gmann | it is working fine for stable branches but on master I am getting strange error https://zuul.opendev.org/t/openstack/build/9e1bf669244e46fe8380d8741234051a | 03:33 |
---|---|---|
gmann | I have not changed any template/job running on master | 03:33 |
*** akahat|rover is now known as akahat | 05:12 | |
gmann | seems master job error is not related to this as I can see the same error in other patches also https://zuul.openstack.org/build/269d2a0d50cb4c55a3fadfc6a58969eb | 05:13 |
gmann | not sure of that is known issue? | 05:13 |
gmann | it seems consistent issue. I think we are hitting I think we are hitting another issue from tox4 https://github.com/tox-dev/tox/pull/2824/ | 06:12 |
gmann | 06:12 | |
*** blarnath is now known as d34dh0r53 | 07:01 | |
*** rlandy|out is now known as rlandy | 11:12 | |
opendevreview | Stephen Finucane proposed openstack/pbr master: Update tox.ini to work with tox 4 https://review.opendev.org/c/openstack/pbr/+/869190 | 11:20 |
opendevreview | Stephen Finucane proposed openstack/pbr master: Remove numpy dependencies https://review.opendev.org/c/openstack/pbr/+/869402 | 11:20 |
*** dviroel|afk is now known as dviroel | 11:24 | |
opendevreview | Stephen Finucane proposed openstack/pbr master: Tie recursion calls to Dist object, not module https://review.opendev.org/c/openstack/pbr/+/869082 | 11:54 |
fungi | gmann: known issue, yes. we tracked it down in #openstack-oslo yesterday, seems to be related to using --skip-missing-interpreters=false with an envlist in tox.ini that includes at least one python interpreter version not present on the test node. pretty sure the main bug report tracking that is at https://github.com/tox-dev/tox/issues/2811 | 13:13 |
fungi | a workaround is to get rid of any specific interpreter versions in envlist and replace them with just "py3" | 13:14 |
opendevreview | Stephen Finucane proposed openstack/pbr master: Update tox.ini to work with tox 4 https://review.opendev.org/c/openstack/pbr/+/869190 | 13:22 |
opendevreview | Stephen Finucane proposed openstack/pbr master: Tie recursion calls to Dist object, not module https://review.opendev.org/c/openstack/pbr/+/869082 | 13:22 |
sean-k-mooney | fungi: so this is impacting multiple projects | 13:33 |
sean-k-mooney | for exampel nova but efoley is also lookign a telemetry projects | 13:33 |
sean-k-mooney | https://zuul.openstack.org/build/905f144095b945e2842991131be01228 | 13:33 |
sean-k-mooney | fungi: we likely need to remove --skip-missing-interpreters=false rahter then move the py3 | 13:35 |
sean-k-mooney | moveing to py3 in the env list is fine too | 13:35 |
sean-k-mooney | but that is not really goign to fix it if you ware trying to use a non default python verion in the vm | 13:36 |
sean-k-mooney | for exampel testign python 3.11 or 3.9 on ubuntu 22.04 and 20.04 respectivly | 13:36 |
efoley | fungi, sean-k-mooney, there are random interpreters being used each time the job is run. I saw py38, py39 and py311 being tried for openstack-tox-py310 | 13:36 |
efoley | FYI: https://github.com/tox-dev/tox/issues/2811 -- skip-missing-interpreters false doesn't work on tox 4. Perhaps we should pin tox<4 on master as well | 13:38 |
frickler | cinder already requires tox>=4, so that won't work. also not a longterm solution | 13:41 |
efoley | Not long-term, no, but unblocks things for now. | 13:42 |
efoley | Anyway, it seems that the issues is solved in tox 4.2.4 | 13:42 |
sean-k-mooney | https://github.com/tox-dev/tox/issues/2754 | 13:43 |
sean-k-mooney | looks like stephin fixed part of it | 13:43 |
sean-k-mooney | we proably need to bump the min tox version to 4.2.4 in requirements.txt | 13:48 |
sean-k-mooney | assuming that when that makes its way into the gate teh issue is resolved on master | 13:48 |
fungi | sean-k-mooney: well, --skip-missing-interpreters=false is there to keep us from successfully running zero tests when a unit test job targets the wrong platform | 14:08 |
fungi | efoley: see scrollback from before you joined, that's the exact github issue we were discussing (but welcome to the party!) | 14:08 |
sean-k-mooney | fungi: yes well more to make sure we dont run with the wrong interperter and have the job fail if the depency is not met | 14:10 |
sean-k-mooney | either via 0 tests or the wrong inteperter being usedif a project incorrectly set that to true in there tox.ini | 14:11 |
fungi | tox -epy37 on an ubuntu-jammy node will successfully run 0 tests if skip_missing_interpreters is true in tox.ini, it won't pick a different interpreter version | 14:13 |
fungi | you may be thinking of basepython + ignore_basepython_conflict | 14:13 |
fungi | they were introduced in tox around the same time, and people tend to conflate them | 14:14 |
*** dasm|off is now known as dasm | 14:14 | |
fungi | https://paste.opendev.org/show/bi1kh9asX31IXpKAIYPx/ | 14:25 |
fungi | that's with skip_missing_interpreters=true in [tox] | 14:25 |
fungi | if we could be sure projects had gotten rid of skip_missing_interpreters (i and others have repeatedly warned it's dangerous for that reason), then the --skip-missing-interpreters=false in the job wouldn't be so necessary | 14:26 |
frickler | well we can be sure that with current tox it is breaking everything. so the decision to drop it seems easy for me | 14:32 |
opendevreview | Stephen Finucane proposed openstack/pbr master: Remove numpy dependencies https://review.opendev.org/c/openstack/pbr/+/869402 | 14:36 |
opendevreview | Merged openstack/openstack-zuul-jobs master: Pin tox<4 for stable branches (<=stable/zed) testing https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/867849 | 15:00 |
fungi | gmann: frickler: sean-k-mooney: efoley: stephenfin's https://github.com/tox-dev/tox/pull/2828 just merged, which addresses most of that i think | 16:05 |
clarkb | you just have to worry about what other new brokeness the new release brings :) | 16:38 |
fungi | tox roulette | 16:38 |
*** dviroel is now known as dviroel|lunch | 16:39 | |
opendevreview | Clark Boylan proposed openstack/pbr master: Run PBR integration on Ubuntu Focal too https://review.opendev.org/c/openstack/pbr/+/869490 | 17:36 |
clarkb | fungi: stephenfin ^ something like that maye | 17:36 |
stephenfin | wfm | 17:37 |
*** dviroel|lunch is now known as dviroel | 17:50 | |
gmann | ack. thanks | 18:01 |
clarkb | stephenfin: fungi: if you grep for TestLTSSupport.test_lts_venv_default_versions in https://zuul.opendev.org/t/openstack/build/caf26d11d9cf4d4682d7813a99fa168d/log/job-output.txt and https://zuul.opendev.org/t/openstack/build/dc9a7dc4f0bb47509c8643a8564be3a3/log/job-output.txt that seems to show my change working. I'll rereview the rest of the stack shortly. | 19:02 |
clarkb | it also shows that we didn't break old pbr with the fix | 19:02 |
fungi | yep, lgtm | 19:02 |
fungi | i think we're good to merge the entire series if you're good with it | 19:02 |
clarkb | yup I just +A'd the stack | 19:04 |
clarkb | though I think there may be one that forks off I need to go find too | 19:05 |
clarkb | fungi: https://review.opendev.org/c/openstack/pbr/+/869402/3 that one I already reviewd it if you have time to review it | 19:05 |
fungi | i'll take a look | 19:09 |
fungi | approved it | 19:09 |
opendevreview | Merged openstack/pbr master: Update tox.ini to work with tox 4 https://review.opendev.org/c/openstack/pbr/+/869190 | 20:41 |
opendevreview | Merged openstack/pbr master: Remove numpy dependencies https://review.opendev.org/c/openstack/pbr/+/869402 | 20:42 |
*** dviroel is now known as dviroel|pto | 21:12 | |
clarkb | re the failure that prevented the other two pbr changes from merging: I wonder if we should drop the coverage job. Coverage makes things slower and we hit a timeout | 21:26 |
clarkb | I'm not sure anyone is looking at those reports these days. Though I have found them useful for various reasons at various points in time | 21:26 |
*** dasm is now known as dasm|off | 21:37 | |
fungi | probably more of a discussion for #openstack-oslo since pbr is an oslo deliverable | 22:53 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!