitry Tantsur proposed openstack/ironic master: Add a prelude for 2025.2  https://review.opendev.org/c/openstack/ironic/+/95938914:34
dtantsurupdated and +214:34
TheJuliaworkflowed, thanks14:54
TheJuliaAnyone screaming to not release 32.0.0 once merged?15:13
JayFahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh15:13
JayF[screaming unrelated to release]15:14
* dtantsur is worried15:14
TheJuliadtantsur: normal worried, or earth shattering oh noes?15:15
* TheJulia is largely just trying to drive the loop here, ensure we release to ensure the feedback engine will loop around again since it is all a self-feeding machine15:16
dtantsurI'm worried by JayF screaming (unrelated to release)15:16
TheJuliaoh, good15:17
JayFdtantsur: does anyone need a specific reason to scream these days? I think there's lots of screaming left unscreamed. Just evening the scales.15:17
dtantsurI don't have anything blocking the release. https://review.opendev.org/c/openstack/ironic/+/958072 and https://review.opendev.org/c/openstack/ironic/+/955762 could be nice but also could be backported15:17
dtantsurJayF: fair enough. I'm not screaming only because if I start I'm not going to end15:18
JayFI have a sore throat so  I had to scream in IRC instead of IRL15:18
JayFapproved both those with a recheck15:20
dtantsurthanks!15:20
TheJuliaWorst comes to worst, we can backport them in15:20
opendevreviewMerged openstack/ironic master: Add a prelude for 2025.2  https://review.opendev.org/c/openstack/ironic/+/95938915:20
fricklerspeaking of releases, did you see https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/7THC2L43DYKEQUURQ6OFIV2P7SLM4SXL/ ? (ipa-builder and network-baremetal)15:36
jovialI'm updating the introspection rules in kayobe to be compatible with the built-in inspector. I can't seem to figure how to replicated the is-empty condition. I've come up with this hacky workaround: https://gist.github.com/jovial/c78cefea45b7eecb2cb5718bfb263abb Is there a better way?15:37
jovialI did try and use one-of: https://gist.github.com/jovial/5ed015a01bc8a23d392176a475392512 but that didn't work15:41
jovialEssentially, I wanted to recreate that logic of not updating a property if a value has already been set15:43
jovialI think the difference is that the old is-empty used to work with missing values: https://github.com/openstack/ironic-inspector/commit/50b989474d874c6004dd2b4f0c1ec598007a2fcb15:47
TheJuliafrickler: I can't load the url right now, but if its ttx's email from 1.5-ish days ago, its literally the next item I'm going to address unless someone runs infront of me with hair on fire. :)16:05
TheJuliajovial: I guess I'm sort of wondering why you'd have such a rule mapping behavior like that anyhow. But that doesn't takeaway from the base question on the match which someone will need to dig into16:07
TheJuliafrickler: yeah, thats the email, it finally loaded here16:08
jovialTheJulia: As in only setting a field if it wasn't already set? For this particular case, I was trying to replicate the old behavior, where you could manually set the property on a handful of nodes and have it automatically set on others. I guess the other option is just always set it.16:12
TheJuliaI guess your making it consistent in a way in your environment, or at least attempting to. But it should not be changing at all across deploys/operations *unless* the hardware is being changed actively16:14
TheJuliaat which point, it sort of makes sense, I guess16:14
TheJuliaA super-quick ipa-b patch: https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/95925516:21
jovialIn particular, I quite like the root device hint as otherwise, I believe, it selects the smallest disk and that could change if your root disk dies and then you end up imaging the wrong disk16:22
TheJuliajust some wireless firmwares to remove16:22
TheJuliajovial: fair, and it is the smallest device greater than ?4 GB? in side.16:22
TheJuliasize16:23
TheJuliaso, looks like ironic, networking-generic-switch, networking-baremetal, ironic-inspector.16:27
TheJuliaand ironic-python-agent-builder16:29
opendevreviewClif Houck proposed openstack/ironic master: Add a new 'category' field to the Portgroup object  https://review.opendev.org/c/openstack/ironic/+/95571316:31
TheJuliaand ironic-python-agent as well it looks like16:41
opendevreviewMerged openstack/ironic master: Replace unused oslo.service options  https://review.opendev.org/c/openstack/ironic/+/95951016:45
TheJuliaheh, already have to rev the stack of releases, woot!16:45
TheJuliafrickler and ironic-cores: https://review.opendev.org/q/topic:%22ironic-releases-2025.2%2216:54
JayFno ironic-inspector in 2026.1, right?17:00
JayFI'm going to run the paperwork to retire metalsmith after the release; curious if I should make it a twofer17:01
TheJuliaI *think* that is what we said, we'd need to doublecheck release notes17:04
clifIt looks like people are already aware of https://bugs.launchpad.net/ironic/+bug/2121702 but it's new as of six days ago, unassigned, and importance is undecided17:05
clifif someone is looking at this please assign yourself and triage17:06
TheJuliaI'm 95% sure its legit in the latest version and is related in different projects being used to admin ironic versus who created the networks17:09
clifJayF: I guess this is the bug you were trying to gently point me at if I needed a break from the yaml config stuff17:09
JayFthat's not the bug I was pointing you at17:12
clifoh17:12
JayFbut it sure as hell would be a baptism by fire into the RBAC/network intersection17:12
JayFand would be a valuable use of time17:12
JayFthis may end up being a "this is a spec and year worth of work" bug rather than something trivially fixable, but it'd still be neat to know17:13
clifyou're certainly selling me on it17:13
JayFwell I've said to you before that I think improving the model between ironic/nova/neutron around advanced networking would be nice17:15
TheJuliaFWIW, the bug which was linked above should be relatively trivial17:19
TheJuliajust requires getting the context and likely adding an extra argument under the hood so we know not to use the requestor context, we've got at least one other place where we've had to do that in the past17:19
JayFoh nice17:27
TheJuliaThe issue is, networking has a few different paths of invocation, and we're grabbing the context and using it when we can17:29
JayFsounds like a perfect place for someone to get familiar with the seams around ironic and neutron :D 17:34
JayF+1 on those release changes17:45
* JayF will be heading out after his meeting in ~hour on a sick day17:45
TheJuliaack ack17:53
TheJuliaI'm eating lunch in a little bit and likely will just take tomorrow off in addition to next week17:54
opendevreviewMerged openstack/ironic master: JSON-RPC: disable server-side logging with rpc_transport=none  https://review.opendev.org/c/openstack/ironic/+/95807218:48
opendevreviewVerification of a change to openstack/ironic master failed: Log how long power sync and sensor collections take  https://review.opendev.org/c/openstack/ironic/+/95576218:48
opendevreviewVerification of a change to openstack/ironic master failed: Add a suggestive warning around power and sensor syncs  https://review.opendev.org/c/openstack/ironic/+/95582118:48
TheJuliasigh18:50

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