| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2025-06-10 | |||
| 16:28:06 | tkajinam | hm ok. I believe the detection feature was merged in the past but I'll recheck if there is anything remaining | |
| 16:28:26 | sean-k-mooney | i know we merged the swtpm and version detection | |
| 16:28:41 | sean-k-mooney | i dont recall if we also merged the patch to auto detect the capsity | |
| 16:28:44 | sean-k-mooney | although maybe we did | |
| 16:29:03 | tkajinam | I believe so but I'll double-check it | |
| 16:29:27 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Address issues with keypairs APIs https://review.opendev.org/c/openstack/nova/+/952214 | |
| 16:29:29 | sean-k-mooney | ack, so your next priority would be the patches to extend that to SEV-ES right | |
| 16:29:52 | tkajinam | these are not really related but yeah sev-es would be the next one I may want to have. | |
| 16:30:15 | sean-k-mooney | ack | |
| 16:30:18 | tkajinam | that helps pending sev-snp as well as arm-cca which some folks from fujitsu have been working on | |
| 16:30:38 | tkajinam | I was supposed to submit spec for sev-snp but I'm not very sure if I can push one quite soon | |
| 16:31:48 | sean-k-mooney | no worries just askign for context | |
| 16:31:54 | tkajinam | :-) | |
| 16:32:33 | Uggla | I guess nobody expects SEV-SNP this cycle. | |
| 16:33:08 | tkajinam | yeah, though I have got a few local pings about my plan about it. | |
| 16:33:48 | tkajinam | anyway we can follow up separately about sev thing and can wrap up the meeting unless anyone else have their topics. (assuming not many people may be interested in such nich feature :-P ) | |
| 16:34:15 | Uggla | yep, something else before moving to bug scrubbing. | |
| 16:34:31 | tkajinam | ahh yes | |
| 16:35:17 | Uggla | #topic Bug scrubbing | |
| 16:35:26 | Uggla | #link: https://etherpad.opendev.org/p/nova-bug-selection-for-triaging#L4 | |
| 16:35:37 | Uggla | #link: https://truc.uggla.fr/ I kicked off this draft page to follow open bugs. | |
| 16:35:45 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for server group APIs https://review.opendev.org/c/openstack/nova/+/952281 | |
| 16:35:45 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for removed APIs https://review.opendev.org/c/openstack/nova/+/952282 | |
| 16:35:46 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for networks API https://review.opendev.org/c/openstack/nova/+/952283 | |
| 16:35:46 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for images APIs https://review.opendev.org/c/openstack/nova/+/952284 | |
| 16:36:47 | Uggla | First: | |
| 16:36:47 | Uggla | https://bugs.launchpad.net/nova/+bug/2112808 - nova-compute restart breaks queued live migration objects (artom) | |
| 16:37:50 | Uggla | I think this is probably valid. sean-k-mooney do you know about it ? | |
| 16:38:54 | sean-k-mooney | so this came form a 16.2 bug report i.e. train | |
| 16:39:24 | sean-k-mooney | i know we did some work to improve this since then so before considering it valid we likely need to see if it can happen in master | |
| 16:39:48 | Uggla | yes it seems to be the case. Artom mentioned that it is probably applicable on master. | |
| 16:39:53 | sean-k-mooney | with that said there probely is a bug in the init_host logic that is ment to heal migrations on start up | |
| 16:40:25 | sean-k-mooney | so it may be but the bug does nto really meet or minium stnadard for reporting a bug :) | |
| 16:40:55 | Uggla | I know. :) | |
| 16:41:00 | sean-k-mooney | https://issues.redhat.com/browse/OSPRH-13122 is public but you should not have to click into downstream tackers to asses the bugs valididity | |
| 16:41:42 | sean-k-mooney | anyway the downstream bug suggest the bug is here https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L1040-L1057 | |
| 16:41:42 | Uggla | Is it possible to create a reproducer with fn tests in that case ? | |
| 16:42:08 | sean-k-mooney | and the logic that its only catherign for runnign mighrations | |
| 16:42:08 | sean-k-mooney | looks correct | |
| 16:42:22 | sean-k-mooney | so this proably is a bug | |
| 16:42:38 | sean-k-mooney | that affect master but ya we need a proper repoducer fo this to fix it | |
| 16:43:06 | sean-k-mooney | and yes our funcitonal tests shoudl be able to repoduce this | |
| 16:43:35 | sean-k-mooney | there is one slight complciation with how restarts work in the functional enve but we have some exampels of this already | |
| 16:44:13 | sean-k-mooney | lets mark it as triaged and add a permalink to the relevnet code https://github.com/openstack/nova/blob/517415b6cbdd1730b5179b714229cfc0a65b1024/nova/compute/manager.py#L1043-L1060 | |
| 16:44:49 | Uggla | ok good. That works for me. I'll do that later. Thanks | |
| 16:45:14 | Uggla | Next one: | |
| 16:45:14 | Uggla | ttps://bugs.launchpad.net/nova/+bug/2112373 - Flavor Extra Specs for SCSI Controller Model Not Honored for Volume Boot | |
| 16:45:27 | Uggla | https://bugs.launchpad.net/nova/+bug/2112373 - Flavor Extra Specs for SCSI Controller Model Not Honored for Volume Boot | |
| 16:47:24 | Uggla | That looks valid too. I may encourage the reporter to submit a patch because he proposed something in the report. | |
| 16:48:32 | sean-k-mooney | that is valid, i fixed that for non BFV instnaces years ago | |
| 16:48:44 | sean-k-mooney | but i never fixed BFV | |
| 16:48:55 | sean-k-mooney | the porblem is with the way we do mappings | |
| 16:49:08 | sean-k-mooney | basiclly we mapp sata and scsi to sd* | |
| 16:49:24 | sean-k-mooney | and as a result we end up with the the wrong type or something like that | |
| 16:49:58 | sean-k-mooney | this is a old know issue | |
| 16:50:12 | sean-k-mooney | we have a downstream tracker for this somewhere i think | |
| 16:50:27 | sean-k-mooney | oh | |
| 16:50:35 | sean-k-mooney | actully this is sligly diffent maybe | |
| 16:50:48 | sean-k-mooney | hw:scsi_model i think is only an imave property | |
| 16:51:16 | sean-k-mooney | what i was descibing was the bug when you have that image peropery on the volume in the volume_image_propties filed | |
| 16:51:34 | sean-k-mooney | i dont recally if we ever added hw:scsi_model and hw:disk_bus | |
| 16:51:43 | sean-k-mooney | as flavor extra specs | |
| 16:51:54 | sean-k-mooney | the were orginaly only image properties | |
| 16:52:09 | sean-k-mooney | that woudl make this a feature request not a bug | |
| 16:52:29 | sean-k-mooney | let me check the valdiation code | |
| 16:52:57 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/api/validation/extra_specs/hw.py | |
| 16:53:04 | sean-k-mooney | so ya this is not soemthing we supprot today | |
| 16:54:18 | Uggla | What do you think about reporter fallback ? | |
| 16:54:25 | sean-k-mooney | we coudl add the rfe tag and make it a wishlest bug | |
| 16:54:40 | sean-k-mooney | but it is a new feature and soemthign we intentiaonly did not suprpot in the past | |
| 16:55:12 | sean-k-mooney | Uggla: historically flavor extra specs were for capstyisty and imave proprites were used for things liek device model or busses | |
| 16:55:29 | sean-k-mooney | the logic is that only the image know what type of virutal device it needs | |
| 16:56:11 | sean-k-mooney | and since some clodus do not show flavor extra specs to user allowing it to be set in the flavor coudl break enduses expcations. | |
| 16:56:33 | sean-k-mooney | with that said we have relaxed that over the years and now allow many thing to be set in either | |
| 16:57:44 | sean-k-mooney | i can write somethign up but really this shoudl be a small specless blueprint or as i said we coudl consider it a wishlist bug with the rfe tag | |
| 16:58:34 | Uggla | or we can close the bug and ask reporter to propose something in next PTG ? | |
| 16:59:46 | sean-k-mooney | we could | |
| 16:59:54 | sean-k-mooney | i added a comment to the bug | |
| 17:00:05 | sean-k-mooney | do we want to revisit it next week after they have time to respond | |
| 17:00:26 | Uggla | sure. | |
| 17:00:34 | Uggla | thanks | |
| 17:01:22 | Uggla | We are on the top of the hour. So closing the meeting. | |
| 17:01:51 | Uggla | Thanks all, see you next week. | |
| 17:01:52 | Uggla | #endmeeting | |
| 17:01:52 | opendevmeet | Meeting ended Tue Jun 10 17:01:52 2025 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | |
| 17:01:52 | opendevmeet | Minutes: https://meetings.opendev.org/meetings/nova/2025/nova.2025-06-10-16.01.html | |
| 17:01:53 | opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/nova/2025/nova.2025-06-10-16.01.txt | |
| 17:01:53 | opendevmeet | Log: https://meetings.opendev.org/meetings/nova/2025/nova.2025-06-10-16.01.log.html | |
| 17:02:22 | sean-k-mooney | Uggla: that other bug i initally descibe is a real bug with bfv or addtional data volumes by the way i just dont have a link to it | |
| 17:02:51 | Uggla | ok | |
| 17:03:17 | Uggla | sean-k-mooney, I will look at your comment and update the bugs, probably tomorrow morning. | |
| 17:03:38 | Uggla | sean-k-mooney, thanks for your help. | |
| 17:03:42 | sean-k-mooney | oh its https://bugs.launchpad.net/nova/+bug/2022093 | |
| 17:04:50 | sean-k-mooney | nova is only ment ot read the prories form the root disk so setting hw_disk_bus=sata is a redherring there but | |
| 17:04:56 | sean-k-mooney | its still incorrect beahvior | |
| 17:05:30 | sean-k-mooney | all disks shoudl bave been added with sata if the disk type was not cdrom | |
| 17:07:45 | Uggla | ok I'll read this bug with a fresh mind. | |
| 17:07:59 | tkajinam | sean-k-mooney, fyi. full series of tpm feature detection was merged https://review.opendev.org/q/topic:%22bp/libvirt-detect-vtpm-support%22 and now nova detects (1) tpm support (2) tpm models (3) tpm versions via libvirt API | |
| 17:08:27 | tkajinam | (1) and (2) requires libvirt 8.0.0 so now I can clean up fallback for these, while (3) requires 8.6.0 so slight logic may be kept | |
| 17:10:21 | sean-k-mooney | tkajinam: ack sound good | |