| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-03-23 | |||
| 15:47:32 | sean-k-mooney | but not sure if we have a tacker for it | |
| 15:47:34 | dansmith | rebuild your venv? | |
| 15:47:38 | dansmith | tox -r | |
| 15:47:57 | outbrito | It was a brand new clone, but I'll try anyway | |
| 15:49:13 | dansmith | ack, usually random oslo things for me are due to a stale venv when reqs had changed, so that's the first thing I try | |
| 15:49:57 | sean-k-mooney | outbrito: is this actully causing test failures for you or just an warning/error message | |
| 15:50:07 | outbrito | 5 test failing | |
| 15:50:13 | outbrito | tests* | |
| 15:50:49 | sean-k-mooney | ok that actully the same as i saw on my mac book air tests yesterday also on a clean clone | |
| 15:51:11 | sean-k-mooney | let me see if i can repoduce on my normal system | |
| 15:53:41 | sean-k-mooney | i updated my ubuntu 20.04 server to 21.10 at the weekend but its running them now | |
| 15:54:17 | sean-k-mooney | so it wont be idential hum so i did get 1 failure with master | |
| 15:54:22 | sean-k-mooney | ill try in a clean clone | |
| 15:54:58 | sean-k-mooney | outbrito: what version of python did you run them under | |
| 15:55:09 | sean-k-mooney | 3.8 i assume? | |
| 15:55:14 | sean-k-mooney | just in case that matters | |
| 15:57:13 | opendevreview | Merged openstack/placement master: tests: Silence noisy tests https://review.opendev.org/c/openstack/placement/+/801101 | |
| 15:58:34 | outbrito | 3.9 | |
| 16:00:26 | outbrito | The only differences I see on requirements is Yappi vs. yappi and PyNaCl vs. pynacl | |
| 16:01:53 | gibi | hm with 20.04 + py38 + fresh master I could not reproduce the lock_path issue | |
| 16:03:54 | outbrito | The tests failing are these btw: | |
| 16:04:02 | outbrito | nova.tests.unit.compute.test_compute_mgr.ComputeManagerUnitTestCase.test_detach_volume_evacuate_mismatch | |
| 16:04:02 | outbrito | nova.tests.unit.compute.test_compute_mgr.ComputeManagerUnitTestCase.test_detach_volume_evacuate_legacy | |
| 16:04:02 | outbrito | nova.tests.unit.compute.test_compute_mgr.ComputeManagerUnitTestCase.test_detach_volume_evacuate | |
| 16:04:02 | outbrito | nova.tests.unit.virt.test_block_device.TestDriverBlockDeviceNewFlow.test_volume_attach_with_shared_targets | |
| 16:04:04 | outbrito | nova.tests.unit.virt.test_block_device.TestDriverBlockDevice.test_volume_attach_with_shared_targets | |
| 16:04:57 | outbrito | All related to volumes for that matter | |
| 16:04:58 | opendevreview | Merged openstack/placement master: tox: Remove psycopg2 warning filter https://review.opendev.org/c/openstack/placement/+/800911 | |
| 16:08:48 | sean-k-mooney | on a clean clone on 21.10 i did not have any failure under python 39 | |
| 16:10:37 | outbrito | py38 also broken for me | |
| 16:11:24 | outbrito | just to be sure, last commit is d2b2516 right? | |
| 16:11:39 | gibi | yepp | |
| 16:11:45 | gibi | I see that as head too | |
| 16:11:51 | gibi | but no luck reproducing the issue | |
| 16:12:07 | gibi | does it happen the same way for you each time you rerun these tests? | |
| 16:12:15 | gibi | or is it unstable? | |
| 16:13:47 | outbrito | consistent | |
| 16:15:30 | gibi | outbrito: do you also get a stack trace? If yes could you past it via https://paste.opendev.org/ (or similar) | |
| 16:17:04 | opendevreview | Merged openstack/placement master: setup: Replace dashes with underscores https://review.opendev.org/c/openstack/placement/+/801102 | |
| 16:17:34 | outbrito | https://paste.opendev.org/show/bZm6XstZ4MTPM7SPVuCR/ | |
| 16:21:38 | gibi | thanks | |
| 16:25:06 | gibi | outbrito: what is the os-brick version in your virtualenv? | |
| 16:40:24 | mfo | sean-k-mooney, hey o/ getting back to you exactly 1 month later :) sorry. i'd appreciate your comments/review when you have a chance (but it's not urgent, as you can see). | |
| 16:40:35 | mfo | we chatted [0] about an issue [1] that the libvirt driver in uefi mode ended up with the ovmf _secboot_ image, which doesn't boot on the pc machine type (required q35); so i proposed to skip the secboot image if on pc. [2,3] | |
| 16:40:45 | mfo | [0] https://pastebin.ubuntu.com/p/jVxcXvDW6h/ | |
| 16:40:45 | mfo | [1] https://bugs.launchpad.net/nova/+bug/1960758 | |
| 16:40:45 | mfo | [2] https://review.opendev.org/c/openstack/nova/+/828980 | |
| 16:40:46 | mfo | [3] https://review.opendev.org/c/openstack/nova/+/828979 | |
| 16:40:53 | mfo | essentially, i would get back to you with: | |
| 16:40:53 | mfo | 1) explain why the secboot image didn't "just work", 2) add unit tests, 3) ensure we only skip secboot image if there's another image to use. | |
| 16:40:53 | mfo | i've explained 1) with technical details/research in the bug's comments, and addressed 2) and 3) in the patches. | |
| 16:40:55 | mfo | thanks! | |
| 16:41:53 | sean-k-mooney | mfo: ack, in a downstream call currently so will review it after | |
| 16:42:19 | sean-k-mooney | mfo: in your opipion are the parces ready for review if so ill try and add them to my backlog for this week | |
| 16:42:46 | mfo | sean-k-mooney, yup, i think it's ready for review. | |
| 16:43:49 | gibi | outbrito: sorry I don't see what is happening on your side with those test and I will soon and my day. | |
| 16:59:33 | outbrito | no worries, i'll keep investigating. Thanks for looking into it | |
| 17:00:38 | outbrito | gibi, os-brick is at 5.2.0 btw | |
| 18:44:31 | melwitt | dansmith: I have a patch proposed to address a func test gate bug I see semi frequently and the patch is based on a NOTE by you in the CellDatabases func test fixture. if you might be interested in reviewing: https://review.opendev.org/c/openstack/nova/+/831084 | |
| 18:53:26 | dansmith | melwitt: yeah I remember that, looking | |
| 19:02:48 | opendevreview | Artom Lifshitz proposed openstack/nova master: Add whitebox-devstack-multinode job to periodic https://review.opendev.org/c/openstack/nova/+/833453 | |
| 19:04:49 | dansmith | melwitt: commented, see what you think | |
| 19:05:11 | melwitt | cool thanks. looking | |
| 19:09:22 | opendevreview | Artom Lifshitz proposed openstack/nova master: Add whitebox-devstack-multinode job to periodic https://review.opendev.org/c/openstack/nova/+/833453 | |
| 19:16:27 | opendevreview | Artom Lifshitz proposed openstack/nova master: Add whitebox-devstack-multinode job to periodic https://review.opendev.org/c/openstack/nova/+/833453 | |
| 19:37:30 | opendevreview | Artom Lifshitz proposed openstack/nova master: Add whitebox-devstack-multinode job to periodic https://review.opendev.org/c/openstack/nova/+/833453 | |
| 19:58:31 | artom | "deprecated_since=versionutils.deprecated.WALLABY" | |
| 19:58:43 | artom | ... in a placement installed in a ussuri tox functional venv | |
| 19:58:48 | artom | *sigh* | |
| 20:00:43 | sean-k-mooney | that is likely down to how we install placment for the placment fixture in the tox env | |
| 20:01:15 | sean-k-mooney | if that is happenign in a clean checkout with new venv | |
| 20:01:28 | sean-k-mooney | https://github.com/openstack/nova/blob/master/tox.ini#L119 | |
| 20:01:53 | sean-k-mooney | that should still be limited by upper constratis for the branch https://github.com/openstack/nova/blob/master/tox.ini#L29 | |
| 20:02:22 | sean-k-mooney | ah hehe | |
| 20:02:26 | sean-k-mooney | https://opendev.org/openstack/requirements/raw/branch/stable/ussuri/upper-constraints.txt | |
| 20:02:39 | sean-k-mooney | placement is not in upperconstrait for stable/ussuri | |
| 20:02:53 | sean-k-mooney | artom: so you will get whatever is on pypi | |
| 20:03:14 | sean-k-mooney | for the most part that should be ok as we are just using it for the fixture in this case | |
| 20:03:35 | artom | Unless it explicitly uses variables named after major releases :P | |
| 20:03:35 | sean-k-mooney | but that is why that happens its installing the latest pypi release of placement | |
| 20:03:56 | artom | Presumably this would fail in the upstream gate as well | |
| 20:04:15 | sean-k-mooney | proably we likely should cap it to the feature level | |
| 20:04:26 | sean-k-mooney | * minor release | |
| 20:04:47 | sean-k-mooney | actully placement is not release independ ot its the major verion | |
| 20:05:13 | sean-k-mooney | that woudl be pretty trival to do just update https://github.com/openstack/nova/blob/stable/ussuri/tox.ini#L91 | |
| 20:05:28 | sean-k-mooney | but we woudl need to do that with a stable only patch for each release | |
| 20:06:19 | sean-k-mooney | so ussuri is 3.0.0 so you need to set it to <4.0.0 | |
| 20:06:58 | sean-k-mooney | so """openstack-placement>=1.0.0,<4.0.0""" i think although im not sure if the , is needed | |
| 20:07:02 | sean-k-mooney | that shoudl fix it | |
| 20:07:34 | sean-k-mooney | or you coudl be explcit and say openstack-placement==3 | |
| 20:08:22 | sean-k-mooney | that will give you any 3.x.y but not 2.* or 4.* | |
| 20:08:47 | sean-k-mooney | artom: hopefully that makes sense | |
| 20:09:03 | sean-k-mooney | im goign to call it a day so chat to you tomorrow o/ | |
| 20:09:32 | artom | sean-k-mooney, yep, many thanks, very useful | |
| 20:10:00 | artom | Will also need to do it for victoria as well, since it doesn't understand WALLABY either | |
| 20:15:02 | clarkb | note I think ==3 means 3.0.0 not 3.x.y | |
| 20:16:14 | sean-k-mooney | im not sure about that but either way we can use min/max version or just pin to the majory with ==3 depending on the semantics of the syntax | |
| 20:16:30 | sean-k-mooney | i tought it woudl select the new 3.x.y | |
| 20:16:36 | sean-k-mooney | but im not sure | |
| 20:19:25 | clarkb | the ~= operator is for that or using * in the versions. missing identifiers seem to expand to .0 | |