Friday, 2026-04-17

opendevreviewChungWon Lee proposed openstack/nova master: Add regression test for bug #2134375  https://review.opendev.org/c/openstack/nova/+/98490000:27
opendevreviewSam Morrison proposed openstack/nova master: Filter out deleted instances when looking for build timouts  https://review.opendev.org/c/openstack/nova/+/88012502:58
opendevreviewchanyeol yoon proposed openstack/nova master: Fix infinite WARNING loop in _reclaim_queued_deletes  https://review.opendev.org/c/openstack/nova/+/98506806:57
opendevreviewKamil Sambor proposed openstack/nova master: Enable threading mode for proxy services  https://review.opendev.org/c/openstack/nova/+/97608907:55
opendevreviewKamil Sambor proposed openstack/nova master: Test nova CLI commands with native threading  https://review.opendev.org/c/openstack/nova/+/98403609:31
Ugglasean-k-mooney, as discussed I will prepare the pci grouping topic because that's an internal high prio topic.09:43
sean-k-mooneythanks i have not really had tiem to propsoe it bvut that partly becasue i dont really knwo what if any the open questiosn are09:44
sean-k-mooneythe spec has obvioulsy been up for a while and its jsut a refinement of the previoulsy appoved spec09:44
sean-k-mooneyso i dont really know what need ptg discussions vs what just need gerrit reivew09:45
bauzasUggla: sean-k-mooney: IMHO we should be discussing this in the cyborg-nova x-p session09:45
sean-k-mooneynope09:46
sean-k-mooneythi is not a cyborg related topic09:46
sean-k-mooneythis is a pure nova proposal and even if cyborg may have simialr functionaltiy in the future it is indepent of that09:46
gibiit is hard to say what are the open question before loading context on the PCI grouping topic. I'm doing that loading now09:46
sean-k-mooneygibi: thanks ++09:47
sean-k-mooneythis is what im suggesting for the nova-cybrog topics https://etherpad.opendev.org/p/cyborg-2026.2-ptg#L29409:48
sean-k-mooneybut honestly i feel like that already too much09:48
gibiUggla: on a related note, I found hints in the pci/stats code that also assumes that VFs under the same parent are equivalent https://github.com/openstack/nova/blob/525158521281cf643a9942522788efd15ba12d6b/nova/pci/stats.py#L164-L169 so this code need to be looked at when we want to remove the restriction of one RC for all VFs under the same PF09:48
Ugglasean-k-mooney, Could you send me the link to the spec if you have it handy?09:49
sean-k-mooneysure ill go grab it09:49
sean-k-mooneyhttps://review.opendev.org/c/openstack/nova-specs/+/97563509:49
UgglaI agree the cyborg content seems already too large. So definitively need a session for PCI grouping.09:50
sean-k-mooneyweired in my brian i pushed that in like december not febuary09:50
sean-k-mooneyi guess because i sated on it just before i went on pto and then picked it up agan breifly in late jan early febuary09:51
Ugglasean-k-mooney thx09:51
sean-k-mooneythe other spec https://review.opendev.org/c/openstack/nova-specs/+/967515 will be taken over by joan thaat is the first cyborg sepcifc spec for nova for this cycle09:52
sean-k-mooneyim not sure if we will have another09:52
sean-k-mooneyit kind of depend on which feature/integration we priorites09:52
UgglaAlso, I’m likely lacking some knowledge on PCI grouping, but I’m counting on all of you to help me out. <309:55
bauzasstephenfin: thanks a lot for your proposals ! Sorry I preferred to have some bootstrapped files but sure I'll review your patch and accept it once I doublecheck about the differences between the markdown file and the restructured one09:57
sean-k-mooneyin essence its just a way to say "provision these togeher and schdule them as a group"  that the core idea, today you can have man pci alias request in a single flavor but they alwas provide you 1 device per aplis, pci grouping is just a way to say thses specific set of device are member of a group and i would like 1 of that type of group in a concise way09:57
sean-k-mooneyeverything beyond that and the usecase that enabels are the details but that is the 10,000 foot view09:58
stephenfinbauzas: Thanks. Can I get core on that too, please? I think sean-k-mooney is after it also09:59
sean-k-mooneyyes i would like to activly be inovled in the development fo that repo10:00
stephenfinor better, just add various `*-core` projects for now, which we can extend with non-cores (ops, other end-users) over time10:00
stephenfin*`*-core` groups from projects10:00
sean-k-mooneyim also reuickly reivewing the scafold patch10:01
stephenfinsean-k-mooney: it's already merged :(10:01
sean-k-mooneyok i can propsoe a revert :P10:01
sean-k-mooneyjust kidding10:01
gibisean-k-mooney: where I feel complication (and digging into) is that today we schedule based on per device parameteres like live-migratable, traits, device_type, numa, etc. But when you have a group of such devices these parameters are not always easily inherited from a diverse set of device to a single group the devs are in10:01
stephenfinbauzas slightly jumped the gun, I'm afraid10:01
sean-k-mooneybut i want to get some agreement on the repo layout ectra before we go too far10:01
stephenfin(I understand his rationale, but it was still under very active review)10:02
sean-k-mooneyi was going to work on some liniting/precommit tooling too10:02
sean-k-mooneyif this is going to be a real repo it need at least basic zuul ci10:02
sean-k-mooneyand that shoudl be doen very early10:02
stephenfintrue10:05
gibisean-k-mooney: stephenfin: which repo are you talking about? :)10:09
stephenfinhttps://review.opendev.org/q/project:openstack/agentic-workflows10:09
sean-k-mooneyhttps://opendev.org/openstack/agentic-workflows10:09
sean-k-mooneywe have a lot of thing to agree on like rst+sphink or markdonw ectra before we start commiting content there10:10
gibiohh, is it a grassroot movement, or was there already a decision on the tc / community level to have it?10:12
sean-k-mooneyit was discssed on the list and propsoed to buidl it under the tact sig10:12
sean-k-mooneybut litrlly in the last day10:12
gibivery strange move before the PTG10:12
sean-k-mooneywhich is why i was not expecting to really merge any content ther until after it was more widely dicssed at the ptg10:12
sean-k-mooneyright10:12
stephenfinyeah10:13
sean-k-mooneyi didnt expect either the repo to be create or content to be merged untile after the tc slot10:13
sean-k-mooneyi fine with propsoing some review ectra10:13
gibiI'm affraid that the content in this repo is seens as the communities agreement about this topic10:13
bauzassorry I was off for 15 mins due to a taxi need10:17
bauzassean-k-mooney: stephenfin: I had to merge some things but we can change anything for sure10:18
sean-k-mooneywell you not really10:18
sean-k-mooneyit could have statyed in gerrit10:18
stephenfinyeah that's what I said šŸ˜…10:19
sean-k-mooneybut ya we can change things10:19
bauzasstephenfin: sean-k-mooney: I'm just adding you both as cores10:20
sean-k-mooneyi need to also go do taxi for "Freya" since my frid does not want to walk in the rain with the dog.10:20
sean-k-mooneybecause of course teh  "dog" does not want to walk :P 10:20
bauzasstephenfin: sean-k-mooney: added you both10:21
bauzasstephenfin: now reviewing your change10:21
stephenfinty šŸ™10:22
bauzasstephenfin: 2 nits and I'm cool10:26
bauzashttps://review.opendev.org/c/openstack/agentic-workflows/+/98507810:26
bauzassean-k-mooney: sorry, as I said with my +W comment,  it was a chicken-and-egg issue, but please review https://opendev.org/openstack/agentic-workflows/src/branch/master/AGENTS.md because of your concerns about the directories10:31
bauzasstephenfin: sean-k-mooney: gibi: would you prefer I revert my merged changes actually ?10:33
stephenfinbauzas: thanks, done10:35
gibibauzas: I prefer to have proper discussion on the PTG10:36
stephenfinRE: changes, I don't see the point of reverting now, but I'd certainly hold off merging any skills yet. PTG is next week: let's wait until after that10:36
gibibefore sending an official looking message via a repo in the openstack namespec10:36
gibinamespace10:36
bauzasI'm like stephenfin here10:36
bauzasI think we can stop merging anything but the AGENTS.md and the README.md 10:37
stephenfinmaybe worth adding a note to the README(.md) that this is not an official project also10:37
bauzasso then we can discuss about the skills10:37
bauzasstephenfin: can do that10:37
bauzasit would only become "official' once https://review.opendev.org/c/openstack/governance/+/984958 is accepted10:37
bauzas(yet another chicken-and-egg issue)10:38
stephenfingibi: bauzas: https://review.opendev.org/c/openstack/agentic-workflows/+/98508210:41
bauzasstephenfin: +Wd both patches (and we need some kind of core policy :-) )10:45
bauzasfor the moment, we bootstrap only README and AGENTS.md but for sure we will need to find a way to have consensus :)10:45
sean-k-mooneybauzas: you know im stilll reviewing the first pathch you merged right10:45
sean-k-mooneyin rst format10:45
sean-k-mooneyim not agianst movign to markdon but i woudl like time to actully leave my feedbac10:46
bauzasdoh sorry10:46
bauzassean-k-mooney: fwiw, stephen removed some sections from the README so I guess it's more important to review the AGENTS.md file now10:48
sean-k-mooneywell the AGENT.md is hopelfully very short10:50
sean-k-mooneybut i was still on the readme ill look at that next10:51
bauzassean-k-mooney: there is a rule about not having more than 500 lines for both skill.md and AGENTS.md :)10:57
sean-k-mooneyok i dissagre on both :)10:57
sean-k-mooneyskills can be longer then AGENTS.md10:57
sean-k-mooneybecuase only frontmatter is loaded into the context window form a skill initally10:58
sean-k-mooneyso the overall leght is only an issue if the skill is actully used10:58
sean-k-mooneywhere as AGENT.md is incldue in every tool call10:58
sean-k-mooneyso it need to be smaller10:58
sean-k-mooneyi dont want to perscibe an actual number right now10:59
sean-k-mooneybut those shoudl be trated differntly10:59
bauzassean-k-mooney: I'm just advocating based on the agentskills best practice :) https://agentskills.io/skill-creation/best-practices#structure-large-skills-with-progressive-disclosure10:59
bauzasanyway, I don't want to be that prescreptive, fwiw the AGENTS.md is pretty small atm, and no skills won't be accepted before we officially agree on the direction11:01
sean-k-mooneyhttps://review.opendev.org/c/openstack/agentic-workflows/+/985021/1/README.rst has my initall comments11:04
sean-k-mooneybauzas: wether we use tox or not was goint to be one of the topic i watned to dicss by the way11:05
bauzasI just wanted a small linter on check 11:05
sean-k-mooneyif we expect to have python script and or unit test then it makes sesne11:05
bauzashence the use ot tox11:06
bauzassean-k-mooney: and thanks for the reviews11:06
sean-k-mooneyif we dont then we may want to condier other options11:06
bauzasthe current is that we use a tox target in check and gate for checking the frontmatters using the 'agentskills validate' recommended tool11:07
bauzashttps://agentskills.io/specification#validation11:09
sean-k-mooneyyep i just saw but again we shoudl dicuss those ways of working first11:10
stephenfinsean-k-mooney: can you re-ack https://review.opendev.org/c/openstack/nova/+/984918?11:24
stephenfinand bauzas, ditto for https://review.opendev.org/c/openstack/nova/+/984956/ (I'd to fix some typos and a comment from dan)11:25
bauzassean-k-mooney: well, I don't see the controversy for the frontmatter linter11:25
bauzassean-k-mooney: what's your concern with running a tox target with it ? that's an easy win to me11:25
sean-k-mooneywell the issue is mroe that you merged anything before the ptg into this repo11:25
bauzasnone of that is skills11:26
sean-k-mooneyi dont think you shoudl have merge anything11:26
sean-k-mooneyat all11:26
sean-k-mooneyas i said creating the reviews is totlaly fine11:26
stephenfinIt's done. Let's keep reviewing now and gather thoughts for the PTG11:27
sean-k-mooneyyep that what im focusing on11:27
bauzasagain I apologize, I didn't wanted to force anything but trying to just have a boostrap that people can review11:27
bauzasI'm cool with reverting for example the tox patch but I found this was not really controversial11:28
stephenfinI'll repeat šŸ˜„ It's done. Let's keep reviewing now and gather thoughts for the PTG. There's a lot more people that are going to have opinions on this than us11:29
sean-k-mooneyits not the content of hte patch that was the issue it was the lack fo time to provide feedback11:29
bauzassean-k-mooney: I think, the easiest if you want is for me to revert https://review.opendev.org/c/openstack/agentic-workflows/+/985027 which was at the top11:49
bauzasand then I'll revert the revert so we can discuss about the AGENTS.md in the new gerrit patch if that's simplied11:50
sean-k-mooneylets not jump right to that11:50
bauzasnow the README.md is pretty small11:50
sean-k-mooneyas steph said its merged but there are large part of it that i want to review in detail and refine11:50
bauzasso I guess we definitely need to discuss about te structure11:50
sean-k-mooneyyes i have a very diffent vission from you of how this repo should be stucureed11:51
sean-k-mooneyadn how the code-review skill shoudl work11:51
sean-k-mooneyfor exampels i do not think we shoudl have any personas in the repo11:51
bauzassean-k-mooney: so if you don't me to revert, please provide a new change for changing what you want11:52
sean-k-mooneyi will but im still reviewing all the open chnages11:52
sean-k-mooneyto do what stephen suggested11:52
sean-k-mooneywhich is building a set of topic to dicusss11:53
bauzassean-k-mooney: stephenfin: want me to create another IRC channel for the project ? I don't want to hit the nova channel to be a lot of discussions for the project and ralonsoh could be join as well12:21
stephenfinprobably premature before next week12:21
sean-k-mooneyam we can12:21
sean-k-mooneybut ya lets wait12:21
sean-k-mooneyi mean we can create it  if you want but i porably have already spent more tiem on this then i should have since i have things to prepare before the ptg12:25
bauzasI just joined #openstack-agentic-workflows so you can join too, but agreed, I wouldn't add it in openstack/project-config until next week12:27
bauzasI'd rather prefer us to discuss there 12:27
ralonsohbauzas, I'm ok with this. We can also, if needed, create a weekly meeting12:38
ralonsohbut that could be discussed in the PTG, in any case12:38
ralonsohbut yes, the channel would be perfect12:38
gibiis this agentic-workflows something that should be a TC driven cross project goal? I think so. But then we need to agree on what is in scope of that goal and propose that in the governance repo12:39
bauzasgibi: see https://review.opendev.org/c/openstack/governance/+/98495812:39
bauzasthat wouldn't be a TC goal, rather a TaCT SIG new project12:40
gibiack12:40
bauzasralonsoh: you can join the channel as I did, we can discuss there until we officially add it in project-config12:40
gibibut that patch does not seems to define the scope12:41
bauzasgibi: there is no real scope per say12:41
bauzasscope is about adding skills and subagents related to OpenStack12:41
gibiand is that good that we are going in without some scope definition?12:41
gibihow do we check if we are going to the right direction, or when we are done?12:42
ralonsoh#join project-config12:42
bauzasbut the project doesn't have scope about how to use it by upstream 12:42
ralonsoharggggg12:42
bauzasralonsoh: nah, join #openstack-agentic-workflows12:42
ralonsohyeah hehehe12:42
gibiscope also helps judging contributions about is it in scope? is it out of scope?12:42
bauzasgibi: there aren't yet conventions about judging contributions for it atm, but there are already some guidelines from agentskills about evaluating skills that I'd like us to do https://agentskills.io/skill-creation/evaluating-skills12:45
gibiOK so one of the implicit scope / goal item is to apply https://agentskills.io/skill-creation/evaluating-skills going forward. 12:47
gibican we collect these implicit things and make them explicit12:47
opendevreviewKamil Sambor proposed openstack/nova master: Test nova CLI commands with native threading  https://review.opendev.org/c/openstack/nova/+/98403612:47
bauzasgibi: the problem with skill evaluations is that you need to be able to use LLMs for evaluating them12:48
bauzasso we can't run CI upstream jobs atm12:48
bauzasI added this item for the TC topic we'll discuss next week12:48
gibiyou can define goal and scope regardless of if an item in the goal is CI verifiable or not. j12:49
bauzasgibi: true13:13
opendevreviewribaudr proposed openstack/nova master: Add reproducer for bug #2141722  https://review.opendev.org/c/openstack/nova/+/98452113:24
opendevreviewribaudr proposed openstack/nova master: Fix os-simple-tenant-usage pagination with duplicate instance_extra  https://review.opendev.org/c/openstack/nova/+/98452213:24
opendevreviewribaudr proposed openstack/nova master: Add reproducer for bug #2111320: soft-deleted services in Instance.services  https://review.opendev.org/c/openstack/nova/+/98267513:50
opendevreviewribaudr proposed openstack/nova master: Filter soft-deleted services from Instance.services ORM relationship  https://review.opendev.org/c/openstack/nova/+/98267613:50
Zhan[m]Hi friends, I have two quick live migration related improvements for review, wondering if I can get some quick feedback on the ideas? Thanks! https://review.opendev.org/c/openstack/nova/+/978255, https://review.opendev.org/c/openstack/nova-specs/+/97631113:54
stephenfinsean-k-mooney: just a reminder about https://review.opendev.org/c/openstack/nova/+/984918, in case you got distacted by the agent stuff :)13:57
stephenfin(if it's already on your list, sorry for the noise)13:58
opendevreviewStephen Finucane proposed openstack/nova master: Fix functional tests and mypy on macOS  https://review.opendev.org/c/openstack/nova/+/93772714:12
winiciusallan[m]folks, I saw a lot of discussion on the mail list about a spot instance service in Nova, some people have even start implementing, but this wasn't go further14:55
winiciusallan[m]CERN has a external service for this https://gitlab.cern.ch/cloud-infrastructure/aardvark14:56
winiciusallan[m]do you know by chance if having this kind of service integrated into Nova is interest for the community?14:56
winiciusallan[m]i'm thinking of implementing this as part of my final thesis. this will make other research regarding billing and charge strategies easier to evaluate15:00
opendevreviewMerged openstack/nova master: tox: Remove override of install_command  https://review.opendev.org/c/openstack/nova/+/98491815:14
opendevreviewMerged openstack/nova master: docs: Update development-environment guide  https://review.opendev.org/c/openstack/nova/+/98495615:14
sean-k-mooneywiniciusallan[m]: we decieed that it should not be in nova15:42
sean-k-mooneywe reivwed the sepec and went back and forth but decieed it shoudl live as its own thing external to nova15:43
sean-k-mooneywiniciusallan[m]: i have tought that this could live in watcher in the future15:43
sean-k-mooneythere is also a host based version in blazar today15:43
sean-k-mooneybut i think watcher can facilatate a more granular approch similar to the orgianl concept15:44
winiciusallan[m]sean-k-mooney: yeah, thinking about watcher it looks more natural considering the architecture of the service15:49
winiciusallan[m]I dropped a message on scientific-sig asking if people have use cases or gaps that they would like to improve15:50
winiciusallan[m]one from NASA talked about GPUs being expensive and it might be a use case for them  15:51
sean-k-mooneyi dont know if i ever wrote up a spec for it there. but the idea was to annotate the isntance via the flavor or instance metadta wiht lifecycle:premtable=ture15:51
winiciusallan[m]others from CERN and nectar-rc use aardvak15:51
sean-k-mooneythen have a new goal for permation and 1-n stragies for how to dop it15:52
winiciusallan[m]sean-k-mooney: +1 with this approach. this is how the first spec was written on nova15:52
sean-k-mooneyso the firction with doing this is nova is  https://docs.openstack.org/nova/latest/contributor/project-scope.html#no-more-orchestration and https://docs.openstack.org/nova/latest/contributor/project-scope.html#no-more-orchestration15:53
winiciusallan[m]probably different strategies for how to find the best instance to preempt and also being possible to extend a custom strategy15:53
winiciusallan[m]sounds good15:53
sean-k-mooneywe have basicly said we do not want to build complex orchstration logic in nova or optimise for batch processing 15:54
sean-k-mooneybut we want to make sure that higher level service can express the policy on top15:54
sean-k-mooneywhich is where aardvark or watcher can play a role15:54
sean-k-mooneywiniciusallan[m]: ya so there are a lot of intersting way to decie which instnace to preempt and what that means  (either delete or shelve) 15:55
sean-k-mooneyfor example delete oldest first15:55
sean-k-mooneydelete bigest first15:55
winiciusallan[m]best fit and so on...15:56
sean-k-mooneydelete instnace on expensive hosts or where there is pressure based on mettrics15:56
sean-k-mooneyso watcher with its plugable goals and stragies15:56
sean-k-mooneywoudl allow you to experiment with those15:56
sean-k-mooneywiniciusallan[m]: https://redhat.atlassian.net/browse/RHOSRFE-142 is a downstream tracker i filed to do this "eventualy"15:58
sean-k-mooneybut its not someting i have had time to activly work on and i dont expect to actully do that any time soon15:58
winiciusallan[m]i'll be happy to help you with this if it is possible15:59
sean-k-mooneyhttps://etherpad.opendev.org/p/r.bf5f1185e201e31ed8c3adeb45e3cf6d#L74716:00
winiciusallan[m]I need to sync from my side what is the scope for my thesis, there is a slighly chance of testing different strategies...16:00
sean-k-mooneywell if you want to talk about this more in the context of watcher feel free to reach out in #openstack-watcher16:00
sean-k-mooneywe are also haveing the PTG next week so we could chat about it more there too if you were interested16:00
sean-k-mooneywiniciusallan[m]: no worries either way, that the most recent context i can provide on this topic so if you do or do not have scope to wrok on this either is fine16:02
winiciusallan[m]at a first moment, i was thinking of discussing this on nova ptg, but since answered me in advance we can discuss this on watcher's ptg16:03
winiciusallan[m]let's take the discussion to #openstack-watcher16:03
winiciusallan[m]joining there16:03
opendevreviewStephen Finucane proposed openstack/nova master: Fix functional tests and mypy on macOS  https://review.opendev.org/c/openstack/nova/+/93772716:27
gmaanbauzas gibi console proxy in threading mode change lgtm now https://review.opendev.org/c/openstack/nova/+/976089 17:04
gibigmaan: ack thanks. Maybe next week before the PTG calls I will have time to get to it17:26
gibiUggla: sean-k-mooney: I filled out the details of the PCI Grouping topic in the etherpad 18:15
sean-k-mooneygibi: ++18:16
gibiand that is my exit today18:16
sean-k-mooneythanks18:16
gibibtw, does AI crawlers know how to read an etherpad?18:17
sean-k-mooneyi expoerted it to a file when i wanted it to fix spelling18:18
gibiand does our Agentic overlord LLMs can read an etherpad?18:18
sean-k-mooneybut proably yes18:18
sean-k-mooneyit can read it in that its just a webpage18:19
sean-k-mooneybut i didnt want to givbe it a link18:19
sean-k-mooneybecause i didnt want it to do random things18:19
Ugglagibi, thanks, but regarding the hour 20:15 looks a bit late and not cool. I wish you a good WE and please rest during the WE. Next week will be intense.19:25
gibiUggla: thanks. It is complicated. :D20:43

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!