| opendevreview | Manideep proposed openstack/manila master: Allow configuring netapp volume retention period to zero https://review.opendev.org/c/openstack/manila/+/963902 | 03:18 |
|---|---|---|
| stephenfin | carloss: Could you grab https://review.opendev.org/c/openstack/python-manilaclient/+/963677/ when you're around? The ruff patch is right above that, so I'm planning to rebase and ask for a quick approval once those two patches below are in | 10:24 |
| opendevreview | Stephen Finucane proposed openstack/manila master: Rework cover tox target https://review.opendev.org/c/openstack/manila/+/965468 | 10:40 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Migrate setup configuration to pyproject.toml https://review.opendev.org/c/openstack/python-manilaclient/+/963677 | 12:08 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: pre-commit: Bump versions https://review.opendev.org/c/openstack/python-manilaclient/+/963678 | 12:08 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Add ruff https://review.opendev.org/c/openstack/python-manilaclient/+/963679 | 12:08 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Fix outstanding ruff issues https://review.opendev.org/c/openstack/python-manilaclient/+/963680 | 12:08 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Fix outstanding hacking issues https://review.opendev.org/c/openstack/python-manilaclient/+/963687 | 12:08 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Remove unnecessary call https://review.opendev.org/c/openstack/python-manilaclient/+/963688 | 12:08 |
| stephenfin | carloss: gouthamr: Also, it looks like the grenade tests are permafailing on master currently? https://zuul.opendev.org/t/openstack/builds?job_name=manila-grenade&project=openstack%2Fmanila&branch=master&skip=0 | 12:12 |
| carloss | stephenfin: just reviewed the python-manilaclient change, sorry for the delay | 12:13 |
| stephenfin | ty | 12:13 |
| carloss | stephenfin: yes, the grenade job is perma-failing and we are on the process of debugging it | 12:13 |
| carloss | ptg week is making it more difficult | 12:13 |
| carloss | we're seeing the message added on this change: https://review.opendev.org/c/openstack/manila/+/936164 | 12:14 |
| carloss | the job started failing right after https://review.opendev.org/c/openstack/grenade/+/961913 | 12:14 |
| carloss | we're assuming we broke something on the service restart, or are not dealing with it correctly | 12:15 |
| carloss | I have a todo to check on the database and increase verbosity on the transactions with a DNM change | 12:15 |
| stephenfin | carloss: took a look at it before lunch. There's definitely an issue with cleanup but I think that's being made worse by the fact that you're trying to start/stop services by executable name instead of unit name. That's causing grenade to fall back to using ps...which is picking up some stale uwsgi processes | 13:16 |
| stephenfin | https://paste.opendev.org/show/bpzsfsk835rVWitVueYL/ | 13:16 |
| opendevreview | Stephen Finucane proposed openstack/manila master: grenade: Start/stop services using unit names https://review.opendev.org/c/openstack/manila/+/965780 | 13:17 |
| opendevreview | Stephen Finucane proposed openstack/manila stable/2025.2: grenade: Start/stop services using unit names https://review.opendev.org/c/openstack/manila/+/965781 | 13:17 |
| stephenfin | carloss: I think that will _obscure_ the issue and get the gate passing. There's obviously still a larger issue here, though your service is using eventlet which suggests a larger rework is going to need to happen this cycle? | 13:18 |
| stephenfin | sorry, I'd been looking for the wrong process in that previous paste. https://paste.opendev.org/show/bMIkDpf2iv8IW62tcutw/ has the correct one. The fix should still help though? | 13:21 |
| carloss | stephenfin: you're a legend | 14:00 |
| carloss | thanks for the fix | 14:00 |
| carloss | I'll watch the gate | 14:00 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Remove unnecessary call https://review.opendev.org/c/openstack/python-manilaclient/+/963688 | 14:06 |
| opendevreview | Stephen Finucane proposed openstack/python-manilaclient master: Re-enable E501 errors https://review.opendev.org/c/openstack/python-manilaclient/+/965786 | 14:06 |
| opendevreview | Merged openstack/python-manilaclient master: Migrate setup configuration to pyproject.toml https://review.opendev.org/c/openstack/python-manilaclient/+/963677 | 15:10 |
| gouthamr | carloss: a ton of work went into organizing the PTG, i think you deserve a couple weeks off for all the pain :D thanks for putting it all together! | 16:09 |
| gouthamr | i've +2/+W'ed stephenfin's grenade script changes | 16:09 |
| gouthamr | ty stephenfin | 16:09 |
| stephenfin | gouthamr: nw | 16:13 |
| stephenfin | though I do feel the need to stress that this is merely masking the issue :) | 16:13 |
| stephenfin | there's definitely something wrong and I still see stale manila-share processes after the systemctl call returns (devstack does not seem to be checking the retcode but I'd guess it's non-zero). You can also see that it's timing out after 60 seconds in CI https://zuul.opendev.org/t/openstack/build/34379f97140748c79e97bd87c61cc3d6/log/controller/logs/grenade.sh_log.txt#2684-2685 | 16:15 |
| gouthamr | ah i see | 16:15 |
| stephenfin | and that seems to be specific to manila, since I don't see e.g. g-api or n-api exhibiting the same behavior | 16:16 |
| stephenfin | _however_, seeing as you need to dump eventlet this cycle, that problem will hopefully become a non-issue | 16:16 |
| stephenfin | and your reproducer is really simple: deploy manila and run 'sudo systemctl stop devstack@m-shr'. It should return almost immediately. If not, you haven't fixed it | 16:17 |
| stephenfin | Happy to stick all this in a bug report if you don't have one already | 16:17 |
| gouthamr | that would be most helpful! :) | 16:17 |
| gouthamr | i'll look at this right after the PTG discussions | 16:18 |
| gouthamr | your script fixes are valid though, and the current voting job's passing, but, we'll need the SLURP job to be voting as well.. i'll look at that next | 16:19 |
| stephenfin | you probably can/should backport that fix as far as you can go | 16:20 |
| gouthamr | ++ | 16:20 |
| stephenfin | gouthamr: https://bugs.launchpad.net/manila/+bug/2130447 | 16:24 |
| gouthamr | thank you stephenfin | 16:25 |
| carloss | ty fow working on this stephenfin, and for the bug report | 16:26 |
| opendevreview | Merged openstack/manila stable/2025.2: grenade: Start/stop services using unit names https://review.opendev.org/c/openstack/manila/+/965781 | 17:28 |
| stephenfin | gouthamr: Potentially for next week at this point, but I addressed all your comments on ruff https://review.opendev.org/c/openstack/python-manilaclient/+/965786 | 17:43 |
| gouthamr | ty stephenfin; will look | 17:43 |
| stephenfin | hopefully my explanations for what ruff is doing/why it's doing it make sense | 17:43 |
| stephenfin | ruff/black | 17:43 |
| gouthamr | stephenfin: yes i thought it was something like that and painful, did it involve manual fixes? | 17:46 |
| stephenfin | yes :( | 17:46 |
| gouthamr | :( painful, ty for pushing it.. i like all the other standardization we're getting as a part of this! | 17:47 |
| opendevreview | Merged openstack/manila master: grenade: Start/stop services using unit names https://review.opendev.org/c/openstack/manila/+/965780 | 18:00 |
| opendevreview | Merged openstack/python-manilaclient master: pre-commit: Bump versions https://review.opendev.org/c/openstack/python-manilaclient/+/963678 | 18:22 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!