Monday, 2026-03-09

opendevreviewDoug Goldstein proposed openstack/nova master: feat: add get_vnic_type() to ComputeDriver and set for Ironic  https://review.opendev.org/c/openstack/nova/+/97935403:22
opendevreviewTaketani Ryo proposed openstack/nova master: mem-enc: adjust requirement checks for mem_encryption guests  https://review.opendev.org/c/openstack/nova/+/96797407:55
opendevreviewTaketani Ryo proposed openstack/nova master: mem-enc: introduce a check between mem_encryption and locked_memory  https://review.opendev.org/c/openstack/nova/+/97130007:55
masahitohello nova team. Please review the lock API 500 error bug fix if you have time. https://review.opendev.org/c/openstack/nova/+/94622308:12
nicolairuckelAre there any updates on https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/2TF4CXQUCP4Y24PAKZQGM2D5BONL5Z34/08:37
nicolairuckelI wasn't able to find any bug report on that on launchpad yet.08:44
nicolairuckelThere's a good chance I'm looking in the wrong place though.08:44
r-taketngibi: bauzas: Could you please take another look at sev refcator series(https://review.opendev.org/q/topic:%22bp/generalize-sev-code%22)? I’ve updated the code and comments based on the feedbacks from gibi last week.11:02
elodillesUggla: hi o/ is the cycle highlights patch ready for merge? or do you want to update it before the merge?11:17
opendevreviewAnton Iacobaeus proposed openstack/nova-specs master: Intel TDX confidential VM support for libvirt driver  https://review.opendev.org/c/openstack/nova-specs/+/97960812:36
opendevreviewAnton Iacobaeus proposed openstack/nova-specs master: Intel TDX support in libvirt driver  https://review.opendev.org/c/openstack/nova-specs/+/97960812:49
opendevreviewElod Illes proposed openstack/placement stable/2024.2: [CI][stable-only] Workaround for missing pkg_resources  https://review.opendev.org/c/openstack/placement/+/97964213:16
opendevreviewElod Illes proposed openstack/placement stable/2025.2: DNM: CI health test  https://review.opendev.org/c/openstack/placement/+/97964413:21
opendevreviewElod Illes proposed openstack/osc-placement stable/2025.2: DNM: CI health test  https://review.opendev.org/c/openstack/osc-placement/+/97964613:24
cardoeDo I need a spec for https://review.opendev.org/c/openstack/nova/+/979354 to be able to be considered?13:27
dansmithcardoe: as it affects all the drivers and sort of alters the compute manager behavior, it might be a good idea.. like to explain if/how another driver might ever use that and what the other options are13:32
dansmithlike, AFAIK, other drivers support _multiple_ VNIC types so a driver method that returns only one might be kinda fundamentally not useful for others (IDK)13:32
dansmithalso, this seems like less of a "feat"ure and more like a bug fix13:33
dansmithso you could try arguing it with a bug, although I'll have the same questions13:33
sean-k-mooneycardoe: one of my consuer is nova and ironci should never modify the vnic type13:38
sean-k-mooneythe existing bevhiaor in ironic is violating that13:39
sean-k-mooneycardoe: this touches on the fact taht ironic is techinly not ment to be modifying the port bindig for nova ironic isntances but is for legacy reaons13:39
sean-k-mooneywe have again for legacy reason restited proviing a way for nova ot create port with a vnic type other then normal13:40
sean-k-mooneyforcing the pre  creation of port for dpdk, sriov or ironic networing13:41
sean-k-mooneyi had suggested allowing the vnic type to be specified when you provide a network in the past and that was rejected or haveing nova determin it so i think this really woudl need a spec to get agreement13:42
dansmith^13:42
sean-k-mooneycardoe: we never blocked --nework form being used with irionic flavors partly due to the upgrade impact but its also techinally incorrect to use that today13:43
sean-k-mooneyit only works becuase ironic "fixes" it eventhough it not really ment to modify the neutron ports13:43
sean-k-mooneyalso to be clear im not really agaisnt the idea of having a hyperviosr specific default vnic type. just pointing out the history of why its not done that way today13:44
cardoedansmith: yeah so this function is more like "default VNIC type"13:45
dansmithI assume that there's also value in having a driver reject a type it doesn't support, or expose the list of supported  ones?13:46
cardoedansmith: that would make sense for me.13:46
dansmithlike libvirt should refuse type "baremetal" as much as ironic should allow it13:46
cardoeessentially Ironic will never work with VNIC_NORMAL13:46
cardoeexactly.13:46
sean-k-mooneythere would be value in reporting this to placement as taits too. i think neutorn may actuly do that in some cases13:46
dansmithcardoe: this is all stuff that the spec could/would shake out, so I think you're making that case for yourself13:47
cardoeyep.13:47
sean-k-mooneyi.e. so you can shudle dpdk vms (with vnic-type=vhsot-user) to dpdk host based on the trait13:47
sean-k-mooneyalthough i woudl expect neutron to provide those required traits not nova13:47
cardoeI would ask for your guys help if that's okay. My focus area is gonna be the Ironic side. But I can see value for libvirt and others advertising what they support.13:47
sean-k-mooneywell maybe not we cant get the info form neuton until we bind the port so nova would have to do that in a prefilter13:48
sean-k-mooneycardoe: the selectionof host based on vnic type could be declared out of scope as a followup but shoudl be noted in the spec13:49
cardoeokay.13:49
sean-k-mooneyit would be nice to exclude hosts we know cant work however13:49
cardoespeaking of specs... https://review.opendev.org/c/openstack/nova-specs/+/471815 :-D13:49
sean-k-mooneybut nice to have rahter then mvp13:49
sean-k-mooneycardoe: ill see if i can get to that again this week. i dont think it has chagne much since i was orginally +213:50
cardoeliterally not a character has changed since it was +2. The only thing that changed is the version its targeting.13:52
cardoeI agree there's ambiguity but the ambiguity that folks have an issue with is fields that currently exist in the metadata and the current spec is ambiguous.13:53
sean-k-mooneywell litrally that not quite ture https://review.opendev.org/c/openstack/nova-specs/+/471815/10..12/specs/2026.2/approved/expose-vlan-trunking.rst but they were just typo/spelling/grammer fixes like adding puncution13:53
sean-k-mooneycardoe: the reason im not restoreing my +2 is i have not re read it to see if any of the changes we have made in the last 2 cycles break any of the assumations or behvioir13:54
cardoeThat's fine.13:54
sean-k-mooneycardoe: im not aware of any of the top of my head13:54
cardoeI just don't want to be on the hook to fix ambiguity in the current metadata format in this spec.13:55
cardoeI'm all for fixing the ambiguity but that's something different because that's touching a different part of the metadata.13:55
sean-k-mooneyright i dont think you are13:55
cardoeSo I would do that in a follow on.13:55
cardoeCause that would involve changing the "eth" type (I might be wrong on the name but whatever the regular network interface name is).13:56
sean-k-mooneyyour fering to modifying the value of link to somethign that make sense to the guest?13:57
sean-k-mooneyor how we ferencie things in the id fields?13:57
sean-k-mooneycardoe: i will not that ironic decieed to implemnt this on there side last cycle or the one before and they now rewite this contnet and use the uuids i belive13:58
cardoeWell the feedback I got was that "id" exposed hypervisor specific info to the guest since it did {"id": "tap143d8632"} but that's how the libvirt driver works today. The only part of the spec that makes mention of that is that we use the id to say "this VLAN is on that interface"14:01
cardoeFolks didn't like that the spec had {"vlan_link": "tap143d8632"} but that's literally just the reference back.14:01
sean-k-mooneyya so that specifc calue is coming form neutron 14:02
cardoeI'm all for changing the value "tap143d8632" but that's a follow on.14:02
cardoeMake sense?14:02
sean-k-mooneyand its the name of the device on the host which wont be visable on the guest14:02
sean-k-mooneyyep14:02
sean-k-mooneythat is fine what im tellign you is that if your primay use case is for ironic they now add some of this info to the metadata and mofiy what nova returns14:03
sean-k-mooneyand i belvie htey brok compatiyblity when they did that and change the value of it14:03
cardoeMy primary use case is Ironic but it's still a valid libvirt use case (honestly when I worked on Xen this would have been a great use case there as well)14:03
sean-k-mooneythat does not mean we shoudl not do this feature without changing that initally for neutorn in general14:03
sean-k-mooneyand then we can decied on normalisting the value seperately14:03
cardoeYep. I'll work with ya on a follow up.14:04
sean-k-mooneyhttps://review.opendev.org/c/openstack/ironic/+/946677 that provides part of the info right, the mtu specificly bu tyour going beyond that and providign the vlan subport configution based on neutron trunks14:06
Ugglaelodilles sorry I missed your earlier message. Can we hold the patch for now. There is something to discuss in today's meeting.14:11
elodillesUggla: sure, no problem, i was just asking to not forget about it o:)14:11
elodillesUggla: also note, that RC1 patches are proposed for nova and placement on openstack/releases repository o:)14:26
Ugglaelodilles 👍14:27
sean-k-mooneycardoe: +1 on https://review.opendev.org/c/openstack/nova-specs/+/471815 can you confirm or document what type:ovs means in the metadata that really my main open quetion14:32
cardoeI can remove that. I believe the author intended for that to be clear that was on the hypervisor side it was being trunked.14:44
sean-k-mooneywell i think you need to represent the guest interface that the vlans are commign form i just dont know what the type shoudl be15:02
sean-k-mooneywe just need ot docuemtn what will be generated and what the expected behvior is 15:04
sean-k-mooneycardoe: ok i have foudn the anser sort of in the schema15:13
UgglaReminder: Nova meeting in ~40mn15:21
ralonsohsean-k-mooney, ping me tomorrow about the vif type virtual thing15:26
ralonsohin a nutshell, we use it for the OVN virtual ports15:26
sean-k-mooneyralonsoh: i think it may be buggy15:27
ralonsohwhy?15:27
sean-k-mooneyor logiclly incorrect15:27
sean-k-mooneywell how do you handel the same ip beigin allowed to fail over between 2 vms15:27
sean-k-mooneywhere im using vrrp or somiler in the guest15:27
ralonsohsean-k-mooney, I don´t get your point15:28
sean-k-mooneywe cant assuem that allowed_address_pair values are uniquce on a network or subnet15:28
ralonsohyes15:28
sean-k-mooneyso we can assuem the port is bound to only one host15:28
sean-k-mooneybut its fien we can chat tomorrow15:28
ralonsohyeah, I need to go to the doctor in some minutes, sorry15:28
sean-k-mooneyill read the bug report15:28
ralonsohadd any comment there, please15:29
ralonsohI'll check it tomorrow morning15:29
opendevreviewElod Illes proposed openstack/placement stable/2025.2: [CI][stable-only] Fix docs job  https://review.opendev.org/c/openstack/placement/+/97968515:35
opendevreviewElod Illes proposed openstack/placement stable/2025.1: DNM: CI health test  https://review.opendev.org/c/openstack/placement/+/97968615:36
opendevreviewElod Illes proposed openstack/osc-placement stable/2025.1: DNM: CI health test  https://review.opendev.org/c/openstack/osc-placement/+/97968715:38
opendevreviewElod Illes proposed openstack/osc-placement stable/2024.2: DNM: CI health test  https://review.opendev.org/c/openstack/osc-placement/+/97968815:38
opendevreviewElod Illes proposed openstack/osc-placement stable/2024.2: [CI][stable-only] Workaround for missing pkg_resources  https://review.opendev.org/c/openstack/osc-placement/+/97969115:52
UgglaHi elodilles15:56
UgglaDo you have a specific deadline for RC1 ? Is Thursday ok ?15:57
opendevreviewElod Illes proposed openstack/osc-placement stable/2025.1: [CI][stable-only] Fix docs job  https://review.opendev.org/c/openstack/osc-placement/+/97969215:57
elodillesUggla: thursday is OK15:58
elodillesthat is the official deadline15:58
elodillesif you want to include something that is not yet merged, then add a -1 to the patch and we can wait a bit15:58
Ugglaelodilles ok it is what I thought, but I'd like to crosscheck.15:58
elodilles+115:58
UgglaWe will discuss that in the meeting.15:59
elodillesACK :)15:59
tkajinamI always feel I get extra bonus when people are talking about "deadline"16:00
Uggla#startmeeting nova16:00
opendevmeetMeeting started Mon Mar  9 16:00:28 2026 UTC and is due to finish in 60 minutes.  The chair is Uggla. Information about MeetBot at http://wiki.debian.org/MeetBot.16:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
opendevmeetThe meeting name has been set to 'nova'16:00
tkajinamthanks to time diff16:00
tkajinamo/16:00
UgglaHello everyone16:00
melwitto/16:00
elodilleso/16:00
dansmitho/16:00
kaiserso/16:01
bauzaso/16:01
UgglaLet's start16:02
Uggla#topic Bugs (stuck/critical)16:02
Uggla#info No Critical bug16:02
lajoskatonao/16:02
Ugglahttps://bugs.launchpad.net/nova/+bug/2143057 is now closed16:03
UgglaThanks tkajinam16:03
Uggla#topic Gate status16:04
Uggla#link https://bugs.launchpad.net/nova/+bugs?field.tag=gate-failure Nova gate bugs 16:04
tkajinamoh the fix is still blocked by CI16:04
Uggla#link https://etherpad.opendev.org/p/nova-ci-failures-minimal16:04
opendevreviewThibaut Démaret proposed openstack/nova master: Add default SSD rotation rate for instance disks  https://review.opendev.org/c/openstack/nova/+/97969316:04
tkajinamlet me rerun CI16:04
Uggla#link https://zuul.openstack.org/builds?project=openstack%2Fnova&project=openstack%2Fplacement&branch=stable%2F*&branch=master&pipeline=periodic-weekly&skip=0 Nova&Placement periodic jobs status16:04
fwieselo/16:04
Uggla#info Please look at the gate failures and file a bug report with the gate-failure tag.16:04
Uggla         #info Please try to provide a meaningful comment when you recheck16:04
UgglaLook quickly at the reports, I have not seen something wrong. Please tell me if I'm wrong?16:05
tkajinamI'm not aware of any problem so far16:06
Uggla#topic Release Planning 16:07
Uggla#link https://releases.openstack.org/gazpacho/schedule.html16:07
Uggla#info Nova deadlines are set in the above schedule16:07
Uggla#info PTG etherpad for 2026.1 is available: https://etherpad.opendev.org/p/nova-2026.1-ptg16:07
UgglaI will probably open 2026.2 this week16:07
Uggla#info This week is RC1 target week.16:08
UgglaWe will probably hold the RC1 until Thursday.16:08
dansmith"hold" it?16:09
dansmithlike delay it?16:09
Ugglayep delay. :)16:09
Ugglareasons in the next points.16:10
dansmithokay16:10
Uggla#topic Review priorities16:10
Uggla#link https://etherpad.opendev.org/p/nova-2026.1-status16:10
UgglaStarting here: https://etherpad.opendev.org/p/nova-2026.1-status#L70 there are interesting bugs to improve live migration. So I would like cores to review them. 16:10
UgglaI know some of them have already been merged.16:10
UgglaAnd others seem in a good shape.16:11
dansmiththe first I clicked on has a -1 from you still16:11
dansmithbut yeah will take a look16:11
opendevreviewThibaut Démaret proposed openstack/nova-specs master: Spec for default SSD rotation rate on instance disks  https://review.opendev.org/c/openstack/nova-specs/+/97969416:11
Uggladansmith the are others with +2 from sean-k-mooney16:12
dansmithack16:12
dansmiththey seem to conflict with each other and aren't stacked though, which is unfortunate16:12
Ugglaso first reason of delaying is to try to introduce these patches.16:12
dansmithor at least, two do16:12
UgglaSecond reason, if about the confidential computing refactoring. gibi is following it.16:13
* sean-k-mooney looks at meeting16:14
UgglaLast reason is what we will discuss later in open topics.16:15
sean-k-mooneyUggla: not sure which patch our refeing to16:15
sean-k-mooneybut i can catch up later16:15
bauzasI can try to look at some16:16
dansmithUggla: what about the refactor? I don't see that on the status page16:16
bauzasthe SEV refactor series is a blueprint so it can't be merged now16:16
dansmiththe status page has quite a list to cram in before thursday.. and I guess I'm wary of the neutron SDK stuff16:16
dansmiththose sorts of things tend to have corner cases and since they're not fixing anything or anything that should be user-visible they seem high risk, low reward to me16:17
Ugglasean-k-mooney I was refering to 965927: Add regression test for bug #2088066 | https://review.opendev.org/c/openstack/nova/+/96592716:17
dansmithbauzas: yeah, anything refactor seems like undue risk given where we are16:17
opendevreviewThibaut Démaret proposed openstack/nova master: Add default SSD rotation rate for instance disks  https://review.opendev.org/c/openstack/nova/+/97969316:17
sean-k-mooneyUggla:ack thanks16:17
Uggladansmith, I don't want to force anything. Just trying to land what is possible.16:18
UgglaSo if you think it is not ready, or risky, it will be for next cycle.16:19
opendevreviewThibaut Démaret proposed openstack/nova-specs master: Spec for default SSD rotation rate on instance disks  https://review.opendev.org/c/openstack/nova-specs/+/97969416:19
sean-k-mooneyat this point if its not a regressioin intoduced this cycle i would wait until we have cut RC1 16:19
dansmithsean-k-mooney: ++16:19
sean-k-mooneythat happeinginn on thurday if i recall correctly16:19
Ugglasean-k-mooney yes.16:20
dansmithI'll have to look at the other fixes here like the volume stuff but I suspect they're all fairly latent state tracking things that aren't hugely recent/relevant16:20
Uggladansmith ok16:21
UgglaI have nothing else. So we can move on unless someone wants to add something else.16:22
opendevreviewThibaut Démaret proposed openstack/nova-specs master: Spec for default SSD rotation rate on instance disks  https://review.opendev.org/c/openstack/nova-specs/+/97969616:22
Uggla#topic Stable Branches 16:23
* Uggla giving the mic to elodilles16:23
elodillesthanks o/16:23
elodilles#info stable/2026.1 branch were cut for libraries16:23
elodilles#info stable/2026.1 branch will be cut for nova and placement this week, along with RC1 releases16:24
elodilles#info nova stable gates should be OK, but placement and osc-placement have blocked stable gates (setuptools/pkg_resources issue), see details in tracking pad16:24
elodillesi forgot about these in the past weeks ^^^^ :S16:24
elodilles#info stable branch status / gate failures tracking etherpad: https://etherpad.opendev.org/p/nova-stable-branch-ci16:24
elodillesUggla: that's all from me about stable things16:24
Ugglathx elodilles.16:25
opendevreviewThibaut Démaret proposed openstack/nova-specs master: Spec for default SSD rotation rate on instance disks  https://review.opendev.org/c/openstack/nova-specs/+/97969616:25
Uggla#topic vmwareapi 3rd-party CI efforts Highlights16:26
* Uggla giving the mic to fwiesel16:26
fwieselHi, no updates from my side.16:26
Ugglaok thx fwiesel16:26
Ugglaskipping next point as gibi is not available. And eventlet was merge already.16:27
Uggla#topic Nova using openstack sdk for neutron16:27
lajoskatonao/16:28
* Uggla giving the mic to lajoskatona16:28
lajoskatonaNo progress since last week, as we agreed to push these patches early 26.216:28
dansmith++ early in a cycle is the time for stuff like this :)16:28
lajoskatona+116:29
Ugglayep it is what we mentionned last week. We will try to merge lajoskatona sdk client stuff earmy next cycle.16:29
dansmithcool16:30
Uggla#topic Open discussion16:30
UgglaWe have a topic from melwitt16:31
* Uggla giving the mic to melwitt16:31
melwitthi, yes, wanted to discuss about the possibility of an exception for vtpm live migration remaining patches16:32
melwittthe 'host' security mode for live migration has merged but the 'deployment' mode is still outstanding16:32
melwittthe 'deployment' mode is the one where the nova service user owns the barbican secrets. 16:33
dansmithI haven't reviewed above what we've merged... well, except for the fixup patch16:33
bauzasit will be difficult to review it for 3 days left given the RC1 deadline16:33
dansmithlooks like the last review from anyone on that patch was almost two years ago (Aug 2024)16:34
dansmithyeah, I'm not sure how we could even consider this an FFE candidate given it's not like "reviewed but just didn't make the deadline"...16:35
dansmithand as sean-k-mooney just said above, candidates for things right now should really be "regression introduced this cycle"16:35
sean-k-mooneyim a littel sad the deployment mode was not done first16:35
sean-k-mooneybut ya i think it woudl be better to finish it early next cycle16:36
dansmithreally? i think host is the most useful mode16:36
sean-k-mooneydeployment requried the least code chage16:36
dansmithI guess we don't know until ops start deciding but I guess I think host is what most of them will want16:37
sean-k-mooneythat why i asked for it to be first since it did nto need to object change16:37
sean-k-mooneybut i didnt object to the current order16:37
sean-k-mooneyjust suggested the other one was less invaisve16:37
dansmith*shrug* I was more focused on getting host in because I _suspect_ it's the thing people will want16:37
dansmithbut yeah I get your point16:37
melwittyeah, the less change makes sense as a subset of that is what would be needed for 'user' mode live migration as well16:38
sean-k-mooneyanyway i woudl not be against re reviewing the spec and appoving it and contining to review/merge the feature16:38
sean-k-mooneyeven before we offically release 2026.116:38
sean-k-mooneyi just think we should take the checkpoint with rc1 and create the sable branch first16:39
dansmithsean-k-mooney: you mean once rc1 is cut and master becomes 2026.2 right?16:39
sean-k-mooneyyep16:39
dansmithyeah, agree, I don't see how it could be any other way16:39
sean-k-mooneyim not commeting on the readyness of the remaining patches because the last time i looked in detail was december16:39
bauzasat least operators can support vtpm live migration by G, which is cool16:40
dansmithlast actual review on the deployment patch was you, from Aug 2024 :)16:40
dansmithsean-k-mooney: ^16:40
bauzasthey will only set it per host16:40
bauzasfor the deployment support, we can make it a short-term priority for Hibiscus given the existing momentum16:41
sean-k-mooney:) i deployed something in decemer and played with the series a bit before going on PTO but i have entirlly ejected that form my memory16:41
dansmithsean-k-mooney: even though I get your point about minimal code change for deployment, it's still the largest perceptual change of the lot16:41
dansmithsean-k-mooney: it puts the users not in control of their TPM key16:41
sean-k-mooneyyep16:41
dansmith_that_ is why it seemed it should go last to me16:42
sean-k-mooneyyep i get that16:42
opendevreviewThibaut Démaret proposed openstack/nova-specs master: Spec for default SSD rotation rate on instance disks  https://review.opendev.org/c/openstack/nova-specs/+/97969616:42
sean-k-mooneyif we are agreed that this is not in socpe of a FFE are we ok to repoease and review the spec/code once rc1 is cut and master is 2026.216:43
sean-k-mooneyi mean the repopposal can happen now16:43
sean-k-mooneyjust are we ok with that directionally so we can move on to other topic or is there anything else to raise16:44
dansmithyeah, are we going to just kick the whole spec to 2026.2 or trim it down to just the "...add deployment" part?16:44
bauzaswe can trim it if we want16:44
dansmithbut that detail aside, yep, good to move on16:44
dansmithI think it  might be nice to mark the first two modes as completed in this cycle16:44
dansmithand then just deployment for 2026.2... I know that's more work for melwitt tho...16:44
sean-k-mooneytrim is proably better but i dont have a strong prefence either way16:44
melwittI wondered about that too. to maybe trim and then make a new blueprint also or?16:44
sean-k-mooneyyep that proably for the best16:45
melwittyeah I was thinking about that about being able to mark the host mode as completed tihs cycle. ok, I will do this16:46
dansmith++16:46
sean-k-mooneythat wasy we can close the curent oen and leave a not the the deployment part was defereed16:46
melwitt++16:46
melwittok, thanks all for the discussion16:47
Ugglamelwitt just ping me when the trim is ok.16:48
melwittwill do, thanks Uggla 16:48
Ugglaso I could adjust BP and spec move patch.16:48
Ugglaok I guess we are good with this topic.16:50
UgglaDoes someone want to discuss something else.16:50
nicolairuckelIs https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/2TF4CXQUCP4Y24PAKZQGM2D5BONL5Z34/ covered by https://review.opendev.org/c/openstack/placement/+/979685? Unfortunately, that bug blocks my backports of the vTPM and NVRAM patches. :/16:50
tkajinamnicolairuckel, I think that config_format problems is different but it was fixed by https://review.opendev.org/c/zuul/zuul-jobs/+/97913616:51
Ugglamelwitt can you answer ^ ?16:51
tkajinamnicolairuckel, but we still need 979685 to fix doc job in 2025.216:52
elodillesyepp, what tkajinam says. i can second that16:52
nicolairuckelthanks, I missed the first fix16:52
elodillesnicolairuckel: i'll review your backport when i get there o:)16:53
nicolairuckelAfter 979685 I can restart the CI jobs?16:53
nicolairuckelThanks. :D16:53
elodillesno problem16:54
tkajinamor you can rebase your backport on 97968516:54
tkajinamI think even depends-on works16:54
tkajinam(to check whether 979685 fixes the problem blocking your backport16:54
tkajinamwe can discuss it outside of the meeting, I think16:55
nicolairuckeloh, that's a good idea16:55
nicolairuckelyeah16:55
Ugglaok some last minute point to discuss ?16:57
Ugglaseems not.16:58
UgglaI'm skipping bug scrubbing part because we are at the top of the hour.16:58
Ugglaso time to close, thanks for joining this meeting. Have a nice day/evening and see you next week.16:59
Uggla#endmeeting16:59
opendevmeetMeeting ended Mon Mar  9 16:59:53 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:59
opendevmeetMinutes:        https://meetings.opendev.org/meetings/nova/2026/nova.2026-03-09-16.00.html16:59
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/nova/2026/nova.2026-03-09-16.00.txt16:59
opendevmeetLog:            https://meetings.opendev.org/meetings/nova/2026/nova.2026-03-09-16.00.log.html16:59
elodillesthanks Uggla! have a nice evening o/16:59
kaisersAfter meeting follow up request: https://review.opendev.org/c/openstack/nova/+/977300 needs a second +2, it’s a small driver re-support change I brought up here recently already, would be great to have this back in sync with Cinder before the release.17:00
Ugglaelodilles thank you, same for you.17:00
kaisersthnx17:00
kaisers & o/17:00
nicolairuckelthanks, have a nice day17:00
nicolairuckelCan I do the depends-on in Gerrit?17:01
tkajinamthanks17:01
tkajinamnicolairuckel, can you provide the link to your backport ?17:01
sean-k-mooneynicolairuckel: in genral that shoudl work on stable as well yes17:01
nicolairuckelhttps://review.opendev.org/c/openstack/nova/+/978908/717:01
opendevreviewThibaut Démaret proposed openstack/nova-specs master: Spec for default SSD rotation rate on instance disks  https://review.opendev.org/c/openstack/nova-specs/+/97969617:01
tkajinamnicolairuckel, Add  Depends-on: https://review.opendev.org/c/openstack/placement/+/979685  to the commit message. Probably right above Signed-Off-By17:02
tkajinamoh, wait17:03
tkajinamnot that one17:03
sean-k-mooneyhttps://review.opendev.org/c/openstack/placement/+/97968517:03
sean-k-mooneyis fixing the docs job in placment17:03
tkajinamI guess we need the fix for nova17:03
sean-k-mooneyright17:03
tkajinamyeah17:03
elodillesnicolairuckel: i think you don't need to do a depends-on as your patch is in nova repo and the broken job is not placement related 17:03
sean-k-mooneythere actul failre is the tox-cover job17:03
sean-k-mooneyand that was fixed globally right17:03
tkajinamoh yes17:04
tkajinamjust recheck17:04
elodillestkajinam: nicolairuckel: i think we don't need the fix as it was already fixed on Friday17:04
nicolairuckelSo just recheck?17:04
elodillesyepp, what sean-k-mooney 17:04
elodillessays17:04
tkajinamAttributeError: 'Parsed' object has no attribute 'config_format'17:04
tkajinamthat's the root cause and that was fixed by the zuul-job update17:04
tkajinamnicolairuckel, yup17:04
sean-k-mooneyya that was fixed in zuul-jobs 17:04
elodillesnicolairuckel: i'll try to look into your patch tomorrow as it's already end of day to me now o:)17:05
nicolairuckelFor me too17:06
nicolairuckelthanks17:06
sean-k-mooneynicolairuckel: https://review.opendev.org/c/zuul/zuul-jobs/+/979136 for context17:06
sean-k-mooneynicolairuckel: while we often have interestign ci issue near feature freeze this cycle has been extra special due to all the upstream packaging changes17:06
nicolairuckelI'm just always confused when jobs that I can't reproduce locally break.17:09
dansmithUggla: this is not a regression, but probably pretty safe and the case it makes is pretty compelling: https://review.opendev.org/c/openstack/nova/+/95855617:09
dansmithI had not seen this before, unfortunately17:09
Uggladansmith, I said there are interesting patches in the list. :)17:13
UgglaAt least for me it is ok to merge things even if it is safe and useful.17:15
dansmithsean-k-mooney: thoughts? ^17:16
sean-k-mooneyUggla: my question would be what benifit does merging them now vs backporting them in a week by?17:16
sean-k-mooneyif its a bug that we feel comfortable merging 2 day before RC1 while we may not consider it valid for an RC217:17
sean-k-mooneywe likely woudl condier backportign it to sable after the release is done17:17
sean-k-mooneyif its not a bug i.e. a small feature we are well past FF17:17
sean-k-mooneyif there was a small spefic patch you tought was worth merging i am not against lookign at that17:18
sean-k-mooneyjsut the tiem to "merge the thign that are close" in my mind was the week before FF17:18
dansmithsean-k-mooney: linked above.. it's sort of a performance bug, 2s query every give minutes in a periodic that goes to 0s with an index17:19
sean-k-mooneydansmith: on where you askign for https://review.opendev.org/c/openstack/nova/+/958556 specificly17:19
dansmithit's also DB so probably not a good one to backport, but yeah maybe should just wait17:19
sean-k-mooneyok ill take a look at that now17:19
sean-k-mooneysorry i tough you were pointing at Uggla message17:19
dansmithnope, sorry17:19
sean-k-mooneyso a db index is pretty safe to add in genral although it may have a upgrade on upgrade17:20
Ugglasean-k-mooney benefit, assuming that's something not risky. Having it the release is cool. People will have it day 1 and I could eventually announced if in the highlights.17:20
sean-k-mooneyi.e. in terms fo time of the deb has to compute the index17:21
sean-k-mooneydansmith: with alembic we nolonger need ot have the placeholer so it proablme more backporable then it was but let me take a look i guess17:21
dansmithsean-k-mooney: I was also going to ask.. I'm not polished with alembic.. is upgrade the synchronous stage?17:21
dansmithsean-k-mooney: yeah, but still not great to backport unless it fixes a big bug or something17:22
Ugglabut again I'm not pushing/forcing to have them by all mean. Relying on your (core) judgement. 17:22
sean-k-mooneyupgrade in our case i think is part of the expand set17:22
sean-k-mooneybut also not an expret on this i would have ot  look it up or read our docs17:22
dansmithyeah, so I think that's the right phase for it17:22
dansmithupgrade is what you can run before you deploy new code, so that's right17:23
dansmithanyway, maybe the fact that we're even discussing means it should wait, it's just a juicy performance bug with a very simple solution17:24
sean-k-mooneyas a mitigation will alembic complain if you manually added the same index?17:24
dansmithlike if the operator has done this manually already?17:24
sean-k-mooneyim not sure how picky it is to schema deviation 17:25
sean-k-mooneyya17:25
dansmithI also don't know that.. I thought it has to look and see if the uuid of the migration is recorded or not, but I also thought it was supposed to be sort of idempotent, so idk17:25
dansmithI think the fact that we're not positive here means we should wait,17:25
sean-k-mooneybasiclly im askign if the down revisiohn has is based on our model defintion or 2903cd72dc14 by looking at the tables in the db17:25
sean-k-mooneyfine by me i normally ask stephenfin about ^ or melwitt 17:26
sean-k-mooneyanyway without diging into those detail the patch looks vaild to me so ill +1 it but beyond that i would have to read up on the detials17:27
dansmithack Uggla let's punt, but I think this is a good one to track and nag people to look at as soon as H opens17:28
Uggladansmith 👍17:29
stephenfinsean-k-mooney: dansmith: if the question is about backporting migrations, if you're going back further than the release the last migration was added in, then you'll end up creating a branch https://alembic.sqlalchemy.org/en/latest/branches.html#working-with-branches17:32
dansmithI don't want to backport this17:33
sean-k-mooneystephenfin: it more how safe to we feel https://review.opendev.org/c/openstack/nova/+/958556 is and what its upgrade impact17:33
dansmithI think sean-k-mooney's question was what happens if someone  has manually done this or something17:33
sean-k-mooneyalso that17:34
sean-k-mooneythe current patch does not have an upgrade release note and i suspect this will take some time to caluate the inde of there are a large number of rows17:34
dansmithalso true17:34
sean-k-mooneyi belive this will be appled as part of the db sync when we do the expand migritons17:34
dansmithbut if it's in the upgrade phase it should be pre-doable17:35
stephenfinWell they're already out of support, but I believe you'll simply end up with a duplicate index with a different name17:35
stephenfinit would be a 5 minute test if you'd a devstack to hand17:35
sean-k-mooneystephenfin: basicly i was wondering if we fix this in master in 2026.217:35
sean-k-mooneycan operator "fix it themselve" by just creating the index on older rellease without breaking our upgrades17:36
sean-k-mooneyor will almebic be unhapply17:36
sean-k-mooneyif its just a duplcit index and it wont break future migration that proably ok17:36
sean-k-mooneyto be clear 2.264 sec to 0.006 sec for 500k rows is nice but for a perodic that runs every 5 minutes its not unworkable at 2.5seconds17:38
sean-k-mooneydefintly worth fixing on master however17:38
dansmithI thought the claim was that this overwhelms conductors when you have a ton of nodes doing it17:38
dansmith"NOTE: missing index for a query that gets regulary executed by every nova-compute node this can overwhelm the nova-conductor service in large scale deployments"17:39
dansmith(from the etherpad, maybe should be copied into the bug)17:39
sean-k-mooneyoh its every 5 mintes per compute17:39
dansmithyeah17:39
sean-k-mooneyok that scaling diffently then i initally tought17:39
dansmithand probably scales like n^2 almost  because each compute doing it is scanning each other compute record that you ad17:39
dansmith*add17:40
sean-k-mooneyalthough ya it makes sense this is per compute17:40
sean-k-mooneywe already have indexs on source_compute and dest_compute and source_node and dest node17:41
stephenfinsean-k-mooney: I can't say, but we've never written our migrations on the assumption that the operator has manually hacked in their own migrations17:41
sean-k-mooneyhttps://review.opendev.org/c/openstack/nova/+/958556/1/nova/db/main/models.py#89517:41
stephenfinso I'm not sure how this would be any different17:41
stephenfinhowever, if they *had* done that, they can always dummy apply the migration using the stamp command17:41
sean-k-mooneyim actully surpised that a quey on teh subset is not coverd by htat17:42
stephenfinhttps://alembic.sqlalchemy.org/en/latest/api/commands.html#alembic.command.stamp17:42
sean-k-mooneystephenfin: well in the very old days we assumed the dba may create indexes ro set coalation types to optimise for there usecase17:42
sean-k-mooneyim not saying that a good thing just that it was a thing17:43
sean-k-mooneythe ops guided uses to have soem recomentions for adding indexes but that was eons ago17:43
dansmithUggla: the other volume bugs there are good things to fix, but I'm not sure in the best way.. I've left some comments18:25
opendevreviewMerged openstack/os-traits master: Remove MANIFEST.in  https://review.opendev.org/c/openstack/os-traits/+/97599218:42
opendevreviewMerged openstack/osc-placement master: Remove MANIFEST.in  https://review.opendev.org/c/openstack/osc-placement/+/97598418:44

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