Tuesday, 2026-04-21

opendevreviewMasanori Kuroha proposed openstack/nova master: Copy applied provider config  https://review.opendev.org/c/openstack/nova/+/94830401:52
opendevreviewMasanori Kuroha proposed openstack/nova master: Copy applied provider config  https://review.opendev.org/c/openstack/nova/+/94830401:53
opendevreviewAshish Gupta proposed openstack/nova master: Run functional test with threading mode  https://review.opendev.org/c/openstack/nova/+/97985005:52
opendevreviewAshish Gupta proposed openstack/nova master: Disable greenlet-based notifier sender check under native threading temporarily  https://review.opendev.org/c/openstack/nova/+/98562105:52
stephenfingibi: sean-k-mooney: Morning 👋 Before the PTG kicks off, could I get some eyes on https://review.opendev.org/c/openstack/nova/+/985506? It's trivial but needed to unblock an u-c bump for o.vo09:03
gibistephenfin: done09:22
stephenfingibi++ ty 🙏09:23
*** sambork_ is now known as sambork09:30
DominikDanelski[m]I missed signing-up for the virtual PTG for 2026.2. Doest that mean that adding https://review.opendev.org/q/topic:%22bp/numa-topology-with-rps%22 (NUMA Topology with Resource Providers) is dead in the water? Are specs / code changes reviewed, if they haven't been added to the agenda?11:15
sean-k-mooneyDominikDanelski[m]: no, the ptg is for gaining concencous not for approving any specific feaute11:20
sean-k-mooneyso 1 it cna still happen via the normal spec review process and 2 if there is time it can still be included in the ptg adgenda11:20
sean-k-mooneythe main issue i have with the current code is the reqrite of numa_fit_instance_to_host11:22
sean-k-mooneyit not clear to me that you fully undersstood what that fucniton was ment to do in the context of numa in placment11:23
sean-k-mooneyin the numa filtier its role is to  determin if a given allcoation candiate can fullifle the the numa request  so we can filter the canidate in the host state object11:25
sean-k-mooneyin the compute manger/resocue tracker claim its role i sgieve a singel allcoation candate generate the full instnac enuma toplogy and confirm its still valid11:26
sean-k-mooneythat code was intentially written in a functional style so that it was a pure input of the parmater with no sideffect which is why the object oriatneted rewrite makes it harder to revew as it break the mental model we had for that code11:27
sean-k-mooneyeven if the code is effectivly equvlent https://review.opendev.org/c/openstack/nova/+/971176/9/nova/virt/hardware.py is an invaisve rewrite11:27
opendevreviewAnton Iacobaeus proposed openstack/nova-specs master: Intel TDX support in libvirt driver  https://review.opendev.org/c/openstack/nova-specs/+/97960811:35
DominikDanelski[m]sean-k-mooney: Did you see my reply in https://review.opendev.org/c/openstack/nova/+/971176/1/nova/virt/hardware.py#b2509? I think the change to the overall logic is not as significant as it may look from the lines of diff, but I asked about your input there and I'm open to changes if you provide direction for them.11:38
sean-k-mooneyi did but again because it fundemtnally change the organstaion of the code i have not had tiem to fully unlopad all my expecations fo how that code shoudl work adn rereview it based on yoru changes11:47
sean-k-mooneyas i said that refactoring makes it a lot harder to review your change11:47
sean-k-mooneyif you wanted to make that easier to review you would do the refactor in its own patch that does not change any exisitng behivor. although im not entirly conveice that is the correct move. it would make it much clear what is jsut a refactorign vs behvior change11:50
opendevreviewEsra Ozkan proposed openstack/nova master: Check volume status before live migration  https://review.opendev.org/c/openstack/nova/+/97375012:13
opendevreviewElod Illes proposed openstack/nova stable/2025.2: fix: device_by_alias should respect config type  https://review.opendev.org/c/openstack/nova/+/98492012:22
sean-k-mooneyUggla: so its a little disapointing that the lack fo openstack client and sdk? supprot for nova share attatchment was not coverd in teh manilla cross project session12:24
sean-k-mooneyUggla: i dont think we shoudl apcpate any new feature there really until that gap is closed12:24
sean-k-mooneyi dont object to any of the feature but without afctully coing that gap so you dont have to use curl im not sure we should keep exteneding it12:24
sean-k-mooneyoh...12:25
sean-k-mooneysorry i msiste it 12:25
Ugglasean-k-mooney, good news the sdk and client were merged, it was the first topic of yesterday meeting.12:25
sean-k-mooneyya i ddint see that initally12:25
sean-k-mooneyall good then12:25
Ugglasean-k-mooney, gouthamr take over this and manage to cross the finish line.12:26
Ugglasean-k-mooney no worries.12:27
Ugglasean-k-mooney, also carloss presented good progress on the testing side.12:29
sean-k-mooneyshoudl i ask if said testing found bugs or not 12:30
Ugglasean-k-mooney, I would say yes at least one I'm aware of it --> 984802: Set instance to ERROR state on start_instance power_on failure | https://review.opendev.org/c/openstack/nova/+/98480212:33
carlosssean-k-mooney: yep, the change Uggla linked was an issue gouthamr ran into. but so far, in the scenario (end to end) tests I have implemented, no further bugs found12:40
carlossthis is a job run using manila's cephfs native: https://zuul.opendev.org/t/openstack/build/f761d9016979435abe69c849a645ebc212:40
carlossI'm still figuring out why LVM (with NFS) is failing on CI because it's passing in a local env I have - so I think there's some mismatch in the env I own compared to the CI12:41
carlossthe share attachment tests are under the TestShareAttachmentCEPHFS class12:41
sean-k-mooneyUggla: oh the resoume on host boot feature12:44
sean-k-mooneyya im betting we didnt re attach the share or something?12:45
sean-k-mooneythat woudl have been easy to miss12:45
opendevreviewMerged openstack/nova master: objects: Prepare for oslo.versionedobjects 3.10.0  https://review.opendev.org/c/openstack/nova/+/98550612:48
DominikDanelski[m]sean-k-mooney:  OK, so I propose that we split the change in 2: first the behaviour change without the signature change (extending the function's behaviour at the expense of making it long), then the split which would end in the current state. It would be hard to show the refactor without the reason it's being done (extending behaviour).12:54
*** clarkb is now known as Guest768313:48
opendevreviewKarol Klimaszewski proposed openstack/nova-specs master: Spec for SPDK-based nova-provisioned storage backend  https://review.opendev.org/c/openstack/nova-specs/+/98567613:56
tkajinamas it's not likely we have time to discuss this... I wonder if that namespace package thing is what we broken when we removed pkg_resources ?13:59
tkajinamI intended to keep that functionality but that was later broken by canonical folks insisting "this doesn't work for us"14:00
tkajinamI might be confused by https://review.opendev.org/c/openstack/nova/+/939827 , though14:00
tkajinamok it seems nova.scheduler.filters  has never been exposed as a namespace package14:01
tkajinamwe probably want to use the entrypoints approach instead14:01
opendevreviewBalazs Gibizer proposed openstack/nova master: DNM: test if CLI testing hitting scatter gather  https://review.opendev.org/c/openstack/nova/+/98568114:20
opendevreviewKarol Klimaszewski proposed openstack/nova-specs master: Spec for SPDK-based nova-provisioned storage backend  https://review.opendev.org/c/openstack/nova-specs/+/98567614:30
*** Guest7683 is now known as clarkb14:57
cardoeThanks all for having the Ironic folks.15:04
opendevreviewGoutham Pacha Ravi proposed openstack/nova-specs master: Add spec for virtiofs hot attach and detach  https://review.opendev.org/c/openstack/nova-specs/+/98381615:15
opendevreviewBalazs Gibizer proposed openstack/nova master: DNM: test if CLI testing hitting scatter gather  https://review.opendev.org/c/openstack/nova/+/98568116:43
cardoedansmith: I think my comments were misconstrued a little bit and I'd like to clear things up.18:07
-opendevstatus- NOTICE: The wiki.openstack.org service will be offline for a few minutes while we perform a database upgrade18:07
cardoeThere's a number of users / operators of the nova-ironic driver that I'm aware of. That driver hasn't been a prime focus of the nova cores team so a lot of times patches kinda stall out a little bit. The testing for that driver in Zuul isn't the best as was highlighted during the call.18:09
cardoeSo what I'm doing is trying to get those users / operators to share some changes together that are important to them and trying to figure out the best way for us to voice that "this bug/issues is important to that group"18:10
cardoeThe reason for the repo idea is that more than once it's happened that I've spoken to one of them and they've said "oh we've iterated on the patch a little bit and use a new version". So just trying to come up with a way for them to do that more in the open.18:11
dansmithcardoe: sounds pretty much like what you said on the call (although the context around it at the time is the important but). Sounds like a fork to me, which doesn't seem very constructive...18:17
dansmithbut, you don't need to convince me, I responded to your comment in the moment in the context of the moment18:18
cardoeCan you give me suggestions how we get fixes that are important to the users of nova-ironic landed?18:21
opendevreviewBryan Fraschetti proposed openstack/nova master: Reject to attach and detach volume while task_state is not None  https://review.opendev.org/c/openstack/nova/+/75891818:23
opendevreviewBryan Fraschetti proposed openstack/nova master: Feat: Add Command for Forceful BDM Deletion  https://review.opendev.org/c/openstack/nova/+/98571718:23
opendevreviewBryan Fraschetti proposed openstack/nova master: add task state for attach volume and detach volume  https://review.opendev.org/c/openstack/nova/+/98571818:23
opendevreviewBryan Fraschetti proposed openstack/nova master: Reject Volume Attachment if task_state not Attaching  https://review.opendev.org/c/openstack/nova/+/98571918:23
sean-k-mooneyok im goign to call it a day. ill be in the watcher session most fo tomorrow but if im meeded for a session ping me and ill try an attend18:25
cardoehttps://bugs.launchpad.net/nova/+bug/1884217 https://bugs.launchpad.net/nova/+bug/1732506 (mentioned by johngarbutt during the session) https://bugs.launchpad.net/nova/+bug/2127087 https://bugs.launchpad.net/nova/+bug/2085432 https://bugs.launchpad.net/nova/+bug/2131095 https://bugs.launchpad.net/nova/+bug/2148059 https://bugs.launchpad.net/nova/+bug/211770018:33
cardoeThat's just the list of bugs that I'm aware of that operators have mentioned18:33
cardoeI don't have a list of gerrit patches handy18:35
melwittgmaan: dunno if you might be able to shed any light on this, but in CI the  nova.tests.functional.test_servers.ServerMovingTestsWithNestedResourceRequests.test_resize_reschedule_uses_host_lists_3_fails test has been flaking out and I can't see why,18:44
melwittthe last messages in the traceback are graceful shutdown stuff and I assume they're not related, but I wondered what you might think. example: https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_af0/openstack/af0437afd3354ef990fcb3d6aa47f80a/testr_results.html18:44
*** erlon0 is now known as erlon19:33
opendevreviewGoutham Pacha Ravi proposed openstack/nova-specs master: Add spec for virtiofs hot attach and detach  https://review.opendev.org/c/openstack/nova-specs/+/98381620:12
gmaanlet me check21:52
gmaanmelwitt: it seems all nova services are shutdown in between of tests.  22:13
gmaanfunctional tests override the graceful shutdown to 0 so shutdown will be immediately without waiting for things to finish and that is why server is stuck in that state22:14
gmaannot sure why shutdown happen22:14
melwittgmaan: thanks for looking. so this is unexpected? that is a good clue for me to continue debugging 22:15
gmaanyeah, shutdown itself should not be initiated in between of test22:16
opendevreviewGoutham Pacha Ravi proposed openstack/nova-specs master: Add virtiofs live migration spec  https://review.opendev.org/c/openstack/nova-specs/+/98380122:57
opendevreviewGoutham Pacha Ravi proposed openstack/nova-specs master: Add spec for virtiofs live migration  https://review.opendev.org/c/openstack/nova-specs/+/98380123:01
opendevreviewGoutham Pacha Ravi proposed openstack/nova-specs master: Add virtiofs cold migration spec  https://review.opendev.org/c/openstack/nova-specs/+/98573823:01

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