Thursday, 2025-11-06

opendevreviewMerged openstack/bifrost stable/2025.2: [stable only] increase test VM mem size  https://review.opendev.org/c/openstack/bifrost/+/96475200:08
opendevreviewMerged openstack/bifrost master: verify creation of test VMs during testenv setup  https://review.opendev.org/c/openstack/bifrost/+/96448500:08
opendevreviewMerged openstack/ironic master: Handle HTTP 400 and 409 race condition in Redfish power operations  https://review.opendev.org/c/openstack/ironic/+/96524800:29
alegacyJayF: ya, that's probably best since gutting it will create a big delta between patchsets which may be confusing.  I'll do that before I upload the new series.  00:56
opendevreviewVerification of a change to openstack/ironic master failed: Add IPMI address validation to node validation  https://review.opendev.org/c/openstack/ironic/+/96445104:33
opendevreviewVerification of a change to openstack/ironic master failed: Add IPMI address validation to node validation  https://review.opendev.org/c/openstack/ironic/+/96445105:57
jandersgood morning Ironic o/07:56
rpittauTheJulia, JayF, dtantsur the real reason why we neglected the bullseye job in ipa-builder is because it has python 3.9 by default, I remember we mentioned it long time ago :)07:58
rpittaudo we want to try and use a more recent python version as we do for CS9 ? not sure it's worth the effort honestly considering that bullseye EOL date was a year ago07:58
rpittauI'd rather abandon bullseye and move to current stable "trixie", keeping compatibility with bookworm as stable-107:58
rpittauand uh.... good morning ironic! o/07:58
dtantsurI concur, it's probably time to drop it (that's pity we could not deprecate it properly)07:59
rpittauindeed, I forgot completely about that :/08:04
dtantsurWe need some sort of a deprecation schedule for IPA-builder and bifrost08:05
rpittauyep08:44
opendevreviewMerged openstack/ironic master: Add IPMI address validation to node validation  https://review.opendev.org/c/openstack/ironic/+/96445108:45
opendevreviewnidhi proposed openstack/ironic master: Add PCIe function fields to redfish inspection  https://review.opendev.org/c/openstack/ironic/+/96317909:03
opendevreviewRiccardo Pittau proposed openstack/bifrost master: Remove tinyipa support  https://review.opendev.org/c/openstack/bifrost/+/96440409:37
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Do not build Debian Bullseye image anymore  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96620410:03
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Use Debian Trixie for debian based ipa ramdisks  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96625310:13
opendevreviewnidhi proposed openstack/ironic master: Add PCIe function fields to redfish inspection  https://review.opendev.org/c/openstack/ironic/+/96317910:51
opendevreviewMerged openstack/ironic-prometheus-exporter unmaintained/2024.1: Update .gitreview for unmaintained/2024.1  https://review.opendev.org/c/openstack/ironic-prometheus-exporter/+/96569611:13
opendevreviewMerged openstack/bifrost unmaintained/2024.1: Update .gitreview for unmaintained/2024.1  https://review.opendev.org/c/openstack/bifrost/+/96566211:42
opendevreviewnidhi proposed openstack/ironic master: Add PCIe function fields to redfish inspection  https://review.opendev.org/c/openstack/ironic/+/96317912:51
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Do not build Debian Bullseye image anymore  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96620413:32
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Use Debian Trixie for debian based ipa ramdisks  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96625313:33
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Do not build Debian Bullseye image anymore  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96620413:36
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Use Debian Trixie for debian based ipa ramdisks  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96625313:37
opendevreviewJakub Jelinek proposed openstack/ironic master: Fix storing inventory and plugin data in Swift  https://review.opendev.org/c/openstack/ironic/+/96625913:37
kubajj^ another bug I have introduced a couple of years ago, sorry13:39
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Use Debian Trixie for debian based IPA ramdisks  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96625313:40
kubajjbtw, is nobody else storing inspection data in Swift?13:45
opendevreviewnidhi proposed openstack/ironic master: Add PCIe function fields to redfish inspection  https://review.opendev.org/c/openstack/ironic/+/96317914:02
TheJuliarpittau: Yeah, time to drop then14:12
iurygregorygood morning ironic14:22
opendevreviewTakashi Kajinami proposed openstack/ironic master: Make sure [cors] allowed_origin accepts a list value  https://review.opendev.org/c/openstack/ironic/+/96627114:25
opendevreviewTakashi Kajinami proposed openstack/ironic master: Make sure [cors] allowed_origin accepts a list value  https://review.opendev.org/c/openstack/ironic/+/96627114:26
rpittaugreat, we're using 2025.1 instead of 2025.2 in the bifrost upgrade jobs mgrrrrr15:05
opendevreviewRiccardo Pittau proposed openstack/bifrost master: [CI] upgrade from 2025.2  https://review.opendev.org/c/openstack/bifrost/+/96628315:11
rpittausome please review when you got a moment, thanks! ^15:11
opendevreviewRiccardo Pittau proposed openstack/bifrost master: Remove tinyipa support  https://review.opendev.org/c/openstack/bifrost/+/96440415:12
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Do not build Debian Bullseye image anymore  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96620415:16
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent-builder master: Use Debian Trixie for debian based IPA ramdisks  https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/96625315:16
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent master: Test advertised ip reachability before assigning it  https://review.opendev.org/c/openstack/ironic-python-agent/+/96367015:28
JayFhttps://review.opendev.org/c/openstack/metalsmith/+/965736 does anyone on our end care about um/2024.1 metalsmith?15:44
JayFif not we should probably get it retired to handed off to um team15:44
opendevreviewRiccardo Pittau proposed openstack/ironic stable/2025.2: Handle HTTP 400 and 409 race condition in Redfish power operations  https://review.opendev.org/c/openstack/ironic/+/96629015:52
cardoekubajj: I'm not storing inspection in swift. Just in the ironic DB.15:52
dtantsuralegacy: btw, I've just had an enormous success asking Claude to turn really messy 4 commits into 2 nicely deliniated ones. I'm quite sure it can be used to split a large commit along a certain line.15:53
JayFclaude is really, really good at stuff you can give examples for too. Like implementing existing interfaces and the like.15:54
JayFI cannot get it to understand how to write a consistent import statement ever, but I guess I need something to do too :P 15:54
dtantsuryeah, it took it a while today to sort Go imports :D15:54
dtantsurbut I'm really happy that I did not have to refactor the commits myself15:55
opendevreviewJakub Jelinek proposed openstack/ironic master: Fix storing inventory and plugin data in Swift  https://review.opendev.org/c/openstack/ironic/+/96625915:58
JayFhttps://review.opendev.org/c/openstack/ironic-python-agent/+/966031 IPA Backport that is simple and could use +2A15:59
opendevreviewRiccardo Pittau proposed openstack/ironic stable/2025.1: Handle HTTP 400 and 409 race condition in Redfish power operations  https://review.opendev.org/c/openstack/ironic/+/96629416:00
kubajjcardoe: isn't the DB massive then? (Our inspection data is roughly 6x as big as the DB)16:00
JayFI'm landing https://review.opendev.org/c/openstack/ironic-specs/+/965865 (work items + new template). Happy to do follow-ups but we need to get our PTG information published.16:00
rpittauthanks for taking care of that JayF :)16:01
JayFI also emailed ptg@ with a photo and count16:01
JayFI'll email about the PTG content once it publishes, then I'm going to record a video summary of the work items post to likely go up on youtube next week16:01
JayFBTW if at least one other -approver would comment on the ML with a +1 to our core team changes that'd be appreciated16:04
opendevreviewRiccardo Pittau proposed openstack/ironic stable/2024.2: Handle HTTP 400 and 409 race condition in Redfish power operations  https://review.opendev.org/c/openstack/ironic/+/96629516:04
zigoAm I right that "openstack baremetal node create" would go in here:16:05
zigohttps://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L297716:05
zigoIf so, how can I tell that I'm in a "node create" request when I'm there?16:05
JayFThis is the most XY question ever asked; what's your high level goal? :D 16:05
JayFhttps://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L2890 POST to /v1/nodes is always a node create16:06
zigoOh, thanks ! :)16:06
JayFhttps://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L3123 patch is for update16:06
JayFwhat is your goal though?16:06
JayFthat's a weirdly oriented question16:07
zigoJayF: I want to transform the api.request.rpcapi.get_topic_for() into a api.request.rpcapi.get_topics_for() (ie: with an s) so that it would iterate on all nodes of the conductor group so it would create the Grub PXE config on all conductor of the group, just like we discussed ...16:07
JayFAh. Hmm. That's not how I would've expected that to be oriented, but I haven't thought much about this so :) 16:07
zigo:)16:07
JayFIf that's really the solution, it's certainly going to be less code than the paths I was thinking16:08
zigoIf I'm right, the code is searching for the conductor that the node is attached to, to send RPC call to the correct, topic, right?16:08
JayFI'd have to look in real detail. The real reason I'd be surprised to see it there is that often the conductor handles the meat of the requests16:09
JayFso I assumed it'd be something like a conductor<>conductor RPC calls to fan them out16:09
JayFbut if you can do it in that layer, maybe that's OK?16:09
zigoGot no clue ... :)16:09
JayFso I think that's more a precursor to making an RPC call16:09
zigoFirst time I'm really trying to contribute to Ironic.16:09
JayFI don't think we have any modeling for "make an RPC call for each" but IMBW16:10
zigoDo you know where's located the code that recieves the create_node on the conductor side?16:10
zigoYou must be right that it should be doing more than just the PXE stuff, and that my approach must be wrong.16:11
JayFconductor/rpcapi.py conductor/manager.py16:11
zigoThanks.16:11
JayFthe usual pattern is rpcapi is where the rpc api calls are16:11
JayFmanager is usually the actual code that gets executed16:12
JayFI'm trying to think if we have any endpoints currently which would ping all the conductors16:12
JayFand I don't think we do16:12
zigoI'll contribute it... :P16:12
zigoOh, there's a def create_node() in the manager indeed ! :)16:12
opendevreviewMerged openstack/ironic master: Apply naming standards to Ironic install guide sections  https://review.opendev.org/c/openstack/ironic/+/96448616:13
opendevreviewMerged openstack/ironic-specs master: Add work items template, including maint checklist  https://review.opendev.org/c/openstack/ironic-specs/+/96524916:15
zigoJayF: I suppose what I should look at is in ironic/objects/node.py then?16:17
zigoOr that's just the db side of things?16:18
zigo(it looks like it)16:18
JayFI don't know how this would be done16:18
zigoSorry to ask dumb question, but it's hard to dive in ...16:18
JayFand won't without significant digging16:18
zigoI'll dig myself then, thanks.16:18
JayFbut conductor manager is more likely to be the right layer16:18
opendevreviewMerged openstack/ironic-specs master: 2026.1 Ironic work items  https://review.opendev.org/c/openstack/ironic-specs/+/96586516:23
opendevreviewMerged openstack/sushy stable/2025.2: Check required credentials in a detailed way  https://review.opendev.org/c/openstack/sushy/+/96487116:38
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent master: Test advertised ip reachability before assigning it  https://review.opendev.org/c/openstack/ironic-python-agent/+/96367017:00
gmaanis this job renamed/removed ironic-tempest-bios-ipmi-direct-tinyipa17:31
gmaanwe used to run this on tempest and devstack gate to avoid any breaking change for ironic17:31
gmaanbut I see it is not running and i cannot find its definition 17:31
gmaanyeah, done in https://review.opendev.org/c/openstack/ironic/+/950192/12/zuul.d/ironic-jobs.yaml17:37
gmaanI think zuul should give some error or warning about jobs usage so that renaming/removal can be more known than just ignore to run the job if not found17:37
* dtantsur agrees17:37
gmaanI just checked today tempest gate and found no ironic job running and it could go unnotice for long17:38
gmaananyways, I will update tempest/devstack gate with new job name17:38
opendevreviewVerification of a change to openstack/sushy stable/2025.1 failed: Check required credentials in a detailed way  https://review.opendev.org/c/openstack/sushy/+/96487217:40
opendevreviewMerged openstack/ironic stable/2025.2: Remove bespoke logic for handling redirects while validating URLs  https://review.opendev.org/c/openstack/ironic/+/96602717:42
opendevreviewJay Faulkner proposed openstack/ironic-specs master: Add 2026.1 workitems to index  https://review.opendev.org/c/openstack/ironic-specs/+/96632217:49
JayF^ is trivial, I accidentally an index on that ptg work items 17:52
TheJuliagmaan: dtantsur: it used to but that changed maybe 2 years ago in the zuul config validation code18:10
JayFgmaan: That explains how a neutron change in devstack broke us unexpectedly last cycle. Thank you for that \o/18:17
TheJuliaCome to think of it, I think we've checked with codesearch as well on whole job removals and it didn't pop up. Guess zuul erroring would actually kind of be more preferable. :\18:19
gmaanyeah, I think at least if it should be zuul error if any of the jobs is missed to run due to any reason. ignoring those jobs to run and gate result as pass is false result18:21
alegacyTheJulia: Jayf isn't keen on the "standalone-experimental" network interface name.  The others are named noop, flat and neutron.  How about just "ironic" or "ironic-standalone" since this is an Ironic alternative to the Neutron-based version of networking?  I'm open to suggestions.18:44
TheJuliaI was thinking standalone originally, or maybe "ironic" could be simple enough18:45
JayFjust ironic is nice, or something like "mercury"18:45
TheJuliaheh, that would be more confusing most likely18:45
JayFif the service name is ironic-networking18:50
JayFcalling it ironic-networking is also on the board18:50
JayFI'll note my primary beef with "standalone-experimental" is the "-experimental" as putting a temporal state into an operator/config/api facing string seems not awesome18:50
TheJuliayeah, I concur with that and I would have preferred something simpler/shorter out of the box. I don't think it was me who demanded -experimental either18:51
alegacythat was my suggestion because there was a worry about being able to delete this driver and move to a newer version once we firmed up the API without going thru deprecation.18:52
alegacy"ironic" is simple.  I like it.18:52
TheJuliaI think the rest API primitives and all are what remains to be figured out18:53
TheJuliaand that will likely just keep a fall back to the old value locations as time moves on18:53
alegacy"ironic-networking" is also a good choice.18:54
TheJulia"internal"18:54
alegacyJayF: Do you have a favorite?  I think my picks (in order) are: "internal", "ironic-networking", "ironic"18:59
JayFoooooh internal19:02
JayFI like ironic-networking, I think19:03
JayFit's the most clear and you can easily document it 19:03
JayFmaking it the same name as the service should IMO remove a decoder ring19:03
JayFbut to be clear: if any of those were the original name, I don't make the review comment :)19:03
alegacyok, how about "ironic-networking" as a winner.19:07
opendevreviewJacob Anders proposed openstack/ironic master: [WIP] Make post-firmware-update reboot conditional on component  https://review.opendev.org/c/openstack/ironic/+/96634419:08
opendevreviewJacob Anders proposed openstack/ironic master: [WIP] Make post-firmware-update reboot conditional on component  https://review.opendev.org/c/openstack/ironic/+/96634419:15
cardoehttps://github.com/dell/iDRAC-Redfish-Scripting/blob/master/Redfish%20Python/DeviceFirmwareMultipartUploadREDFISH.py is completely generic. Has anyone played around with adding a "push" version of firmware uploading?19:22
TheJuliaI know janders has been trying to get it standardized19:24
TheJuliaacross vendors, that is19:24
cardoezigo, JayF: So we're gonna make the grub config on all conductor nodes... are we then gonna generate the custom ISO for virtual-media on all of the conductor nodes?20:39
zigocardoe: No, that's not the idea. The idea is: when one doesn't control DHCP, and we still want HA, then one may setup 2 conductors in a single conductor group, with a VIP. In such environment, we want the Grub config to be written on both nodes, every time, so PXE can be done on any of the 2 nodes, anytime.20:41
zigoThough I've spent hours this afternoon reading Ironic code, and couldn't figure out where to change it... :(20:42
JayFYeah, I don't think you need a small change, this is an entirely new concept, that's what I was trying to point at earlier21:01
cardoezigo: so essentially when we PXE boot like boot.ipxe it would be on all the conductors21:25
TheJuliaIts a change in the driver code which launches a new task on the remote conductor21:40
TheJuliasome of the building blocks are there21:41
opendevreviewDoug Goldstein proposed openstack/ironic master: fix: local_link_connection inspection hook does not fail on missing port  https://review.opendev.org/c/openstack/ironic/+/96637322:16
opendevreviewDoug Goldstein proposed openstack/ironic master: fix: local_link_connection inspection hook does not fail on missing port  https://review.opendev.org/c/openstack/ironic/+/96637322:17
cardoeokay great I can now sit down to start working on what I wanted to do first thing this morning... hmm what time is it? 4:46pm.22:46
TheJulia... I know this feeling22:48
JayFI thoguht I had hit that point, then realized I have a giant (OSS but not OpenStack) PR to review :( 22:48

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