| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2024-08-29 | |||
| 01:48:42 | melwitt | cc stephenfin ^ | |
| 01:49:05 | melwitt | (sphinx expert) | |
| 03:06:38 | melwitt | stephenfin: update: I found what the problem is, sphinx 8.0.0 made overwriting of user files an error by default and the sphinx-feature-classification extension is doing an overwrite. it needs to use force=True to maintain its behavior, I have proposed a patch based on local testing https://review.opendev.org/c/openstack/sphinx-feature-classification/+/927449 | |
| 03:07:20 | melwitt | that ^ will need a version requirement on 8.0.0 or some other way to only pass force=True if >= 8.0.0. so I'm not sure how to handle that | |
| 03:09:35 | melwitt | stephenfin: I suspect it will be best to roll back the sphinx version bump in upper-constraints for now | |
| 03:35:08 | melwitt | cc bauzas ^ | |
| 05:25:30 | opendevreview | Amit Uniyal proposed openstack/nova master: VMware: updates resource provider trait list https://review.opendev.org/c/openstack/nova/+/926522 | |
| 05:25:31 | opendevreview | Amit Uniyal proposed openstack/nova master: zvm: updates resource provider trait list https://review.opendev.org/c/openstack/nova/+/926523 | |
| 05:25:31 | opendevreview | Amit Uniyal proposed openstack/nova master: Ironic: updates resource provider trait list https://review.opendev.org/c/openstack/nova/+/926524 | |
| 06:56:34 | mikal | melwitt: when is feature freeze? I'm feeling like I'm unlikely to get this patch series landed before then? | |
| 07:04:17 | mikal | Yeah, I think its this week which means this thing isn't going to happen. | |
| 07:28:32 | opendevreview | Jens Harbott proposed openstack/nova master: Delete ./doc/source/_static/support-matrix.css file https://review.opendev.org/c/openstack/nova/+/927471 | |
| 07:30:02 | frickler | melwitt: ^^ maybe this is the better solution? sphinx tries to copy .tox/docs/lib/python3.10/site-packages/sphinx_feature_classification/support-matrix.css over it anyway. at least that seemed to fix the error for me locally | |
| 07:31:41 | frickler | but if that doesn't work for some reason I have missed so far, I'm also fine with reverting the sphinx bump until next cycle | |
| 08:14:19 | opendevreview | Balazs Gibizer proposed openstack/nova stable/2023.2: hardware: Correct log https://review.opendev.org/c/openstack/nova/+/927473 | |
| 08:18:50 | opendevreview | Balazs Gibizer proposed openstack/nova stable/2023.1: Functional test test_boot_reschedule_with_proper_pci_device_count https://review.opendev.org/c/openstack/nova/+/926558 | |
| 08:18:51 | opendevreview | Balazs Gibizer proposed openstack/nova stable/2023.1: Fix PCI passthrough cleanup on reschedule https://review.opendev.org/c/openstack/nova/+/926559 | |
| 08:20:40 | bauzas | frickler: saw the email from melwitt about the sphinx update that hits the docs CI ? | |
| 08:20:54 | bauzas | oh shit, just saw your points | |
| 08:23:41 | opendevreview | Kaloyan Kotlarski proposed openstack/nova master: libvirt: Fetch the available vCPUs from the respective cgroup https://review.opendev.org/c/openstack/nova/+/927474 | |
| 08:23:43 | bauzas | frickler: do you know if others projects are impacted ? | |
| 08:36:58 | bauzas | I'm just afraid we would find other bugs so close to FF | |
| 09:14:35 | frickler | I didn't see other failures yet, but also didn't explicitly look for them so far | |
| 09:42:09 | opendevreview | Stephen Finucane proposed openstack/python-novaclient master: tox: Add Python-specific functional envs https://review.opendev.org/c/openstack/python-novaclient/+/927250 | |
| 09:42:36 | opendevreview | Stephen Finucane proposed openstack/python-novaclient master: Fix Python 3.12 compatibility https://review.opendev.org/c/openstack/python-novaclient/+/927251 | |
| 11:08:50 | opendevreview | sean mooney proposed openstack/nova master: update nova-next to use ubuntu 24.04 https://review.opendev.org/c/openstack/nova/+/922492 | |
| 12:14:17 | opendevreview | sean mooney proposed openstack/nova master: [DNM] testing uc cap https://review.opendev.org/c/openstack/nova/+/927500 | |
| 12:18:47 | sean-k-mooney | melwitt: i think https://review.opendev.org/c/openstack/openstackdocstheme/+/927496/3/openstackdocstheme/page_context.py will fix the openstackdocstheme | |
| 12:19:05 | sean-k-mooney | and i think https://review.opendev.org/c/openstack/requirements/+/927499 is what allowed the newer sphinx build | |
| 12:19:18 | sean-k-mooney | well the bump to 4.0.0 | |
| 12:20:26 | frickler | sean-k-mooney: sphinx is explicitly pinned in u-c. but also please check https://review.opendev.org/c/openstack/nova/+/927471 | |
| 12:21:10 | sean-k-mooney | that not the error we are seeign in the gate as far as i can see | |
| 12:21:20 | sean-k-mooney | or its not related to it | |
| 12:22:30 | frickler | sean-k-mooney: then maybe we are talking about two different issues, can you point me to one of your failures? | |
| 12:22:56 | sean-k-mooney | there is a new sphinx removal warning for sphinx 9 | |
| 12:23:13 | frickler | yes, but afaict that is non-fatal | |
| 12:23:14 | sean-k-mooney | we treat warnings as errors | |
| 12:23:24 | sean-k-mooney | so in the gate it is | |
| 12:24:24 | frickler | in the docs build? are you telling me my patch passes in check but will fail in gate pipeline? that would be a pretty bad design IMO | |
| 12:25:02 | sean-k-mooney | no | |
| 12:25:19 | sean-k-mooney | i thnk you test willl fail now that the new verison of sphinx is allwoed | |
| 12:25:29 | frickler | no, it ran after that | |
| 12:27:26 | sean-k-mooney | maybe we are exluding | |
| 12:27:28 | sean-k-mooney | /home/zuul/src/opendev.org/openstack/nova/.tox/docs/lib/python3.10/site-packages/openstackdocstheme/page_context.py:70: RemovedInSphinx90Warning: Sphinx 9 will drop support for representing paths as strings. Use "pathlib.Path" or "os.fspath" instead. | |
| 12:27:30 | sean-k-mooney | src_file = full_src_file[len(str(app.builder.env.srcdir)):].lstrip('/') | |
| 12:27:42 | auniyal | but shouldn't we just delete too, if its problametic, and deleting fixes the issue also, like what this patch is doing https://review.opendev.org/c/openstack/nova/+/927471 | |
| 12:27:46 | sean-k-mooney | form what is fatal | |
| 12:28:04 | auniyal | or it can cause any other error ? | |
| 12:28:20 | frickler | check https://ca0e8b358c93d8e83ff8-d2c9747c19735a1cd74d2b5484a5c4b3.ssl.cf5.rackcdn.com/927471/1/check/openstack-tox-docs/8034b4b/job-output.txt , using sphinx-8.0.2, lots of warning, no failure | |
| 12:30:13 | frickler | the failure is this: WARNING: Aborted attempted copy from /home/zuul/src/opendev.org/openstack/nova/.tox/docs/lib/python3.10/site-packages/sphinx_feature_classification/support-matrix.css to /home/zuul/src/opendev.org/openstack/nova/doc/build/html/_static/support-matrix.css (the destination path has existing data). [misc.copy_overwrite] | |
| 12:30:18 | frickler | build finished with problems, 1 warning. | |
| 12:31:18 | sean-k-mooney | so your current change is passign so im just going to merge it but i think we regruest how our docs job works at some point | |
| 12:32:28 | bhola | Hi. Looking for some help. I have installed openstack following the installation guide on openstack site. Everything seems to be working good but I cannot launch an instance. It throws the error "Block Device Mapping is Invalid: failed to get image" if I use image. If I use pre-defined volume based on image it throws error "Block Device Mapping is Invalid: failed to get volume" | |
| 12:32:43 | bhola | Any idea how to fix this issue? | |
| 12:32:59 | frickler | sean-k-mooney: regruest = regressed? (I can parse most of your typos but not sure on this one) | |
| 12:34:29 | sean-k-mooney | yes regressed | |
| 12:35:02 | sean-k-mooney | we run "sphinx-build -W --keep-going -b html -j auto doc/source doc/build/html" | |
| 12:35:13 | sean-k-mooney | -W is fail-on-warnings | |
| 12:37:48 | auniyal | bhola can you check nova and cinder logs for detailed traceback | |
| 12:40:29 | sean-k-mooney | bhola: as a general rule we do not provide support on this channel. it for development but if you can point ot a spefici error messge then we may be able to point you in the right direction | |
| 12:40:56 | sean-k-mooney | bhola: also without know which version of openstack you installed we cant really say what might be the issue | |
| 12:41:25 | sean-k-mooney | and which distro | |
| 12:46:58 | sean-k-mooney | frickler: there are some followups we likely shoudl do for https://review.opendev.org/c/openstack/nova/+/927471 | |
| 12:47:15 | sean-k-mooney | it should be in .gitignore for one | |
| 12:47:22 | sean-k-mooney | but lest unblock things and do that after | |
| 12:47:54 | bauzas | I was awaiting people to arrive before deciding anything | |
| 12:48:12 | stephenfin | sean-k-mooney: frickler: yeah, https://review.opendev.org/c/openstack/nova/+/927471 looks like the proper approach. I'm +2 on that also. | |
| 12:48:56 | bauzas | I tend to be conservative as we're close to the FF and probably ask ourselves to temporarly avoid the sphinx bump | |
| 12:49:34 | bauzas | Dalmatian is a non-SlURP release, I feel less obliged to have all our libs being upgraded, if we don't really need the upgrade itself | |
| 12:49:40 | bauzas | and other issues may arise | |
| 12:49:50 | sean-k-mooney | i dont think that a good approch to take | |
| 12:50:08 | sean-k-mooney | we shoudl not devalue non-slurp release | |
| 12:50:16 | sean-k-mooney | they are just as imporant as SLURPS | |
| 12:50:32 | bauzas | is there any reason for nova to use sphinx 8.0 ? | |
| 12:50:51 | frickler | IMHO we have the freeze coming up soon in order to still have enough time before the release to fix issues, not in order to avoid bumping things even before that | |
| 12:50:52 | sean-k-mooney | i belive we have been for some time | |
| 12:50:56 | bauzas | sean-k-mooney: I'm not devaluating the non-SLURP releases, I'm just evaluating the risks | |
| 12:51:07 | frickler | why not use the latest and greatest version as we do for everything? | |
| 12:51:16 | sean-k-mooney | bauzas: this is blockign seveeral feature that are close to merging in this release | |
| 12:51:33 | sean-k-mooney | so the trade of to fixing this now is FFE into the early rc period | |
| 12:52:14 | sean-k-mooney | there are several small feature and bugfix pathces that i dont think would be reasonabel to just defer to the next release | |
| 12:52:33 | bhola | auniyal, Just a sec. Let me see in the log. | |
| 12:52:45 | bauzas | sean-k-mooney: for the sake of evaluating the need, which features are depending on the very recent sphinx bump ? (I may have missed them) | |
| 12:52:55 | sean-k-mooney | none | |
| 12:53:03 | sean-k-mooney | but this is currently break our ablity to merge patches | |
| 12:53:20 | bauzas | hence my proposal to skip the bump, for sure | |
| 12:53:24 | stephenfin | bauzas: sphinx isn't a runtime dependency, the docs are not a "feature" (so talk of FFE doesn't make sense), and Nova is doing something wrong and Sphinx is only pointing it out now | |
| 12:54:02 | sean-k-mooney | stephenfin: i ment for all the feature patches that failed to meger last night and that we cant merge today until this is fixed | |
| 12:54:04 | bhola | sean-k-mooney, The versioning system of openstack always confuses me. I think there are different versions of every service. Are you referring to openstack client toll version? | |
| 12:54:20 | bauzas | stephenfin: I'm not against docs, for sure, I'm just evaluating the risk of discovering a late bug that would require some RC2 | |
| 12:54:31 | stephenfin | I don't know if folks like frickler that work on deps would agree, but IMO we should need a *very* good reason to hold back a dep bump | |
| 12:54:41 | sean-k-mooney | bhola: no we have a unified release when you install you normally install that unified release. | |
| 12:54:58 | sean-k-mooney | so while ech componet may have a diffent release number you isntall one release of openstack | |
| 12:55:03 | bauzas | a regression sounds a good reason to me :) | |
| 12:55:12 | stephenfin | bauzas: Right. But again, sphinx is not a runtime dependency. So talk of an RC2 doesn't make any sense. | |
| 12:55:58 | bauzas | this is true, but I wouldn't want our dalmatian branch to be hit by some unknown yet regression | |
| 12:56:27 | bauzas | after RC1, you know that we would have this 2 weeks period when no backports can be done on stables | |
| 12:57:31 | bauzas | anyway, seems to me the ship has sailed. https://review.opendev.org/c/openstack/nova/+/927471 is +Wd | |
| 12:57:45 | frickler | so IMO https://review.opendev.org/c/openstack/nova/+/927471 is a very simple, no-risk fix, and much better than reverting the u-c bump for everyone | |
| 12:58:00 | stephenfin | +1 | |