Monday, 2026-02-16

*** ykarel_ is now known as ykarel04:17
opendevreviewTakashi Kajinami proposed openstack/placement master: Drop unused bandit target  https://review.opendev.org/c/openstack/placement/+/97690608:56
opendevreviewTakashi Kajinami proposed openstack/placement master: Skip functional tests for git config update  https://review.opendev.org/c/openstack/placement/+/97690708:58
opendevreviewLajos Katona proposed openstack/nova master: Use SDK for Neutron extensions  https://review.opendev.org/c/openstack/nova/+/96227010:14
sean-k-mooneygmaan: we are down to the last 8 patches for the openapi serise and im +2 on all of them. 6 of those are the servers api which is pretty heavy lifting but what are the chances we can get those merged today or this week?11:46
sean-k-mooneygmaan: it woudl be really nice to check that off the list and complete that work11:47
opendevreviewStephen Finucane proposed openstack/nova master: db: Move regex helpers to utils  https://review.opendev.org/c/openstack/nova/+/95874512:13
opendevreviewStephen Finucane proposed openstack/nova master: tests: Clean up flavors tests  https://review.opendev.org/c/openstack/nova/+/95874612:13
opendevreviewStephen Finucane proposed openstack/nova master: api: Simplify API version check for flavor description  https://review.opendev.org/c/openstack/nova/+/95874712:13
opendevreviewStephen Finucane proposed openstack/nova master: api: Add ability to filter flavors by name  https://review.opendev.org/c/openstack/nova/+/95874812:13
opendevreviewStephen Finucane proposed openstack/nova master: api: Remove dead fields from flavors response  https://review.opendev.org/c/openstack/nova/+/95874912:13
opendevreviewStephen Finucane proposed openstack/nova master: api: Restrict additional query string arguments  https://review.opendev.org/c/openstack/nova/+/95875012:13
opendevreviewStephen Finucane proposed openstack/nova master: tests: Fix typo  https://review.opendev.org/c/openstack/nova/+/95885512:13
opendevreviewStephen Finucane proposed openstack/nova master: api: Add runtime check for query additionalProperties  https://review.opendev.org/c/openstack/nova/+/95885612:13
opendevreviewStephen Finucane proposed openstack/nova master: tests: Add missing test coverage  https://review.opendev.org/c/openstack/nova/+/95885712:13
opendevreviewStephen Finucane proposed openstack/nova master: WIP: tests: Replace manual mocking with fixture  https://review.opendev.org/c/openstack/nova/+/97216212:13
opendevreviewStephen Finucane proposed openstack/nova master: api: Restrict additional query string arguments  https://review.opendev.org/c/openstack/nova/+/95875012:23
opendevreviewStephen Finucane proposed openstack/nova master: tests: Fix typo  https://review.opendev.org/c/openstack/nova/+/95885512:23
opendevreviewStephen Finucane proposed openstack/nova master: api: Add runtime check for query additionalProperties  https://review.opendev.org/c/openstack/nova/+/95885612:23
opendevreviewStephen Finucane proposed openstack/nova master: tests: Add missing test coverage  https://review.opendev.org/c/openstack/nova/+/95885712:23
opendevreviewStephen Finucane proposed openstack/nova master: WIP: tests: Replace manual mocking with fixture  https://review.opendev.org/c/openstack/nova/+/97216212:23
stephenfinUggla: jfyi the server share command support for OSC still hasn't merged https://review.opendev.org/c/openstack/python-openstackclient/+/88154012:29
opendevreviewPierre-Samuel Le Stang proposed openstack/nova master: Reset file position on Glance upload retry  https://review.opendev.org/c/openstack/nova/+/97676613:44
dansmith /query fungi 14:48
opendevreviewTakashi Kajinami proposed openstack/placement master: Drop sphinxcontrib-actdiag  https://review.opendev.org/c/openstack/placement/+/97698214:59
opendevreviewTakashi Kajinami proposed openstack/placement master: Fix sphinxcontrib-actdiag/seqdiag  https://review.opendev.org/c/openstack/placement/+/97698215:03
tkajinam^^^ this is yet another pkg_resources problem...15:07
sean-k-mooneytkajinam: wont that break our curren tuseage of those or have we already removed all usage of them?15:15
opendevreviewTakashi Kajinami proposed openstack/placement master: Remove sphinxcontrib-actdiag/seqdiag  https://review.opendev.org/c/openstack/placement/+/97698215:15
tkajinamsean-k-mooney, quick grep shows no actual usage of these15:15
sean-k-mooneyi am not sure if we have sequence diagrams in placment docs15:15
sean-k-mooneyack i guess the docs job will fial if they are used15:16
tkajinamyeah15:16
sean-k-mooneyill check back once the jobs report but removing the unused deps makes sense even if pkg_resocuse was not a thing given there maintance status15:17
r-taketnUggla: gibi: bauzas: Hello, and sorry for the repeated request. Could you please take another look at the SEV refactoring patches(https://review.opendev.org/q/topic:%22bp/generalize-sev-code%22)? 15:20
UgglaHello r-taketn  sure15:21
r-taketnI have updated the code and comments based on your recent feedback. 15:21
r-taketnUggla: Thank you very much.15:21
*** haleyb is now known as haleyb|out15:26
UgglaNova upstream meeting in ~30mn15:28
opendevreviewElod Illes proposed openstack/nova stable/2025.1: [CI][stable-only] Fix docs job and set skip level grenade non-voting  https://review.opendev.org/c/openstack/nova/+/97699115:30
dansmithbauzas: around?15:55
bauzasyes, but working on my new laptop15:55
dansmithI've pinged you a couple times on this which is holding up the TPM set.... if you're not reviewing that for a reason, please let us know https://review.opendev.org/c/openstack/nova/+/95262815:55
dansmithit's basically only used by the host mode patch, which is +2+W15:56
bauzasdansmith: sorry, missed your pings, I'll work hard on reviewing the vtpm series indeed15:56
dansmiththere's a lot queued to merge, waiting on that trivial object patch is all15:57
bauzas(on friday, was on PTO)15:57
Uggla#startmeeting nova16:03
opendevmeetMeeting started Mon Feb 16 16:03:09 2026 UTC and is due to finish in 60 minutes.  The chair is Uggla. Information about MeetBot at http://wiki.debian.org/MeetBot.16:03
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:03
opendevmeetThe meeting name has been set to 'nova'16:03
UgglaHello everyone16:03
lajoskatona1o/16:03
bauzaso/16:03
tkajinamo/16:03
sean-k-mooneyo/16:04
elodilleso/16:04
fwiesel\016:04
dansmitho/16:04
UgglaLet's go16:06
Uggla#topic Bugs (stuck/critical) 16:06
Uggla#info No Critical bug16:06
Uggla#topic Gate status16:06
Uggla#link https://bugs.launchpad.net/nova/+bugs?field.tag=gate-failure Nova gate bugs 16:06
Uggla#link https://etherpad.opendev.org/p/nova-ci-failures-minimal16:07
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 status16:07
tkajinamI've not filed a bug for this, but doc job for placement repo is broken due to pkg_resources removal. https://review.opendev.org/c/openstack/placement/+/976982 should fix it.16:07
Uggla#info Please look at the gate failures and file a bug report with the gate-failure tag.16:07
Uggla         #info Please try to provide a meaningful comment when you recheck16:07
UgglaLast week we identified several gates problems around devstack.16:08
Ugglatkajinam would you like to add more on this ?16:08
sean-k-mooneyit passes ci and the deps seam to be unused16:09
tkajinamUggla, I think devstack job was unblocked by disabling horizon, and xstatic package fix is still on-going16:09
tkajinam(I hope that I understand "this" correctly16:09
tkajinamsean-k-mooney, yup :-D16:09
sean-k-mooneyi just +2'd it16:10
sean-k-mooneywe proably shoudl backport that as well16:10
sean-k-mooneyi assume the same jobs are broken on the stabel branches16:10
sean-k-mooneyit would be nice to have a bug for that but not strictly requried in this case16:10
tkajinamah, yes. we likely need that backported to stable branches16:11
elodillesthe other option is to pin setuptools==81.0.0 on stable gates.16:11
elodillesfor docs job16:11
sean-k-mooneyya given this is just removing unsued requirement i feel like the backport is better16:11
sean-k-mooneybut either is fine16:12
tkajinamthat should be also ok, though I tend to try backporing it instead of adding hack for the unused requirements16:12
* tkajinam is repeating mostly the same thing :-P16:12
Ugglagood thanks for the updates16:12
Ugglamoving on to next topic unless you want to sahre something else ?16:13
Uggla#topic Release Planning 16:13
Uggla#link https://releases.openstack.org/gazpacho/schedule.html16:14
Uggla#info Nova deadlines are set in the above schedule16:14
Uggla#info Feature freeze is next week.16:14
Uggla#info PTG etherpad for 2026.1 is available: https://etherpad.opendev.org/p/nova-2026.1-ptg16:14
Uggla#topic Review priorities 16:14
Uggla#link https://etherpad.opendev.org/p/nova-2026.1-status16:15
UgglaThis week Thursay (20260219) is spec review day.16:15
Uggla#info I have updated Launchpad and the above doc. Please ping me if you spot something missing.16:15
tkajinamI know that people are busy but I'd appreciate any additional review for firmware selection bp. https://review.opendev.org/q/topic:%22bp/libvirt-firmware-auto-selection%2216:15
tkajinam(will put links to the etherpad later16:16
tkajinamalso https://review.opendev.org/c/openstack/nova/+/909635 may be nice to have, because it's related to mem encryption refactoring.16:16
Ugglatkajinam will review it again, but imho it was almost good.16:17
* tkajinam is too greedy :-P16:17
tkajinamUggla, thanks !16:17
UgglaUnless if I miss something, both confidential computing features are ok.16:18
tkajinamyup :-D16:18
Ugglabtw thank tkajinam and r-taketn for these features.16:20
Ugglamoving on16:20
Uggla#topic OpenAPI 16:20
Uggla#link: https://review.opendev.org/q/topic:%22openapi%22+(project:openstack/nova+OR+project:openstack/placement)+-status:merged+-status:abandoned16:20
Uggla#info still 11 remaining atm.16:20
UgglaI hope we will manage to close this before FF.16:21
UgglaThat looks in a good shape.16:21
sean-k-mooneyUggla: i think its actlly less then that16:21
sean-k-mooneyi think 7-8 to be compelte and the rest are minior fix ups16:21
sean-k-mooneyso ya im hoping those could land this week16:22
Ugglasean-k-mooney 👍16:22
Ugglawould be great. So we can close this long topic.16:22
Uggla#topic Stable Branches16:23
* Uggla giving the mic to elodilles16:23
elodillesthx16:23
elodilles#info stable/2025.2 gate is working again (thanks to grenade job fixes)16:23
elodilles#info stable/2025.1 gate is still broken (docs job, skip-level grenade): workaround: https://review.opendev.org/c/openstack/nova/+/97699116:23
elodilles#info stable/2024.2 gate is broken (py312 based jobs). fix: https://review.opendev.org/c/openstack/nova/+/97637316:23
elodillesi know it's hard to compete with feature patches these days, but i'd appreciate some stable core reviews on these two patches ^^^ o:)16:24
elodilles#info stable branch status / gate failures tracking etherpad: https://etherpad.opendev.org/p/nova-stable-branch-ci16:24
elodillesand that's all from me about stable branches16:24
* elodilles is passing back the mic16:25
Ugglathx elodilles16:25
UgglaI'll try to have a look at those patches, despite this week is shorter for me.16:26
Uggla#topic vmwareapi 3rd-party CI efforts Highlights16:26
* Uggla passing the mic to fwiesel16:26
fwieselHi, no updates from my side here.16:26
Ugglaok fwiesel no pb.16:26
UgglaSkipping Gibi's topic because he is on PTO.16:27
Uggla#topic Nova using openstack sdk for neutron16:27
* Uggla passing the mic to lajoskatona116:27
lajoskatona1Hi, not much progress the last week16:28
lajoskatona1I updated the 1st 3 commit, fixed few nits, that's it for this topic from my side16:28
Ugglalajoskatona1 no pb. Thanks for the update16:29
Uggla#topic Open discussion16:29
UgglaWe have a topic: (fwiesel) Should rebuild & resize reschedule like instance spawn & Live-migration on failure? (i.e. https://blueprints.launchpad.net/nova/+spec/flexible-evacuate-scheduler)16:30
Ugglafwiesel the floor is yours16:30
fwieselSorry, that has clarified itself. I misunderstood the code.16:30
fwieselInstead:16:30
fwieselImageMetaProps and different hypervisors: hw_disk_bus.16:30
fwieselSo, on the image, we have some properties, like the hw_disk_bus. Which makes sense, because in particular windows, one has to install certain drivers that they work (i.e. virtio).16:31
sean-k-mooneyso resize does reschudle rebuild is not a move operation so it cant evacuate is a move op built on rebuild and it always must select a diffent node16:32
sean-k-mooneyas far as im aware evacutate does not require a host today16:33
sean-k-mooneyso the start of that bluepritn confuses me16:33
fwieselsean-k-mooney: Sorry, I had different topic. I think, the blueprint is simply severely outdated.16:33
sean-k-mooneyoh ok16:34
sean-k-mooneysorry proceed with your topic16:35
fwieselSo, the issue is: We have that parameter, that depends on the hypervisor (and the drivers installed in the image), but I can only set it to one value.16:35
fwieselThe only option I see is to upload the same image to glance to give it different metadata.16:35
fwieselBut usually an OS can have various drivers installed. 16:36
fwieselSo, if I as a user of openstack have a windows with vmware paravirtual drivers and virtio drivers installed, and I could leave it to the driver to chose the best option.16:37
sean-k-mooneyhw_disk_bus is specified on the image not just because of drivers but also becasue there are diffent performance charcteriestincs and fucntionaltiy depending on the value16:38
sean-k-mooneyfwiesel: by default the hypervior dirver is allowed to choose the best value if you dont define the metadta16:38
fwieselBut the best defines on knowledge I need from the image.16:39
sean-k-mooneyin the case of the libvirt driver there is even the ableity to consult libosinfo for a recomenmdation although i do not encourage using that 16:39
sean-k-mooneylibosinof does not guarentee so the behvior can change after a package udpate16:40
sean-k-mooneywell what bests depend on the workload adn the image but yes its a trade off16:40
fwieselsean-k-mooney: Would you see a point in moving from a single value for things like hw_disk_bus to a list?16:40
sean-k-mooneyi dotn think we shoudl do that16:41
sean-k-mooneyhow woudl we chosoe the correct value in the dirver or ensure it does nto change when live migrating ectra16:41
fwieselRight now, I cannot create an image that can be booted with paravirtual drivers on vmware as well as on libvirt, simply because the values are conflicting16:42
tkajinamfwiesel, do you mind explaining a bit more details, especially the actual value ofr hw_disk_bus you want to set for libvirt+qemu and vmware ?16:42
sean-k-mooneywe coudl possible add new image proeprties to express a prefernce but the resulting vm defintion idelaly shoudl not change based on teh hypervior it lands on16:42
fwieseltkajinam: Sure, in case of vmware, I would like pvscsi, in libvirt+qemu, I might want virtio-scsi, in case of libvirt+ch I might virtio-block16:43
sean-k-mooneyso that actuly the releated parmater which is the hw_scsi_model16:43
sean-k-mooneyhttps://github.com/openstack/glance/blob/master/etc/metadefs/compute-libvirt-image.json#L7016:44
sean-k-mooneyfwiesel: you want to have 1 iamge that can be booted on both correct16:45
tkajinamI guess so16:45
fwieselsean-k-mooney: Well, even all three16:45
fwieselI mean, assuming libvirt+ch is coming. So, it isn't necessarily a VMware only thing.16:46
fwieselWe also have hw_vif_model with virtio and vmxnet3.16:46
sean-k-mooneythat is hypervior depened if not defiend16:48
sean-k-mooneynormally you shoudl not set hw_vif_model on your images16:48
fwieselsean-k-mooney: But yes, I'd be looking for a way to formulate the metadata in a way to make Nova clear, that actually that windows has support for virtio driers installed, and also the vmware drivers16:48
sean-k-mooneythat not what the existing image properties are for16:49
sean-k-mooneythey are not for expressing supprot16:49
sean-k-mooneythey are for requesting the explcity use of a bus or model16:49
sean-k-mooneysupprot coudl be express today with traits16:49
fwieselOn an image?16:50
sean-k-mooneyin placment we supprot the in operation16:50
sean-k-mooneyand you can have required traits on an image16:50
tkajinambut that does not affect xml, right ?16:50
sean-k-mooneyi am unsuer if we have testse trais:required=in:virtio,vmwarething16:50
sean-k-mooneyright the vm value  woudl be determin by the driver16:51
sean-k-mooneybased on what it landed on16:51
tkajinamI think the core problem is that for windows sometimes virtio (which is the libvirt+qemu's default) is not preferred.16:51
sean-k-mooneywe could add new option that took a hypverior as a prefix i guess16:51
tkajinamand requesting no-default bakes your "request" which looses flexibility16:51
fwieseltkajinam: It cannot be preferred, because it isn't ensured that it is installed.16:52
sean-k-mooneyhw_disk_bus_preference=qemu:virtio:vmware:scsi but im not sure if we should16:52
tkajinamfwiesel, I know it might be a off-topic but I'm just curious but how you handle different tools (qemu-ga vs vmware-tools) in that case ?16:52
tkajinamI mean, I though you may need different agent installed for libvirt+qemu and vmware and for that you eventually need different images, but I could be wrong16:54
fwieseltkajinam: Well, right now we only have the vmware-tools, and prefer not to use any tools in the os, unless forced to.16:54
tkajinamok16:54
sean-k-mooneyi dont know what others think16:55
sean-k-mooneyi can see the pain point16:55
sean-k-mooneybut we have genrealy assume diffent images per hypervior16:55
tkajinamyeah16:55
sean-k-mooneymainly because historicly outside of raw the image format has not been protabl16:56
UgglaI agree.16:56
fwieselWell, that is the fun thing. Qemu-img converts it quite fine on import16:56
sean-k-mooneyit does although converting vmdk image has secirty implciation and we have dicussed if it should do that at least by default in the future16:58
dansmiththis seems like something I care about, but I've been distracted here16:58
sean-k-mooneyyep i was thinkign that ^16:58
UgglaWe are almost at the top of the hour. Can we try to conclude ?16:58
sean-k-mooneyi think we need to dicuss it more maybe on a spec/ptg/mailing list or outside the meeting16:59
tkajinam+116:59
dansmithyeah, image properties with hypervisor prefixes seems uncool16:59
Ugglawe can discuss it next week too.16:59
sean-k-mooneyi.e. to understand fi we really want to open the door to multi value hypverisor denpened image porpeties16:59
dansmithbut definitely good to have a targeted discussion16:59
fwieselAny preferred communication venue?17:00
fwieselmailing list?17:00
fwieselI'll take that then as a yes to mailing list :) 17:02
tkajinamstart with mailing list but I feel like we eventually need it discussed at pgt (to get agreement from enough cores)17:02
tkajinamptg17:02
fwieseltkajinam: Understood.17:02
Ugglaare we good ?17:03
fwieselFrom my side yes. Thanks17:03
Ugglaso time to close, thanks for joining this meeting. Have a nice day/evening and see you next week.17:04
Uggla#endmeeting17:04
opendevmeetMeeting ended Mon Feb 16 17:04:21 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)17:04
opendevmeetMinutes:        https://meetings.opendev.org/meetings/nova/2026/nova.2026-02-16-16.03.html17:04
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/nova/2026/nova.2026-02-16-16.03.txt17:04
opendevmeetLog:            https://meetings.opendev.org/meetings/nova/2026/nova.2026-02-16-16.03.log.html17:04
tkajinamthanks !17:04
fwieselThanks!17:04
elodillesthanks o/17:05
lajoskatona1Bye17:05
tkajinamas a reminder we need https://review.opendev.org/c/openstack/placement/+/976982 to fix doc build in placement.17:05
tkajinamdansmith, do you have a minute to throw this to gate ? ^^^17:06
dansmithtkajinam: yup17:10
tkajinamdansmith, thanks !17:11
opendevreviewMerged openstack/nova master: Add vtpm_secret_(uuid|value) to LibvirtLiveMigrateData  https://review.opendev.org/c/openstack/nova/+/95262818:28
opendevreviewMerged openstack/nova master: TPM: prepare to bump service version for live migration  https://review.opendev.org/c/openstack/nova/+/96205118:30
opendevreviewMerged openstack/nova master: TPM: support live migration of `host` secret security  https://review.opendev.org/c/openstack/nova/+/94148318:30
opendevreviewMerged openstack/placement master: Remove sphinxcontrib-actdiag/seqdiag  https://review.opendev.org/c/openstack/placement/+/97698218:55
opendevreviewGhanshyam Maan proposed openstack/nova master: WIP: Use 2nd RPC server in compute operations  https://review.opendev.org/c/openstack/nova/+/97558820:20
gmaansean-k-mooney: ack, I will try to check this week20:28
opendevreviewGhanshyam Maan proposed openstack/nova master: Add manager graceful shutdown, timeout, and wait  https://review.opendev.org/c/openstack/nova/+/97558621:47
mikaldansmith / sean-k-mooney: I have a crazy plan for making image conversion safer, but it's not ready for public view and would take a while to land all the pieces. I haven't had any luck finding anyone who is happy to talk about the idea before release, so I'll just just one day show up with a thing when it's readyish.21:58
dansmithmikal: I'm happy to talk about it22:01
dansmithcan we save it for ptg?22:02
mikalIf you send me a GitHub username I'll invite you to a private repo. When is the ptg again?22:02
dansmithend of april22:02
mikalYeah, I think I will have shipped a prototype by then. I am aspiring to get something out the door by the end of this month. Assuming y'all can find a time which also works for my timezone I'd be happy enough to chat.22:03
mikalEven if my prototype is pure genius, it's going to take a while to stabilise and for distros to package etc so I think this is more of a long term play than anything.22:05
mikalAnd my prototype might be... Stupid. I haven't had much luck validating the idea outside of my own head.22:05
dansmithwell, we also have some "delete this and start over" feelings about the whole of the imagebackend, so it probably makes sense to discuss both in terms of that22:06
mikalOh god yes. imagebackend has needed a rewrite for many many years. So confusing. That would be an interesting opportunity. There are some interesting things which could be done with distributed caching as well, but it's all too hard with the current implementation.22:07
*** erlon6 is now known as erlon23:10

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