Thursday, 2025-06-12

« Wednesday, 2025-06-11 Index Friday, 2025-06-13 »
opendevreviewMerged openstack/nova master: Allow autopep8 to fix more things  https://review.opendev.org/c/openstack/nova/+/94929203:13
opendevreviewMerged openstack/nova master: reorder and extend pre-commit hooks  https://review.opendev.org/c/openstack/nova/+/94929903:13
opendevreviewMerged openstack/nova master: Return HTTP400 for multi spec pci alias if PCI in Placement  https://review.opendev.org/c/openstack/nova/+/95100703:28
opendevreviewMerged openstack/nova master: Multiple spec per PCI alias limitation  https://review.opendev.org/c/openstack/nova/+/94406203:28
opendevreviewMerged openstack/nova master: Validated that PCI alias has proper ids  https://review.opendev.org/c/openstack/nova/+/95112903:28
opendevreviewMerged openstack/nova master: Validate [pci]alias at service startup  https://review.opendev.org/c/openstack/nova/+/95114303:31
opendevreviewMerged openstack/nova master: Cache [pci]alias parsing  https://review.opendev.org/c/openstack/nova/+/95115103:31
opendevreviewRajesh Tailor proposed openstack/nova master: Add support for showing instance-action finish_time  https://review.opendev.org/c/openstack/nova/+/92893306:38
Ugglamelwitt, I think you are right, we forget about this patch.06:46
opendevreviewRajesh Tailor proposed openstack/nova master: Fix instance vm_state during shelve  https://review.opendev.org/c/openstack/nova/+/93429406:58
*** sambork_ is now known as sambork08:40
opendevreviewRajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show  https://review.opendev.org/c/openstack/nova-specs/+/92978010:22
opendevreviewRajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show  https://review.opendev.org/c/openstack/nova-specs/+/92978011:08
opendevreviewRajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show  https://review.opendev.org/c/openstack/nova-specs/+/92978011:47
opendevreviewRajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show  https://review.opendev.org/c/openstack/nova-specs/+/92978012:02
opendevreviewJohannes Beisiegel proposed openstack/nova master: Adds packing_host_pci_numa_cells_allocation_strategy option  https://review.opendev.org/c/openstack/nova/+/95245112:58
opendevreviewBalazs Gibizer proposed openstack/nova stable/2025.1: Add functional reproducer for bug 2102038  https://review.opendev.org/c/openstack/nova/+/95245213:25
opendevreviewBalazs Gibizer proposed openstack/nova stable/2025.1: Return HTTP400 for multi spec pci alias if PCI in Placement  https://review.opendev.org/c/openstack/nova/+/95245313:25
opendevreviewBalazs Gibizer proposed openstack/nova stable/2025.1: Multiple spec per PCI alias limitation  https://review.opendev.org/c/openstack/nova/+/95245413:26
opendevreviewBalazs Gibizer proposed openstack/nova stable/2025.1: Validated that PCI alias has proper ids  https://review.opendev.org/c/openstack/nova/+/95245513:26
opendevreviewBalazs Gibizer proposed openstack/nova stable/2025.1: Validate [pci]alias at service startup  https://review.opendev.org/c/openstack/nova/+/95245613:26
opendevreviewBalazs Gibizer proposed openstack/nova stable/2025.1: Cache [pci]alias parsing  https://review.opendev.org/c/openstack/nova/+/95245713:26
opendevreviewMerged openstack/nova master: api: Add response body schemas for hypervisors APIs (3/3)  https://review.opendev.org/c/openstack/nova/+/93724713:57
opendevreviewMerged openstack/nova master: api: Address issues with hypervisors APIs  https://review.opendev.org/c/openstack/nova/+/95086714:16
stephenfinsean-k-mooney: Every +W'd openapi patch is now merged and gmaan has most of the outstanding patches reviewed. The ball's in your court now if you're happy/eager for us to keep moving on these 🙏14:29
opendevreviewMoritz Wanzenböck proposed openstack/nova master: Implement extend_volume for local devices  https://review.opendev.org/c/openstack/nova/+/87876314:38
sean-k-mooneystephenfin: ack. not today but ill try and do a pass on some/all them tomorrow14:38
sean-k-mooneythats really good progress14:39
stephenfinthanks14:39
stephenfinand yes, great progress14:39
sean-k-mooneyso there are 16 patches i think for nova outstanding. do you stil have another batch on your github once this next bactch is done or is this the full set - any followups needed14:40
sean-k-mooneystephenfin: context im tryign to decied if i think we can get this compleded by m2 or not14:41
sean-k-mooneyit woudl be nice if we can but im not sure how much review time i can dedicate between now and then14:42
stephenfinAh, there's quite a few patches still outstanding. 20 in total, targeting 14 APIs (I've split the servers one up into 6 patches, as you've already seen)14:43
stephenfinWhen is m2? Assuming it's less than ~6 weeks away, that sounds optimistic14:43
sean-k-mooneyso the 16 that are open and ~20 after. ok well i hope we can land the first batch of 16 before m2 and we can then continut with the rest.  its july 3rd i think14:44
sean-k-mooneyso 3 weeks time14:44
sean-k-mooneythat not a deadline by any means i just wanted to knwo were we were since i knew you proviosly had more patches that were pending14:45
sean-k-mooneyi dont knwo what gmaan PTO plans are like and im also aware that they have the policy work to do this cycle so im just trying to guage it we are likely to get this compelte this cyle. i think we are still on track to do that14:46
opendevreviewsean mooney proposed openstack/nova master: [DNM] testing rdo experimental jobs  https://review.opendev.org/c/openstack/nova/+/92776214:47
sean-k-mooneygibi: as an fyi whiel that "works" we get some volume related test failures if i recall corectly14:48
sean-k-mooneyim not really sure why14:48
sean-k-mooneygibi: ^ we also need to clean up the fr2 branch of the operator to remvoe the duplicate definiton of the pragm that may or may not cause that job to fail14:49
sean-k-mooneygibi: if it does nto work as exepcted let me knwo an ill take a quick look14:49
gibisean-k-mooney: thanks. I just told the story about this possibility to artom and then I wanted to see if it still works14:53
gibiI will report back14:53
sean-k-mooneywe have the watcher-operator reporting on changes to master opendev.org prs14:54
sean-k-mooneywe had to swap it to use c10s contianers14:54
sean-k-mooneyso its defintly possibel btu this is actully going to break because we dropped 3.9 supprot14:54
sean-k-mooneygibi: we can chat downstream about how to prot the test change form watcher-operator to nova-operator to make this work if your interested14:55
sean-k-mooneyalthough maybe not this evening14:55
sean-k-mooneythe short version is https://github.com/openstack-k8s-operators/watcher-operator/blob/main/.zuul.yaml#L308-L344 defiens a c10s based job that will run on all changes ot opendev master for watcher14:57
sean-k-mooneywe have both watcher and epoxy version fo those jobs so that we will be able to do some testing on stable/2025.1 to if needed14:57
sean-k-mooneyhttps://github.com/openstack-k8s-operators/watcher-operator/blob/main/.zuul.yaml#L153-L27014:57
sean-k-mooneychandan got all that working in the last month or so, if we want to improve our upstream nova testign we can definetly do that in the future14:58
gibithanks for the pointers15:00
gmaansean-k-mooney: no PTO planned until Sept. for policy, yes that is what I will be pushing now. but I will keep eyes on stephenfin series if anything new or need re-review15:16
sean-k-mooneygmaan: just as a follow up any removal of the tempest policy tests wont happen until 2025.1 goes to unmaintaed right15:17
sean-k-mooneyso it will be removed in 2026.2 or  2027.1 ish15:18
gmaansean-k-mooney: there are no specific policy tests there but idea is to move the tempest tests to move to new defaults (as default) and we have one job testing tempest on old defaults also15:18
sean-k-mooneysorry not policy15:19
sean-k-mooneyi ment the schema validation15:19
sean-k-mooneyso updating tempest to nolonger needign to validate the repsocne once stephens seriss is supported on all stable branches15:19
sean-k-mooneyassuming 2025.2 is the first release with fully coverage we need to keep the nova schema test until 2025.1 moves to unmainted right?15:20
gmaanok. we can discuss that if we need to remove but anyways that can only happen after 2026.1 to unmaintained (if Nova implement all in 2025.2). But we discussed in last PTG, we might keep both side15:21
sean-k-mooneywell once we have the coverage in tree15:21
sean-k-mooneywe shoudl not need to ask people to extend the tempest schdma validation for new changes15:21
sean-k-mooneythat really what im poking at. only doing that in one place, in nova going forwarwad instead o both in nova and tempest15:22
sean-k-mooneybut your right we can discuss that at the next ptg15:23
gmaanI am ok for that but let's discuss during 2026.1 unmaintained status change time. if no objection then I am ok to remove15:23
sean-k-mooneywell its 2025.1 not 2026.115:23
sean-k-mooneyanyway not todays problem :)15:24
gmaannext PTG might be too early, anyways we need to keep them until 2026.1 (after 2025.2 EOL but I am considering 2026.1 as immediate SLURP after Nova has all schema)15:24
sean-k-mooneyright but its not when the first slurp that supprot it that is imporant15:24
sean-k-mooneyit when the last slurp that does not that matters15:24
sean-k-mooneyso its when 2025.1 goies unmaintined that tempest could consider removing the vlaidation15:25
sean-k-mooneywhich will likely be during the 2026.2 or 2027.1 cycle15:25
gmaanyeah, SLURP thing does not need to be mandatory as such. 2025.1 unmaintained is also ok15:27
gmaanI will say let's discuss during 2026.1/2027.1 PTG when 2025.1 is moving to unmaintained15:28
stephenfingmaan, sean-k-mooney: Joining late, but we can/should just copy-paste the schemas from Nova to Tempest15:57
gmaanyeah, if Nova required schema to implement during API change then it is not big deal to keep Tempest one also up to dated15:58
gmaanas Tempest hats on, I can make sure those are syncup and tempest has all schema covered. I already started that for current schema https://review.opendev.org/q/topic:%22latest-microversion-testing%2216:09
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.2: Add functional reproducer for bug 2102038  https://review.opendev.org/c/openstack/nova/+/95247816:45
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.2: Return HTTP400 for multi spec pci alias if PCI in Placement  https://review.opendev.org/c/openstack/nova/+/95247916:45
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.2: Multiple spec per PCI alias limitation  https://review.opendev.org/c/openstack/nova/+/95248016:45
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.2: Validated that PCI alias has proper ids  https://review.opendev.org/c/openstack/nova/+/95248116:45
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.2: Validate [pci]alias at service startup  https://review.opendev.org/c/openstack/nova/+/95248216:45
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.2: Cache [pci]alias parsing  https://review.opendev.org/c/openstack/nova/+/95248316:45
JayFTheJulia: jfyi https://review.opendev.org/c/openstack/ironic/+/949385/6 apparently needs rebasing to land. As long as it's a trivial rebase ping me and I'll reapply +2A16:47
TheJuliaheh, wrong channel :)16:48
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.1: Add functional reproducer for bug 2102038  https://review.opendev.org/c/openstack/nova/+/95248416:48
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.1: Return HTTP400 for multi spec pci alias if PCI in Placement  https://review.opendev.org/c/openstack/nova/+/95248516:48
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.1: Multiple spec per PCI alias limitation  https://review.opendev.org/c/openstack/nova/+/95248616:48
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.1: Validated that PCI alias has proper ids  https://review.opendev.org/c/openstack/nova/+/95248716:48
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.1: Validate [pci]alias at service startup  https://review.opendev.org/c/openstack/nova/+/95248816:48
opendevreviewBalazs Gibizer proposed openstack/nova stable/2024.1: Cache [pci]alias parsing  https://review.opendev.org/c/openstack/nova/+/95248916:48
* JayF has been doing that a lot today16:51
TheJuliano worries16:53
sean-k-mooneyJayF: so on that17:16
sean-k-mooneynot inclduing the MTU in the metadta was specificly request by ironic a number of cycles ago17:17
JayFI meant to tell you: welcome to the claude-code club :) It was nice to see someone not-me using it :D 17:17
JayFSounds like we've corrected a previous mistake, then.17:17
sean-k-mooneywe do it condtionally17:17
sean-k-mooneyif the network has dhcp we dont incldue it and allow it to be managed vi dhcp17:17
JayFJulia is the expert in this bit moreso than I am :)17:18
sean-k-mooneybtu if there is no dhcp enabled we includde it for people doing static ip17:18
sean-k-mooneyack i was just wonderign why its only coming up now it think we didi it 2 or we release ago17:18
sean-k-mooneyTheJulia: context was includign the MTU infroamtion in teh metadat made it much harder to go form ml2/ovs to ml2/ovn becacuce that requried reducing the mtu17:19
sean-k-mooneyso neuton asked us to only set it if there was no dhcp17:20
sean-k-mooneyJayF: on claude-code ya i decied to pay for it for a year and see where the state of the art is vs local llms ectra17:21
JayFsean-k-mooney: so, you want some advice since I've been poking at it for a few weeks? I made a openstack-dev dir, put a claude.md in it, and checked out my other projects under it17:21
sean-k-mooneyso far im likeing the terminal first approch not that i have done a lot with it17:21
TheJuliasean-k-mooney: we've had reports of mtu weirdness and we're seeing a huge increase in not using dhcp so just makes sense to address it.17:21
JayFthe pycharm ide integration is nice, it uses the jetbrains diff stuff17:21
sean-k-mooneyJayF: you mean like this https://github.com/SeanMooney/grian-dev17:22
TheJuliasince, the mismatch between what is known and what an OS can assume otherwise introduces a huge amount of risk17:22
JayFif you wanna use mine as a start https://www.irccloud.com/pastebin/sL4RIxrg/CLAUDE.md17:22
sean-k-mooneyJayF: i also have created https://github.com/SeanMooney/openstack-ai-style-guide17:22
sean-k-mooneywhich i might rename because i want ot evolve that ot be more hten just style17:22
JayFokay I think yours wins for quality lol17:22
JayFbut I will say: the pattern of "start claude in a dir that has all your openstack stuff checked out" has been really nice17:23
sean-k-mooneyi got claude to help write it17:23
TheJuliaplus, cases where dhcp may not even be accessible without network data out of the box and where config drive must be relied upon to get started17:23
JayFI gave it a sample thing during trying it out, pointing it to a feature that'd require implementation in Ironic + modification of hte nova virt driver, and it did a decent job of navigating it17:23
JayFTheJulia: sean-k-mooney: plus the uncomfortable reality that often Ironic knows more about the physical network situation than neutron/nova :( 17:24
sean-k-mooneyTheJulia: so that odd becasue we do still include the mtu info if dncp is disabled on neutron17:24
sean-k-mooneyTheJulia: if we are not then that a regresion we shoudl fix in nova too17:24
TheJuliasean-k-mooney: dhcp only matters if it is reachable out of the box, that is not always true. It is clear cut for VMs, but you can have greater hurdles with physical machines17:24
sean-k-mooneyTheJulia: ah so this is also related to prot groups and bond/vlan netowrking17:25
JayFsean-k-mooney: everything is also related to that :)17:25
JayFsean-k-mooney: there's very few people running ironic in the real world not using at least one of those17:25
sean-k-mooneywell if we know its ironic, which we shoudl be cause it vnic type=baremetal17:25
sean-k-mooneywe may be able ot make ti work better,17:25
sean-k-mooneybut if you have already fixed it on the ironic side17:26
sean-k-mooneythen thats fine too17:26
JayFHaving Ironic own some of the newtork metadata is a feature we'd need for standalone anyway17:26
JayFand gives us more flexibility since (as I said before) we often know more than you all do about the physical world17:26
sean-k-mooneyya that makes sense17:26
TheJuliaYeah, we're seeing more asks for smoothing that general experience so, just makes sense to spot/address the highly specific items to make things better17:26
sean-k-mooneyi also think thie idea of ironic haveign the ablity to generate or enrish the config drive is also useful for that case17:26
TheJuliaSo... 17:27
TheJuliaFunny you should mention that!17:27
JayFand honestly, the more I think about ^^ (the physical world not always being perfectly modelled in nova) is a ... feature not a bug? The entire benefit of ironic+nova integrated is keeping the experience simple ... and the real world is complex as hell lol17:27
TheJuliawe've semi-wondered outloud about "should we offer ssh key injection?!"17:27
sean-k-mooneywell keyparis works with ironic already right17:28
TheJuliaoh yeah17:28
sean-k-mooneybut you mean in standalone17:28
TheJuliayeah17:28
sean-k-mooneyi sort of feel like the keypari really shoudl be in keystone17:28
JayFwe already support it in a supremely roundabout way17:28
sean-k-mooneyand hten we jsut get it form there or in barbican17:28
JayFsean-k-mooney: notably that doesn't help most of our standalone users who use basic auth :) 17:28
TheJuliasean-k-mooney: agree, and that would make it sort of easier17:28
sean-k-mooneybtu legacy17:28
TheJuliayup17:28
JayFyou can do arbitrary file injection in a deploy template17:28
JayFbut that's a lot of crap to do just to get an ssh key in17:28
TheJuliayeah17:29
sean-k-mooneyi mean teh format of the config drive/metadata is pretty standarsied 17:29
sean-k-mooneyso having the ablity for ironic to do that is not a bad thing17:29
JayFTheJulia: the killer potential feature hiding under the surface there: Ironic somehow being able to inject troubleshooting-keys into IPA at boot time :D 17:29
sean-k-mooneywoudl we use it in the combined case proably not but maybe17:29
TheJuliaJayF: I'd have concerns about doing that, but yeah, I've had asks for it in the past17:30
sean-k-mooneyJayF: cant it od that or do you have to buidl them in17:30
JayFTheJulia: I have concerns about the whole thing but I'm assuming we can find a way to have the cake and eat it too :D 17:30
sean-k-mooneyits been a long time since i deployed with bifrost17:30
TheJuliaSo one thing on this entire subject, I'd <3 if nova just shipped us the dicts, but I'd have to double check the constraints and modeling for user input in nova again17:30
sean-k-mooneybut i remeber sshing into the ipa in the past to debug somthing17:30
TheJuliaits been years since I've walked that code path17:30
sean-k-mooneyTheJulia: JayF may rememebr this better then i but i think we just send you json blobs today17:31
sean-k-mooneyi woudl have to look but i dont think we provide a filesytem but i oculd be wrong17:31
TheJuliasean-k-mooney: I don't think y'all actually do17:31
JayFI remember nothing :)17:31
JayFbut we can check!17:31
TheJuliaI think we get handed an iso9660 file right now17:31
sean-k-mooneyoh ok17:32
sean-k-mooneyi mean we can always change that if you come up with a design17:32
TheJuliathe factory which builds it in the overall virt utils, if memory serves just hands us the file pre-formed17:32
sean-k-mooneyi might be conflating this with the auto leasee feature17:32
* TheJulia looked at that relatively recently17:32
TheJuliayeah17:32
sean-k-mooneyso the reaosn i was quistioning it is we currently store it in memcache 17:33
sean-k-mooneyif you turn that one17:33
sean-k-mooneyso we ahve it in a dict like serisable datascuture17:33
TheJuliaIsent there a mechanism on the nova side to do file injection?17:34
sean-k-mooneyyes its deprecated and ahs been for about a decaade or so17:34
TheJuliaoh, good17:34
sean-k-mooneyits called personality files17:34
sean-k-mooneyyou can also pass a user-data cript for cloud init that is what was ment to repalce arbity file injection17:35
TheJuliaokay, as I mentioned, it has been ages17:35
TheJuliain the user-data payload right?17:35
sean-k-mooneyya, either by passing somethign like a valut cred and downloading the afctul data on first boot or inline17:35
sean-k-mooneyit depens on the usecase17:36
TheJuliayup, unfortunately I've done stuff like that long ago17:36
sean-k-mooneyhttps://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L1108-L114117:37
TheJuliaYeah, we don't allow file passing today, but if that is the personality files stuff, maybe since its been deprecated so long it might make sense to drop and refactor?17:38
sean-k-mooneyso righ now we create a tempory file generate the config data gzip it, base64 encode and then decode it and pass that ot ironci 17:38
TheJuliadunno17:38
TheJuliawe only permit a pass-in of meta-data, user-data, and network-data via configdrive api on our provision endpoint17:39
TheJuliaI could have sworn that was structured a little differnetly in nova, but thanks!17:40
sean-k-mooneyit is17:40
sean-k-mooneythis is the ironic customtioation17:40
sean-k-mooneyall the other virt drviers share common code. the ironic driver  uses the common code but then modifyes it before sending to to the ironci api17:41
sean-k-mooneyTheJulia: the ironic driver enriches the network metadta for exmaple https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L1045-L110617:41
sean-k-mooneythe rest use https://github.com/openstack/nova/blob/master/nova/api/metadata/base.py#L11217:42
sean-k-mooneyso we have an objet in memoy thaat can redner subpaths into json for the metadta api https://github.com/openstack/nova/blob/master/nova/api/metadata/base.py#L31717:44
sean-k-mooneyor produce the content for the config drive https://github.com/openstack/nova/blob/master/nova/api/metadata/base.py#L59017:45
TheJuliaYeah, for portgroups. we can likely just rip that out and address it locally in ironic at some point17:45
sean-k-mooneyso it would not take much to jsut add a method to this obejc tthat wil return the data in a singel json dict17:46
sean-k-mooneyfilepath -> file content17:46
TheJuliayeah, that would be fairly straight forward17:46
sean-k-mooneythat acctully what that generator returns already17:46
sean-k-mooneyits just a mateter of passing it to ironic in a way that would work better for ye17:47
TheJuliasomething for some point down the road17:47
sean-k-mooneyyep, we can always imrpove things like that if we know about the pain point and someone has time to od the work17:50
sean-k-mooneyTheJulia: speaning of improment https://review.opendev.org/c/openstack/nova-specs/+/47181517:51
sean-k-mooneyi need to get back to that but tha tis aimed at making the data we pass more useful for the vlan trunking case17:51
sean-k-mooneyalthough that not nessisarly ironic speicific it shoudl be useful for any dirver the support neutron trunk ports17:52
TheJuliaHmmmm17:52
TheJuliaI'll add that to my list to look at when I don't have a migraine17:52
TheJuliaUnfortunately, I woke up with one today :(17:52
sean-k-mooneythat sucks, i often get eystrain headaches or if i forget to take my blodd pressure medication but i fortunate to never had to deal with migraines17:53
sean-k-mooneythat actully is propsoing inlcudign the mtu info for the vlan sub intereface by the way 17:55
TheJuliaoh good, becuase that would be a logical requirement.17:56
« Wednesday, 2025-06-11 Index Friday, 2025-06-13 »

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