| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-09-06 | |||
| 16:07:24 | bauzas | gibi: would you propose a patch to nova's requirements.txt or do we automatically pull the latest ? | |
| 16:07:38 | sean-k-mooney | https://review.opendev.org/c/openstack/requirements/+/856044 | |
| 16:07:39 | gibi | we are not constrained in noca | |
| 16:07:43 | gibi | nova | |
| 16:07:50 | sean-k-mooney | i think that is the requiremtns bump | |
| 16:08:05 | bauzas | gibi: ok, so we'll benefit from it | |
| 16:08:17 | gibi | yep, automatically | |
| 16:08:20 | sean-k-mooney | we shoudl once that merges yep | |
| 16:08:27 | bauzas | https://github.com/openstack/nova/blob/master/requirements.txt#L34 | |
| 16:08:33 | bauzas | yup, we're not constrained | |
| 16:08:46 | bauzas | ok, let's wait for the req patch to be merged then | |
| 16:09:11 | bauzas | for stable/yoga, seems we don't cap too | |
| 16:09:12 | bauzas | https://github.com/openstack/nova/blob/stable/yoga/requirements.txt#L30 | |
| 16:09:26 | sean-k-mooney | correct | |
| 16:09:31 | gibi | coo | |
| 16:09:31 | gibi | l | |
| 16:09:41 | sean-k-mooney | we can either pull in the latest oslo release there or cap fasterners | |
| 16:09:42 | bauzas | ok, we're on the good way | |
| 16:09:55 | bauzas | moving on to the next bug then | |
| 16:10:09 | bauzas | #link https://bugs.launchpad.net/nova/+bug/1988316 Will be set to High now that the gate is skipping the test | |
| 16:10:17 | bauzas | people agree with the proposal ? | |
| 16:10:23 | gibi | sean-k-mooney: I would bump this https://github.com/openstack/requirements/blob/stable/yoga/upper-constraints.txt#L26 | |
| 16:10:24 | bauzas | => High | |
| 16:10:45 | bauzas | gibi: oh you're right about yoga's u-c | |
| 16:11:32 | gibi | on the unshelve stuff I think we need a nova fix to return a better error message, but I'm OK to drop it from Critical to High | |
| 16:11:40 | bauzas | yeah | |
| 16:11:45 | gibi | it is not blocking the gate at the moment | |
| 16:11:47 | bauzas | not saying there will be no bug to fix | |
| 16:11:51 | sean-k-mooney | ya its a latent bug | |
| 16:12:06 | bauzas | but this isn't longer critical as the gate is now working back | |
| 16:12:07 | dansmith | unshelve thing meaning the to-host cell one? | |
| 16:12:13 | sean-k-mooney | yes | |
| 16:12:15 | bauzas | yup | |
| 16:12:29 | bauzas | saying we have to restrict to the same cell | |
| 16:12:35 | dansmith | yeah, that's not new just recently exposed so doesn't seem like we should be too critical in the bug status there | |
| 16:12:55 | sean-k-mooney | im saying its latent as its the same error we would have got with unshelve to AZ we just dont have test coverage for that in tempest that was also corss cell | |
| 16:13:23 | dansmith | right, and I'm saying being latent it's not a regression, so really shouldn't be >high :) | |
| 16:13:31 | sean-k-mooney | +1 | |
| 16:13:45 | bauzas | honestly, I'm in favor of saying "sorry but we don't support *yet* cross-cell unshelve to host" that's it | |
| 16:13:59 | bauzas | so the fix could be just docs | |
| 16:14:06 | bauzas | and a better exception handling | |
| 16:14:09 | sean-k-mooney | yep we can do both | |
| 16:14:18 | bauzas | anyway, set to High, there is | |
| 16:15:13 | dansmith | well, the test needs fixing too | |
| 16:15:24 | dansmith | it's just being skipped there right now right? | |
| 16:15:29 | bauzas | yup | |
| 16:15:34 | sean-k-mooney | the test is not broken | |
| 16:15:40 | sean-k-mooney | it just should not be run in a cross cell env | |
| 16:15:41 | bauzas | yeah, the test shouldn't be assuming we could | |
| 16:15:57 | sean-k-mooney | no this is a job config issue | |
| 16:15:59 | bauzas | if we would want a test, this would have to be negative | |
| 16:16:05 | sean-k-mooney | you cant tell if its cross cell form the api | |
| 16:16:09 | bauzas | like "I know this can't work" | |
| 16:16:12 | sean-k-mooney | so tempest cannot detech that | |
| 16:16:20 | sean-k-mooney | so this has to be done via job config | |
| 16:16:24 | bauzas | ahah true | |
| 16:16:47 | bauzas | anyway, bug report is still open, feel free to comment it out for resolution | |
| 16:16:50 | bauzas | moving on | |
| 16:16:57 | opendevreview | Rajat Dhasmana proposed openstack/nova-specs master: Clarify client changes in rebuild spec https://review.opendev.org/c/openstack/nova-specs/+/856164 | |
| 16:17:00 | bauzas | #link https://bugs.launchpad.net/nova/+bug/1988482 Proposed patch on the fly https://review.opendev.org/c/openstack/nova/+/855658 | |
| 16:17:13 | bauzas | as said, reviews are welcome ^ | |
| 16:17:23 | dansmith | sean-k-mooney: shoudn't the test be using an az? | |
| 16:17:25 | bauzas | I exceptionally relaxed my review needs | |
| 16:17:32 | dansmith | anyway, we can discuss elsewhere | |
| 16:17:42 | bauzas | so I gave +2 for a patch without testing | |
| 16:17:43 | sean-k-mooney | dansmith: i dont think so be we can follow up after ya | |
| 16:18:21 | bauzas | tl;dr: the problem is with PrettyTable having a changed behavior | |
| 16:18:43 | sean-k-mooney | yes so they have now fixed this by reverting the behaivor | |
| 16:18:49 | bauzas | gibi: sean-k-mooney: wants to address this issue now ? | |
| 16:18:56 | sean-k-mooney | i tested with the previous release broken release and new release with the revert | |
| 16:19:19 | sean-k-mooney | i woudl prefer to merge and backprot https://review.opendev.org/c/openstack/nova/+/855658 | |
| 16:19:22 | bauzas | so https://review.opendev.org/c/openstack/nova/+/855658 wouldn't be needed ? | |
| 16:19:25 | bauzas | ack | |
| 16:19:27 | sean-k-mooney | to yoga so we never need to thnk about htis again | |
| 16:19:51 | sean-k-mooney | its not needed unless the decided to reinstate the feature or change the default | |
| 16:20:01 | sean-k-mooney | or we can hardcode the alignmend and not worry | |
| 16:20:18 | gibi | I agree we sean-k-mooney | |
| 16:20:31 | bauzas | gibi: didn't know this verb | |
| 16:20:43 | bauzas | I sean-k-mooney, you sean-k-mooney | |
| 16:20:49 | bauzas | :) | |
| 16:20:53 | sean-k-mooney | :) | |
| 16:20:57 | gibi | but as an extra: in general we don't have py39 lines in the global requirements so this can hit us with different packages | |
| 16:21:00 | gibi | bauzas: :) | |
| 16:21:17 | gibi | I think there was a discussion in relmgmt about adding those lines | |
| 16:21:40 | gibi | in yoga we should have been upper constrained | |
| 16:21:44 | sean-k-mooney | it could hit distos if the distro had 3.4.0 | |
| 16:21:45 | elodilles | (rather on requirements, but yes) | |
| 16:21:52 | bauzas | yeah | |
| 16:22:10 | gibi | elodilles: then there :) same folks :) | |
| 16:22:12 | sean-k-mooney | anyway i think we can just review this normally | |
| 16:22:14 | sean-k-mooney | not critical | |
| 16:22:18 | gibi | agree | |
| 16:22:23 | sean-k-mooney | it was just annoying as it was blocking some backports | |
| 16:22:25 | sean-k-mooney | its not now | |
| 16:23:07 | bauzas | ok, so chaning the prio to High ? | |
| 16:23:16 | bauzas | because of the requirements patch ? | |
| 16:24:04 | gibi | I agree that this is not critical now | |
| 16:24:20 | bauzas | ok, punting the prio then | |
| 16:24:35 | bauzas | sean-k-mooney: please just update the bug report with your thoughts please | |
| 16:24:41 | sean-k-mooney | ack | |