Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2024-07-03
22:02:46 sean-k-mooney melwitt: ah good point
22:11:54 fungi just got back
22:12:52 fungi assembling the list now
22:23:09 fungi mmm, this is going to be tricky because i have to pass in both the project name and the change,revision for each change i evict and re-add, so a simple loop won't suffice
22:23:12 dansmith yeah, that one is required for the glance on the same branch
22:25:21 dansmith poor glance hasn't merged anything yet, but the head change is in gate right now
22:26:19 fungi should https://review.opendev.org/c/openstack/glance/+/923433 be preserved too? it switches a job to non-voting
22:26:39 dansmith fungi: this got kicked out of gate but should be able to right back in: https://review.opendev.org/c/openstack/nova/+/923258/3
22:26:43 dansmith and the parent of it is in gate now I think
22:27:02 fungi i can insert it between the dequeue and reenqueue steps, sure
22:27:30 dansmith fungi: I think abhi still wanted that mostly to insulate them from job timeouts, but it's not strictly required and I think it'd be better if we don't drop all that testing now that tempest is merged
22:27:43 fungi wfm
22:27:43 dansmith so you'd have my blessing to kick that out and make space, or I can if you want
22:28:27 fungi adding it to the list of stuff i'll move to the back of the line
22:28:32 dansmith ++
22:36:34 fungi okay, surgery complete
22:37:10 fungi though in retrospect, as none of the changes at the top were being kept, it ended up being the same as if i'd done a promote of the ossa changes i guess
22:45:50 fungi i guess i expected there to be more changes from the ossa already approved
22:46:11 dansmith meaning the stable ones?
22:46:18 fungi turned out there were only 5 in the gate (plus the one that was kicked out and re-added)
22:46:20 dansmith all the master ones should be approved I think
22:46:46 dansmith given how things have been going with the insane cross-dependencies I think we've been kinda trying to make sure they are merging in release order
22:47:01 dansmith but if you really want them all +Wd we can probable just do that
22:47:15 dansmith I've heard great things about melwitt's +W finger
22:47:15 fungi nah, it's okay
22:47:27 melwitt lol
22:47:43 melwitt yup, I'm really great at clicking buttons
#openstack-nova - 2024-07-04
03:06:23 opendevreview Merged openstack/nova master: Additional qemu safety checking on base images https://review.opendev.org/c/openstack/nova/+/923257
06:13:40 opendevreview Merged openstack/nova stable/2023.1: Fix disk_formats in ceph job tempest config https://review.opendev.org/c/openstack/nova/+/923344
07:19:25 opendevreview Maxime Lubin proposed openstack/nova-specs master: USB over IP https://review.opendev.org/c/openstack/nova-specs/+/920687
07:25:43 frickler more gate resets happening all the time. I'll regate https://review.opendev.org/c/openstack/nova/+/923258 now, but not promote in order to give the glance stack a chance first
07:28:37 abhishekk thank you!
07:32:37 opendevreview Maxime Lubin proposed openstack/nova-specs master: USB over IP https://review.opendev.org/c/openstack/nova-specs/+/920687
07:49:25 amorin hey team, just wondering if there is any way to start nova-conductor when nova-placement is down? It seems nova-conductor refused to start if placement API is not up
07:49:49 amorin openstack.exceptions.NotSupported: The placement service for keystone.x.y.z xists but does not have any supported versions.
07:55:21 frickler amorin: what would be the use case for that? I don't think the conductor can do anything useful without placement?
07:56:58 amorin I mean, maybe it can start anyway, and try to reconnect later, at least it will start and answer to computes that need some data
08:06:52 frickler amorin: hmm, maybe, but then let me ask a different question: in what scenario would you expect your placement service to be down for an extended period? or would just retrying a couple of times at startup be enough?
08:19:09 amorin agree that this scenario of placement down is wrong
08:20:44 amorin I am just surprised that this is the behavior of nova-conductor.
08:27:03 amorin my good old stein openstack was not complaining about this :)
08:55:15 frickler did that even have the split placement service yet? ;)
08:57:00 amorin yes, placement was splitted (at least in our deployment)
09:01:40 gibi amorin: I think by https://github.com/openstack/nova/commit/232684b44022f1bc4d72b07045900780de456e63
09:01:49 gibi it shoud not happen any more
09:02:02 bauzas this is indeed an expected behaviour
09:11:35 amorin perfect!
09:12:08 amorin will try cherry-picking that to my antelope openstack then
09:12:21 amorin but wait, it should be there
09:15:44 amorin the exeption raised is the following: exceptions.NotSupported
09:16:17 amorin which is not in the first except block, so it's conesidered as a fatal one
09:16:47 amorin my service is stop and returning http 503
09:26:17 gibi amorin: do you have placement in openstack endpoint list?
09:32:55 frickler amorin: that sounds like the keystone lib might be mishandling that error
09:35:12 amorin yes gibi, it's all set correctly, as soon as I start placement, everything works fine :)
09:37:01 frickler frickler: can you try to run something like "openstack --debug resource provider list" while placement is down and show the output on paste.opendev.org? (and possibly let's move this to #openstack-sdks)
09:37:09 frickler ehm amorin ^^
09:51:05 sean-k-mooney amorin: this is a similar but slightly diffent issue
09:51:43 sean-k-mooney the orginal fix was to allow it to start when there was a keyston outaage not placment actully down
09:51:58 sean-k-mooney i think we have a diffent patch to make loading the placement client lazy
09:52:15 sean-k-mooney so that provided you dont make an api call that need it to talk to palcment it wont try
09:52:38 sean-k-mooney we worked on that for the schduler too but i dont recall whicih serviecs got that change and whihc did not
09:53:36 sean-k-mooney https://review.opendev.org/c/openstack/nova/+/878238 that was the scheduler change
09:56:20 sean-k-mooney i guess it thet same apprch https://github.com/openstack/nova/commit/232684b44022f1bc4d72b07045900780de456e63 make the report clint a property and lazy loads it too
09:58:28 sean-k-mooney my guess is that a "503 Service Unavailable" responce does not count as ConnectionFailure and we just need to add that to the first expection block to cover your edgecase
10:10:53 frickler seems there is another different set of error classes for that, like in this case keystoneauth1.exceptions.http.ServiceUnavailable
10:15:01 sean-k-mooney ya so we just need to add them
10:15:23 sean-k-mooney with that said nova does expect placment ot be aviabel when wee start up we just added this lazy loading to make it more graceful
10:16:05 sean-k-mooney if placement is down for a protacted period of tiem we want to allow you to manage things that dont require interaction with placement but obvioulsy you wont be abel to boot new vms, evaucate them or move them in any way
10:16:36 sean-k-mooney start/stop/snapshot/attache port ectra is ok
10:17:02 sean-k-mooney but we would consdier the openstack to be in a degreated state and expect some operations ot fail
10:20:52 opendevreview Stephen Finucane proposed openstack/placement master: tox: Simplify functional testenv definitions https://review.opendev.org/c/openstack/placement/+/919578
10:20:53 opendevreview Stephen Finucane proposed openstack/placement master: pre-commit: Add sphinx-lint https://review.opendev.org/c/openstack/placement/+/919580
10:20:53 opendevreview Stephen Finucane proposed openstack/placement master: Integrate pre-commit https://review.opendev.org/c/openstack/placement/+/919579
10:20:54 opendevreview Stephen Finucane proposed openstack/placement master: Fix typos https://review.opendev.org/c/openstack/placement/+/923491
10:20:54 opendevreview Stephen Finucane proposed openstack/placement master: pre-commit: Add autopep8 https://review.opendev.org/c/openstack/placement/+/923490
11:37:25 sean-k-mooney sigh.... https://github.com/al45tair/netifaces netifaces is not buildable on debian testin on python 3.12... we need to drop that depency form nova
11:37:33 sean-k-mooney it works on 3.11
11:42:24 opendevreview sean mooney proposed openstack/nova master: port format insepctor tests from glance https://review.opendev.org/c/openstack/nova/+/923504
11:42:55 opendevreview sean mooney proposed openstack/nova master: port format inspector tests from glance https://review.opendev.org/c/openstack/nova/+/923504
11:52:10 frickler hmm, seems at least canonical managed to make it work https://launchpad.net/ubuntu/+source/netifaces
11:52:50 sean-k-mooney it works on 3.11 but it raise an error where it say "you must add supprot for your plathform"
11:52:58 sean-k-mooney ill get the exact message in a sec
11:53:07 sean-k-mooney it might be because im using debian testing
11:53:18 sean-k-mooney but the project is archivbed and not maintaned anymore
11:53:25 sean-k-mooney so we need to replace it anyway
11:54:00 sean-k-mooney i dont thnk nova really should have thsi dep anymore anyway it was mainly for nova-net before and os-vif on windows i think
11:54:07 frickler ack. I'm just wondering why the py312 tests don't seem to notice it
11:54:09 sean-k-mooney so we can likely jsut remove the code that is using it
11:55:54 sean-k-mooney https://paste.opendev.org/show/bzSoczPxEGhAlGJPfcCc/
11:56:03 sean-k-mooney thats what im seeing on 3.12
11:56:13 opendevreview sean mooney proposed openstack/nova master: port format inspector tests from glance https://review.opendev.org/c/openstack/nova/+/923504
11:57:51 sean-k-mooney ok this is what is firing https://github.com/al45tair/netifaces/blob/master/netifaces.c#L206-L211
11:59:22 frickler hmm, can you check why pip is wanting to build the wheel, are you on arm? on noble I just get "Using cached netifaces-0.11.0-cp312-cp312-linux_x86_64.whl"
11:59:49 sean-k-mooney debian testing on x86
12:00:29 sean-k-mooney x86_64 to be precise
12:00:45 sean-k-mooney how would i check that
12:01:01 frickler good question, add some "-v"'s?
12:01:16 sean-k-mooney sure i can add some to tox
12:01:29 sean-k-mooney well the install line in tox.ini

Earlier   Later