| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-03-21 | |||
| 18:49:29 | sean-k-mooney | the zuul change that is https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/833286/2/zuul.d/project-templates.yaml | |
| 18:54:34 | sean-k-mooney | stephenfin: ok done.# | |
| 18:54:43 | stephenfin | spot on, thanks! | |
| 18:57:42 | sean-k-mooney | did you also do os-resource classes | |
| 18:59:01 | sean-k-mooney | stephenfin: dont see it here https://review.opendev.org/q/topic:add-python3-unversioned-jobs but i think its independent too https://github.com/openstack/releases/blob/master/deliverables/_independent/os-resource-classes.yaml | |
| 18:59:21 | stephenfin | Oh, my hacky script probably just missed that | |
| 19:00:16 | sean-k-mooney | i need to add a PTG topic about this by the way | |
| 19:00:32 | sean-k-mooney | we might want to merge those back into placment | |
| 19:00:53 | sean-k-mooney | we have a cirular dep issue with testing and releases | |
| 19:01:29 | sean-k-mooney | we can either relax the test that assert the amount of traits/resouce classes or we can put them all in the same repo | |
| 19:02:06 | sean-k-mooney | stephenfin: im going to go eat something. if you push a patch add me and ill look at it tomorrow | |
| 19:02:54 | opendevreview | Stephen Finucane proposed openstack/os-resource-classes master: Update CI to use unversioned jobs template https://review.opendev.org/c/openstack/os-resource-classes/+/834537 | |
| 19:03:01 | stephenfin | sean-k-mooney: Cool. It's there btw ^ | |
| 19:04:08 | opendevreview | Merged openstack/os-traits master: Update CI to use unversioned jobs template https://review.opendev.org/c/openstack/os-traits/+/833826 | |
| 19:13:18 | gmann | stephenfin I missed it for os-traits but please update setup.cfg too to reflect the tested versions https://review.opendev.org/c/openstack/os-resource-classes/+/834537 | |
| 19:14:30 | stephenfin | gmann: I'd rather not mix those two things though I'll admit they're related. Do we want to bump 'python-requires' also? | |
| 19:14:38 | gmann | also, like i mentioned in ML reply, I have added to remove the release specific templates itself which will fix all these issue and less work during releases https://etherpad.opendev.org/p/tc-zed-ptg#L108 | |
| 19:15:37 | gmann | stephenfin: yes we can bump python-requires also. They are related as if we do not test py3.6/7 then we should update it in classifier too. | |
| 19:16:01 | gmann | in zed release template we will do the same for branched repo also | |
| 19:16:24 | stephenfin | Oh, that's a good idea | |
| 19:16:55 | gmann | in Zed: python-requires >= py3.8 and remove py3.6 and py3.7 from list | |
| 19:17:26 | stephenfin | _Personally_, I'd question the value the classifiers bring. They aren't used by any tooling that I'm aware of, unlike python-requires. I wonder if we should get rid of them? | |
| 19:19:14 | gmann | I think someone might be checking it from pypi doc but yes I do not have anyone asked about it explicitly | |
| 19:25:45 | dansmith | gmann: yeah, maybe some centralized docs would be good | |
| 19:25:57 | dansmith | gmann: I definitely think we need PTLs to get the word out though | |
| 19:26:56 | gmann | dansmith: +1 on PTL one. I like to neutron way to track it in their weekly meeting. that is something we can suggest PTL to do. | |
| 19:27:24 | gmann | I have added it in TC+PTL PTG etherpad | |
| 19:30:58 | dansmith | yeah | |
| #openstack-nova - 2022-03-22 | |||
| 10:17:57 | opendevreview | Stephen Finucane proposed openstack/os-resource-classes master: setup: Update Python testing classifiers https://review.opendev.org/c/openstack/os-resource-classes/+/834643 | |
| 10:17:57 | opendevreview | Stephen Finucane proposed openstack/os-resource-classes master: setup: Replace dashes with underscores, add links https://review.opendev.org/c/openstack/os-resource-classes/+/834644 | |
| 10:51:30 | zigo | Is there a way to evacuate a host that has 3 VMs that have affinity? Can I somehow tell nova "migrate them together" ? | |
| 12:41:50 | opendevreview | anguoming proposed openstack/nova master: fix the bug of the log line has no request_id info at source host when live migration https://review.opendev.org/c/openstack/nova/+/834677 | |
| 12:47:34 | opendevreview | anguoming proposed openstack/nova master: fix the bug of the log line has no request_id info at source host when live migration https://review.opendev.org/c/openstack/nova/+/834677 | |
| 12:50:53 | stephenfin | sean-k-mooney: This isn't hugely important, but could you look at https://review.opendev.org/c/openstack/nova/+/723572/ and https://review.opendev.org/c/openstack/nova/+/723573/ today? | |
| 12:54:49 | opendevreview | anguoming proposed openstack/nova master: fix the bug of the log line has no request_id info at source host when live migration https://review.opendev.org/c/openstack/nova/+/834677 | |
| 12:58:54 | opendevreview | Stephen Finucane proposed openstack/nova master: objects: Don't use generic 'Field' container https://review.opendev.org/c/openstack/nova/+/738239 | |
| 12:58:55 | opendevreview | Stephen Finucane proposed openstack/nova master: objects: Remove unnecessary type aliases, exceptions https://review.opendev.org/c/openstack/nova/+/738240 | |
| 12:58:55 | opendevreview | Stephen Finucane proposed openstack/nova master: objects: Use imports instead of type aliases https://review.opendev.org/c/openstack/nova/+/738018 | |
| 12:58:56 | opendevreview | Stephen Finucane proposed openstack/nova master: objects: Remove wrappers around ovo mixins https://review.opendev.org/c/openstack/nova/+/738019 | |
| 12:58:56 | opendevreview | Stephen Finucane proposed openstack/nova master: WIP: add ovo-mypy-plugin to type hinting o.vos https://review.opendev.org/c/openstack/nova/+/758851 | |
| 13:00:01 | sean-k-mooney | stephenfin: sure ill take a look at them now while i have context on this they look reasonably short and i see gmann has already reviewed them | |
| 13:00:29 | sean-k-mooney | getting rid fo the dict compat layer has been long overdue | |
| 13:00:47 | sean-k-mooney | i woudl be nice not to have to review for new usease of them as a dict | |
| 13:02:21 | opendevreview | Stephen Finucane proposed openstack/nova master: doc: Remove useless contributor/api-2 doc https://review.opendev.org/c/openstack/nova/+/828599 | |
| 13:54:03 | EugenMayer | When deploying via terraform it and changing an flavor (thus replacing it) it seems like the old flavour was removed but not yet 'removed from the instance it has been used' and then it all failed. Now i'am stuck with Unable to retrieve instance size information. Details Flavor 384bc436-a0cb-4e4a-80d1-26dd03743061 could not be found. (HTTP 404) | |
| 13:54:03 | EugenMayer | (Request-ID: req-7c68445d-a8b5-4ef6-a11d-6f037402d92a) - so basically one of my instances references a flavor that no longer exists. Is there a way to somehow fix this? | |
| 14:48:12 | artom | Anyone able to run functional tests on ussuri? | |
| 14:48:28 | artom | Trying to figure out if it's something local to me, or more widespread | |
| 14:48:48 | artom | Seems to be hanging/timing out on: | |
| 14:48:49 | artom | functional installdeps: -chttps://releases.openstack.org/constraints/upper/ussuri, -r/home/artom/src/nova/requirements.txt, -r/home/artom/src/nova/test-requirements.txt, openstack-placement>=1.0.0 | |
| 14:50:14 | sean-k-mooney | i can try it one sec | |
| 14:50:17 | artom | Seems to be doing... something? | |
| 14:50:19 | artom | Looping on https://paste.opendev.org/show/b45jbgPA429f5iKFJSEq/ | |
| 14:52:44 | sean-k-mooney | looks like we are missing a fixture | |
| 14:52:48 | sean-k-mooney | from that trace | |
| 14:52:58 | sean-k-mooney | we shoudl not eb doing ioctl;calls in general | |
| 14:53:47 | sean-k-mooney | like that implies we are doing file io or network configuration | |
| 14:56:40 | sean-k-mooney | its running fine for me | |
| 14:56:46 | sean-k-mooney | were you having a failing test? | |
| 14:56:50 | sean-k-mooney | or just would not install | |
| 14:57:48 | sean-k-mooney | i did locally change the psycopg2 to psycopg2-binary in my test-requirements.txt but that is just because i dont have or want postgress installed on my laptop | |
| 14:58:09 | sean-k-mooney | so i dont have the headers to build psycopg2 form source | |
| 15:01:04 | artom | So there is a backport in progress | |
| 15:01:12 | artom | *repo | |
| 15:01:30 | artom | But... it's not running any tests (yet), it's on installdeps... | |
| 15:02:26 | sean-k-mooney | got the gerrit linke i can try that explictly if you want | |
| 15:03:54 | artom | sean-k-mooney, only local for now | |
| 15:04:11 | artom | Backporting https://review.opendev.org/c/openstack/nova/+/796907/2/nova/tests/functional/libvirt/test_pci_sriov_servers.py#73 to ussuri | |
| 15:08:07 | sean-k-mooney | i had one failure | |
| 15:08:18 | sean-k-mooney | FileNotFoundError: [Errno 2] No such file or directory: 'openssl' | |
| 15:08:35 | sean-k-mooney | which is likely just down to the fact im runing this on nixos | |
| 15:09:57 | artom | Seems to be the same problem with a pristine ussuri... | |
| 15:10:10 | artom | I should try on Ubuntu I guess? | |
| 15:10:22 | artom | Although func tests should be platform-independant | |
| 15:15:04 | sean-k-mooney | running them on macos last night not as much as you woudl think | |
| 15:15:16 | sean-k-mooney | we have a bunch that fail because they detach its not linux | |
| 15:15:55 | sean-k-mooney | maybe pass -r | |
| 15:16:01 | sean-k-mooney | or delete the .tox dir | |
| 15:16:17 | sean-k-mooney | incase you have some leftover issue form a previous run | |
| 15:17:02 | artom | Yep, tried with -r, same | |
| 15:17:32 | sean-k-mooney | odd what distro are you currently using | |
| 15:17:44 | sean-k-mooney | i can try on ubunu if you like i also have a centos 9 vm | |
| 15:18:18 | artom | F35 | |
| 15:19:21 | bauzas | reminder : nova meeting in 41 mins here at #openstack-nova | |
| 15:19:53 | bauzas | fwiw, DST is not impacting our meeting, as we use UTC | |
| 15:20:10 | clarkb | artom: sean-k-mooney: pip installs taking forever likely indicates a dependency resolver problem | |
| 15:20:35 | clarkb | we've seen that happen when the solver can't find a valid answer. However constraints tends to fix that and you supply constraints so maybe not that | |
| 15:20:39 | sean-k-mooney | clarkb: i dont think it was the resolver | |
| 15:20:48 | sean-k-mooney | clarkb: i think artom is gettign stack traces | |
| 15:21:01 | artom | sean-k-mooney, no, just spinning in the void | |
| 15:21:07 | sean-k-mooney | oh have you added -v | |
| 15:21:09 | artom | The paste was a `strace -p` output | |
| 15:21:12 | sean-k-mooney | so you can see what actully happening | |
| 15:23:28 | artom | sean-k-mooney, *facepalm* lemme try that | |
| 15:35:47 | sean-k-mooney | artom: f35 has a much newer gcc libffi and kernel then ussuri was developed with by the way so that strace was refering to ffi presumable as part of compiling some of the c python modules so there might be issues with tyrign to install ussuri on f35 to run the func tests | |
| 15:42:24 | zigo | I just noticed that if a host is over its CPU ratio (because it has been reduced), then live-migrations are silently failing (only the scheduler gives a clue). Is this known? Is this considered a bug? Should I file the bug? | |
| 15:43:02 | zigo | The workaround is obviously to temporary up the CPU overcommit ratio temporarily, but that's still kind of annoying to do. | |
| 15:43:19 | sean-k-mooney | zigo: yes its a know issue | |
| 15:43:38 | sean-k-mooney | it has to do with how placement currently validates allocation candiates | |
| 15:43:50 | sean-k-mooney | if its the issue i think it is | |