Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2022-08-16
16:20:46 sean-k-mooney after it was released
16:20:54 elodilles (i mean, with the intermittent failures - volume operations timeout)
16:21:09 elodilles sean-k-mooney: yes
16:21:25 elodilles sean-k-mooney: and at that time it was not really needed for train
16:21:45 sean-k-mooney ack so now that it is are you ok to proceed in this direction?
16:22:01 elodilles but now, somehow it seems it is not facing with the issue we have with the old devstack-gate jobs
16:22:10 elodilles sean-k-mooney: yes
16:22:20 elodilles and that means we can unblock train at least
16:22:21 sean-k-mooney ack then ill try and find time to review it
16:22:47 elodilles stein rocky and queens will be still blocked, though
16:22:55 elodilles sean-k-mooney: ++
16:23:13 elodilles but probably we can even EOL those branches
16:23:26 elodilles as there are no interested parties around those branches :/
16:23:35 gibi yeah it seems
16:23:47 elodilles anyway, this is all i think
16:24:56 gibi thanks elodilles
16:25:02 elodilles np
16:25:04 gibi #topic Open discussion
16:25:09 gibi there is noting on the agenda
16:25:16 gibi so speak up now if you have someting
16:25:37 elodilles -
16:26:28 gibi dont be so negative elodilles ;)
16:26:38 elodilles gibi: ++
16:26:39 elodilles :D
16:26:40 JayF Hey Nova folks; I have a reasonably old Ironic driver backport to stable/victoria, I just found out someone has an install that could use it, if we could land it, I'd appreciate it. https://review.opendev.org/c/openstack/nova/+/800873 (elodilles reviewed it a while back, but I didn't update it for a while because I was between openstack-y jobs)
16:26:50 JayF whoops, didn't mean to send that until the meeting was over, sorry
16:27:17 gibi JayF: that is a good open discussion topic so no worries :)
16:28:04 gibi JayF: I will hit that tomorrow morning and I hope elodilles will do it too
16:28:34 sean-k-mooney JayF: looks like its already merged in stable wallaby
16:28:38 JayF thank you both very much! I will be doing a review of anything that might need backporting to victoria. If that one merges, you might get more :)
16:28:41 sean-k-mooney and that the victoria backport
16:28:46 sean-k-mooney so i dont think that is contoverial
16:29:25 sean-k-mooney the ci logs are mising but they were green
16:29:47 sean-k-mooney so provide this does not need a rebase i think we can just review it as normal
16:29:47 JayF That's what I'm hoping, but I know folks are busy and getting attention is often the difficult part. Just had no real reason to promote it until I met an Ironic user today who is having this specific issue and is already mid-upgrade to Victoria :)
16:30:06 sean-k-mooney ack
16:30:24 sean-k-mooney ill add it to my list
16:32:01 gibi I feel we are done with the today's meeting
16:32:06 gibi thank you all for joining
16:32:08 opendevmeet Log: https://meetings.opendev.org/meetings/nova/2022/nova.2022-08-16-16.00.log.html
16:32:08 opendevmeet Minutes (text): https://meetings.opendev.org/meetings/nova/2022/nova.2022-08-16-16.00.txt
16:32:08 opendevmeet Minutes: https://meetings.opendev.org/meetings/nova/2022/nova.2022-08-16-16.00.html
16:32:08 opendevmeet Meeting ended Tue Aug 16 16:32:08 2022 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)
16:32:08 gibi #endmeeting
16:32:32 sean-k-mooney JayF: its a clean cherry pick and the previous branch is already merged so +2
16:33:23 sean-k-mooney its been a month since zuul last ran but we have not been merging code to victoria much so i expect this should still apply
16:33:37 JayF I'll watch it, if it kicks out V-2 I will take a look
16:36:03 elodilles JayF: thanks for the update, +2'd it
16:36:32 sean-k-mooney elodilles: care to +w too
16:36:51 sean-k-mooney or did you want to hold if for some reason
16:37:19 sean-k-mooney also do assert not fire in our functional tests by default
16:39:52 elodilles sean-k-mooney: nope, just haven't seen that you +2'd already :) but we can wait for the tests to pass again
16:40:07 sean-k-mooney ack
16:40:13 sean-k-mooney oh gibi rechecked
16:40:25 sean-k-mooney just saw that now
16:40:58 elodilles sean-k-mooney: on the other hand, +W probably doesn't hurt, so that it can merge quicker (and we won't forget about +W it later)
16:41:02 gibi sean-k-mooney: yeah I kicked it as I figured I will look at it tomorrow but if you are both agree I trust you so feel free to +A it
16:41:24 sean-k-mooney well the ci will reject it if there is an issue
16:41:32 sean-k-mooney but we can take a look later anyway
16:41:43 elodilles sean-k-mooney: what do you mean by the functional test thing?
16:42:01 sean-k-mooney im debuging why one of my func tests is not working
16:42:12 sean-k-mooney i think the libvirt fixture is incomplete
16:42:27 sean-k-mooney and its not setting something that is always set with real libvirt
16:42:39 sean-k-mooney but my assert is not firing so i think the funciton im looking at is mocked
16:44:19 elodilles but i guess that's a different patch (on master branch?)
16:44:48 sean-k-mooney yes differnt patch
16:45:04 sean-k-mooney its what i was lookign at before the meeting started
16:45:32 sean-k-mooney if dev_path not in (None, ''):
16:45:33 elodilles i see, i thought you want to -W this patch due to some failure, but then it's a different thing
16:45:34 sean-k-mooney raise TypeError()
16:45:36 sean-k-mooney assert dev_path not in (None, ''), (
16:45:38 sean-k-mooney "dev_path is required to be a non empty path "
16:45:40 sean-k-mooney f"called with: {dev_path}")
16:45:42 sean-k-mooney no
16:45:48 sean-k-mooney so the assert there has no effect in our functional tests
16:45:55 sean-k-mooney but adding the raise breaks them
16:46:08 sean-k-mooney so at least i know the fucntion is being called but asserts are off
16:46:28 sean-k-mooney i had expected tox/stestr to trun asserts on by default
16:46:33 sean-k-mooney i wonder if we should change that
16:47:06 sean-k-mooney and jsut to be 100% clear the ironic patch looks fine no reason to -w
16:47:25 elodilles sean-k-mooney: ack
16:49:25 elodilles sean-k-mooney: i haven't used 'assert', ussually self.assert* is used everywhere
16:50:36 sean-k-mooney i wanted to put it in real code rather then in the tests
16:50:47 sean-k-mooney but its fine this is just for me to fidn where its broken
16:51:54 elodilles oh, i thought you added it in a functional test
16:53:28 gibi sean-k-mooney: interesing find. I agree that we should turn on assert in func test
17:27:52 JayF https://review.opendev.org/c/openstack/nova/+/800873 seems to have cleared CI; if someone wants to toss a workflow+ on it
17:59:25 mloza2 hi, is there way for nova to autostart specific VMs when compute comes up from a crash?
18:04:18 melwitt gibi: ack re: triage this week, I will do it. thanks!
18:24:55 sean-k-mooney gibi: i think there is a python debug envar or something like that to enable them in a non interactive interperter
18:25:48 elodilles JayF: actually it wasn't at that time ('recheck' does not clear the Verified+1), but now it is and +W'd (fun fact, that when I saw in the console log that tests have passed, I +W'd it and gate jobs started to run simultaneously with check jobs, because of the fact that i wrote: 'recheck' does not clear Verified+1)
18:25:55 sean-k-mooney but if asserts do fire in func/unit tests it would allow use to capture preconditions in the code with asserts and if they are ever violated in a unit or func test it would be caught.
18:26:10 elodilles JayF: anyway, hopefully it merges soon :)
19:10:16 sean-k-mooney mloza2: there is a host level config option
19:10:40 sean-k-mooney mloza2: https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.resume_guests_state_on_host_boot
20:26:40 opendevreview Elod Illes proposed openstack/nova stable/train: [CI] Fix gate by using zuulv3 live migration and grenade jobs https://review.opendev.org/c/openstack/nova/+/795435
20:46:29 JayF elodilles: ...has that always been the behavior? That recheck doesn't clear a V+1?
20:46:57 sean-k-mooney its not recheck the clears it
20:47:01 sean-k-mooney its zuul
20:47:09 sean-k-mooney it only does it when the build set starts

Earlier   Later