| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2023-07-18 | |||
| 12:46:26 | gibi | yeah I agree with sean-k-mooney on both patch | |
| 13:03:30 | opendevreview | Kashyap Chamarthy proposed openstack/nova master: Bump MIN_{LIBVIRT,QEMU} for "Bobcat" https://review.opendev.org/c/openstack/nova/+/887255 | |
| 13:49:07 | stephenfin | sean-k-mooney: gibi: The issue is that we are currently SQLAlchemy 2.x compatible and don't want to regress. However, people don't look at non-voting jobs so they're effectively useless for anti-regression purposes IMO | |
| 13:50:06 | stephenfin | There is of course a risk that SQLAlchemy could introduce a breaking change but if we're serious about being ready for SQLA 2.x by C then we'll need to handle those issues. Also, SQLA is super mature so the chances of that seem minute, while we can simply toggle it to non-voting if they inadvertently introduce a bug | |
| 13:52:54 | stephenfin | Also also, I've proposed the same thing for ironic, cinder, placement, and a few others. Neutron are running all their functional tests with sqlalchemy master for some time now. I think the benefits outweigh the concerns | |
| 13:56:21 | bauzas | sean-k-mooney: I quite agree with stephenfin, let's KISS | |
| 13:56:28 | bauzas | we also have other jobs | |
| 13:58:52 | bauzas | but -1 b/c of the gate | |
| 13:59:56 | dansmith | could we not just make the 310 job SA 2.0 and leafve the 3.9 one as it is? having another job seems excessive to me | |
| 14:00:36 | sean-k-mooney | i would prefer to just raise our min version to 2.0 | |
| 14:00:40 | auniyal | hi dansmith, sorry for not responding on your question on time before new revision, yes as you mentioned in comment, on melwitt suggestion, I passed the context from caller and did not created new context and I added the assert. as per suggestion. Actually I had the reply on draft but was waiting for CI job to pass that time. | |
| 14:00:54 | sean-k-mooney | but ya i can also live with dansmith suggestion i guess | |
| 14:01:01 | gibi | yeah, if we want to track 2.0 then lets bump now to 2.0 | |
| 14:01:05 | dansmith | we're really running three functional jobs and it seems like we could just re-use one to move forward | |
| 14:01:34 | sean-k-mooney | could we bump the uc for py3.10 specificaly to 2.0 | |
| 14:01:55 | sean-k-mooney | then we could follow the same pattern in all project | |
| 14:02:09 | dansmith | can we do that? if so that'd also work I suppose | |
| 14:02:12 | sean-k-mooney | or as i said just move nova to 2.0 | |
| 14:02:23 | dansmith | I mean I know requirements.txt can do it, but will a bot complain? | |
| 14:02:33 | sean-k-mooney | dansmith: we can we used ot do it for py2.7 vs py 3 | |
| 14:02:42 | dansmith | sean-k-mooney: we can't move nova to *require* 2.0 right now | |
| 14:02:55 | dansmith | because it would mean nova can't be installed on the same machine as keystone, for example | |
| 14:03:17 | stephenfin | we can't bump to 2.0 in requirements.txt - the requirements check job won't let us (co-installability) | |
| 14:03:27 | stephenfin | i.e. what dansmith said | |
| 14:03:46 | sean-k-mooney | i ment in the reqiuremetns repo | |
| 14:04:00 | sean-k-mooney | can we increases the cap to 2.0 for py3.10 | |
| 14:04:07 | dansmith | yeah, can't do that until everyone can use 2.0 but we're not ready for that | |
| 14:04:09 | sean-k-mooney | i guess that could break devstack jobs | |
| 14:04:17 | sean-k-mooney | ack | |
| 14:04:19 | stephenfin | sean-k-mooney: manila still isn't compatible, nor is masakari and a few other projects | |
| 14:04:31 | dansmith | nor keystone, last I checked | |
| 14:04:47 | sean-k-mooney | ok so i guess looping back to dansmith suggestion of just doing this in one of the existing job | |
| 14:04:55 | sean-k-mooney | its that prefer or having the extra job | |
| 14:05:03 | bauzas | ok cool with me too | |
| 14:05:16 | gibi | OK with me to force py310 to run with 2.0 | |
| 14:05:32 | dansmith | I just think the actual likelihood of breakages between 3.9 and 3.10 is minor, so using 3.10 as our 2.0 job makes frugal sense, | |
| 14:05:40 | stephenfin | Also, we can't just reuse a job unless we want to get rid of our use of job templates | |
| 14:05:44 | dansmith | and we can probably spot the difference between SA problems and some python base issue | |
| 14:05:54 | dansmith | stephenfin: ah, you're right | |
| 14:06:05 | dansmith | fine, fine, just stack another job I guess | |
| 14:06:06 | stephenfin | because it's the job templates that specifies our use of openstack-tox-whatever | |
| 14:06:24 | dansmith | yup | |
| 14:06:33 | bauzas | ~o~ | |
| 14:07:09 | stephenfin | ack, with any luck it should be gone in 3 months or so 🤞 hopefully sooner but definitely in C | |
| 14:07:09 | stephenfin | ack, with any luck it should be gone in 3 months or so 🤞 hopefully sooner but definitely in C | |
| 14:07:25 | sean-k-mooney | ok | |
| 14:07:32 | sean-k-mooney | so we will proceed with the patch as is | |
| 14:07:43 | dansmith | I will bet you one unit of currency that it will not be gone in three months :P | |
| 14:08:36 | stephenfin | I value my currency and will not take that bet | |
| 14:08:43 | dansmith | hehe | |
| 14:09:04 | gibi | :) | |
| 15:14:34 | bauzas | reminder : nova meeting in 45 mins here | |
| 15:36:32 | opendevreview | Merged openstack/nova stable/wallaby: Reproducer for bug 1983753 https://review.opendev.org/c/openstack/nova/+/853273 | |
| 15:48:07 | dansmith | sean-k-mooney: got another guest panic this morning: https://27abb73f6979e2fd6932-3e445c1a5b8a8c84f71b5a61cff5a09b.ssl.cf1.rackcdn.com/879499/10/check/nova-multi-cell/5c92eb8/testr_results.html | |
| 15:50:01 | dansmith | looks like *very* early in kernel boot | |
| 15:55:20 | gibi | its smells like APIC https://bugs.launchpad.net/nova/+bug/1939108 | |
| 15:57:53 | dansmith | gibi: I dunno that bug trace made it much farther than mine did.. note all the kernel timestamps are 0.0 | |
| 15:58:25 | dansmith | but yeah I guess the timer failed messages before the crash look the same | |
| 15:58:39 | dansmith | so maybe it's just a difference in the kernel from then to now | |
| 15:59:32 | dansmith | and in check_timer() where we panic, so okay yeah | |
| 16:00:12 | dansmith | and another report from 2022 at the end of that report that looks more like mine | |
| 16:00:23 | bauzas | starting the meeting now folks | |
| 16:00:29 | gibi | both calls setup_IO_APIC that was my matching point | |
| 16:00:31 | dansmith | kashyap: I wonder if it might be worth chasing a kernel bug with that now | |
| 16:00:44 | bauzas | let's discuss this gate failure in a sec | |
| 16:00:50 | bauzas | #startmeeting nova | |
| 16:00:50 | opendevmeet | Meeting started Tue Jul 18 16:00:50 2023 UTC and is due to finish in 60 minutes. The chair is bauzas. Information about MeetBot at http://wiki.debian.org/MeetBot. | |
| 16:00:50 | opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | |
| 16:00:50 | opendevmeet | The meeting name has been set to 'nova' | |
| 16:00:55 | bauzas | hey folks | |
| 16:01:02 | gibi | o/ | |
| 16:01:04 | bauzas | #link https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting | |
| 16:01:45 | elodilles | o/ | |
| 16:01:49 | bauzas | who's around ? | |
| 16:02:59 | bauzas | okay, we can softly start | |
| 16:03:14 | bauzas | #topic Bugs (stuck/critical) | |
| 16:03:20 | bauzas | #info No Critical bug | |
| 16:03:31 | bauzas | #link https://bugs.launchpad.net/nova/+bugs?search=Search&field.status=New 40 new untriaged bugs (+1 since the last meeting) | |
| 16:03:41 | bauzas | thanks melwitt for the triaging | |
| 16:04:20 | bauzas | any bug to discuss ? | |
| 16:04:25 | Uggla | o/ | |
| 16:05:00 | bauzas | looks not, moving on | |
| 16:05:04 | bauzas | #info Add yourself in the team bug roster if you want to help https://etherpad.opendev.org/p/nova-bug-triage-roster | |
| 16:05:23 | bauzas | IIRC, artom is on PTO this week, so he can't take the baton | |
| 16:05:37 | bauzas | Uggla: wants to have it ? | |
| 16:06:06 | Uggla | yes it is ok | |
| 16:06:11 | bauzas | cool thanks | |
| 16:06:19 | bauzas | #info bug baton is being passed to Uggla | |
| 16:06:51 | bauzas | next topic | |
| 16:06:53 | bauzas | #topic Gate status | |
| 16:06:57 | bauzas | #link https://bugs.launchpad.net/nova/+bugs?field.tag=gate-failure Nova gate bugs | |
| 16:07:03 | bauzas | #link https://etherpad.opendev.org/p/nova-ci-failures | |
| 16:07:15 | bauzas | gibi: wants to discuss the kernel panic bug you found ? | |
| 16:07:26 | gibi | that is dansmith finding | |
| 16:07:42 | gibi | I just matched it to an older report | |
| 16:08:04 | gibi | I guess we can try the apic debug kernel param | |
| 16:08:31 | bauzas | ack | |
| 16:08:58 | gibi | as the kernel error message suggests | |
| 16:08:59 | gibi | Kernel panic - not syncing: IO-APIC + timer doesn't work! Boot with apic=debug and send a report. Then try booting with the 'noapic' option. | |
| 16:09:52 | bauzas | any link to share . | |