| ttx | I'll fast-track openstacksdk 4.20 as the more we wait the more of a problem it creates | 07:12 |
|---|---|---|
| opendevreview | Merged openstack/releases master: openstacksdk 4.20.0 https://review.opendev.org/c/openstack/releases/+/1002790 | 07:25 |
| ttx | Hmm that release failed ... will probably need reenqueing | 08:02 |
| frickler | hmm, it sounds like we maybe shouldn't run ensure-pip in post again if it already ran in pre, but this looks like hopefully there was only a temporary error from pypi. I verified that the release wasn't uploaded, going to reenqueue now | 08:15 |
| frickler | oh, I missed the module_stderr before: "/bin/sh: 1: sudo: not found | 08:24 |
| frickler | so the error is reproducible, someone will need to dig deeper into why this is broken, not sure I'll get to that soon | 08:25 |
| frickler | the delta to a passing run 3 days ago is that "TASK [ensure-pip : Check if pip is installed]" is now failing with "/usr/local/bin/python3: No module named wheel", triggering the "Install pip from packages" task, which fail since we did remove-sudo in pre | 09:09 |
| frickler | I assume something has changed in noble's python pkgs, since I didn't find any related recent change on our end | 09:11 |
| elodilles | frickler: should we stop approving release patches? if i'm not mistaken this affects all the other releases, right? | 09:12 |
| frickler | I would assume so, yes, unless there is something special about the sdk build that triggers this. you could create a release-test release if you wanted to verify | 09:23 |
| elodilles | ACK, will try then | 09:56 |
| opendevreview | Elod Illes proposed openstack/release-test master: Add py315 classifier https://review.opendev.org/c/openstack/release-test/+/1003029 | 10:18 |
| frickler | I've set up a hold and will trigger the reenqueue for the sdk release once again to further debug | 10:21 |
| elodilles | +1 | 10:22 |
| opendevreview | Merged openstack/release-test master: Add py315 classifier https://review.opendev.org/c/openstack/release-test/+/1003029 | 10:32 |
| frickler | oh, the hold doesn't help me, the playbook is run on localhost, i.e. on the executor | 10:41 |
| opendevreview | Elod Illes proposed openstack/releases master: Release test release release-test https://review.opendev.org/c/openstack/releases/+/1003031 | 10:42 |
| frickler | so the issue is likely triggered by https://review.opendev.org/c/opendev/system-config/+/1002425 which rolled out over the weekend. let's move this to #opendev | 10:44 |
| elodilles | +1 | 10:45 |
| frickler | meh, I would, if my matrix server wouldn't be undergoing maintenance right now :( | 10:47 |
| elodilles | /o\ | 10:47 |
| elodilles | at least it's not Friday o:) | 10:48 |
| *** croeland1 is now known as croelandt | 13:27 | |
| cardoe | frickler: stephenfin: I've +W'd the ironic fixes for ksa. | 13:49 |
| stephenfin | cardoe: ty 🙏 forgot I could ask you 😅 | 13:54 |
| frickler | looks like nova also merged the change, so we should be able to proceed with the reqs bump soon | 13:55 |
| elodilles | \o/ | 13:57 |
| cardoe | ttx: frickler: should openstacksdk 4.20.0 still not be on pypi? | 14:43 |
| fungi | cardoe: correct, it hit a problem with the new python3.14-based zuul ecexutor images | 14:51 |
| fungi | executor | 14:51 |
| fungi | we're holding before we decide on a fix and reenqueue | 14:51 |
| fungi | discussion is in progress in the #opendev:opendev.org matrix room | 14:52 |
| cardoe | ah | 14:52 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!