| *** bauzas_ is now known as bauzas | 01:17 | |
| *** bauzas_ is now known as bauzas | 04:10 | |
| *** bauzas_ is now known as bauzas | 06:29 | |
| opendevreview | Merged openstack/releases master: Add libpcre3-dev to bindep https://review.opendev.org/c/openstack/releases/+/926860 | 07:01 |
|---|---|---|
| frickler | I'm going to rerun the other failed tag-release job from yesterday to test ^^ https://zuul.opendev.org/t/openstack/build/643cd70dca7546b28a1192c60a090a0a | 07:17 |
| frickler | hmm, wait, propose-update-constraints is also failing in a very weird way https://zuul.opendev.org/t/openstack/build/bad50a3664bd4bed96d95ea3bd6e8b67 | 07:26 |
| opendevreview | Slawek Kaplonski proposed openstack/releases master: Release final os-ken for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926767 | 07:26 |
| frickler | ok, rerunning the above and holding a node for debugging the propose job | 08:05 |
| elodilles | it seems that 'python3 setup.py --name' gave some garbage instead of the distname | 08:20 |
| frickler | yes, well not instead, but in addition to it | 08:20 |
| elodilles | i remember seeing this error in the past, but don't remember the fix :/ | 08:21 |
| frickler | publish-tox-docs-releases was still failing :( https://zuul.opendev.org/t/openstack/build/6a748d48e2684ab1aab8d173a59ce214 | 08:22 |
| elodilles | yeah, that is what clarkb was talking about some hours earlier: https://meetings.opendev.org/irclogs/%23openstack-release/%23openstack-release.2024-08-21.log.html#t2024-08-21T22:23:35 | 08:26 |
| frickler | bindep didn't install the dev lib. either something weird is happening or the library is there and the fix must look different | 08:27 |
| frickler | elodilles: yes, that's why I merged https://review.opendev.org/c/openstack/releases/+/926860 earlier, but didn't seem to help | 08:27 |
| elodilles | but that is for releases, and not for the job directly. at least that is how i understood clarkb's explanation | 08:29 |
| frickler | hmm, the job does use bindep from releases repo https://zuul.opendev.org/t/openstack/build/6a748d48e2684ab1aab8d173a59ce214/console#1/0/14/ubuntu-noble | 08:32 |
| elodilles | oh. hmm. but then i don't see git-review | 08:34 |
| elodilles | libre2-dev | 08:34 |
| elodilles | python3-all-dev | 08:34 |
| elodilles | ahh, copy buffer... | 08:34 |
| elodilles | so. then i don't see libpcre3-dev in stdout. so it wasn't installed, right? | 08:35 |
| frickler | elodilles: that may be because it might have been present already? iiuc bindep only lists pkgs that need installing | 08:40 |
| frickler | guess I need to setup another hold to check in place and we need to approve another release, then | 08:41 |
| frickler | elodilles: oslo.config seems like a good candidate, if you would be so kind? hold is in place https://review.opendev.org/c/openstack/releases/+/926544 | 08:43 |
| elodilles | frickler: looking | 08:47 |
| elodilles | there you go | 08:49 |
| tkajinam | https://zuul.opendev.org/t/openstack/build/ff8f60902f4d4b128eeff466bae769e4 | 08:51 |
| tkajinam | 2024-08-22 08:33:27.483588 | ubuntu-noble | + git checkout -b '[pbr] Generating ChangeLog | 08:52 |
| tkajinam | 2024-08-22 08:33:27.483667 | ubuntu-noble | python-glanceclient-4.7.0' | 08:52 |
| tkajinam | 2024-08-22 08:33:27.483706 | ubuntu-noble | fatal: '[pbr] Generating ChangeLog | 08:52 |
| tkajinam | 2024-08-22 08:33:27.483897 | ubuntu-noble | python-glanceclient-4.7.0' is not a valid branch name | 08:52 |
| tkajinam | just in case you are not aware of this ... | 08:52 |
| frickler | tkajinam: that's one of the issues we were talking about earlier, I have a held node to debug it | 08:54 |
| tkajinam | ah, yes. you mentioned propose-update-constraints . sorry I overlooked that | 08:54 |
| frickler | np, thx for the heads up, easy to not notice such things | 08:56 |
| opendevreview | Merged openstack/releases master: Release oslo.config 9.6.0 for 2024.2 dalmatian. https://review.opendev.org/c/openstack/releases/+/926544 | 08:58 |
| elodilles | frickler: i found my dummy workaround for a similar case: https://review.opendev.org/c/openstack/releases/+/879065 | 09:07 |
| elodilles | for the pbr changelog thing | 09:07 |
| elodilles | i can propose the same workaround to project-config if we don't have a better fix | 09:11 |
| frickler | elodilles: ah, nice, that looks pretty similar. I need to afk for a bit, will verify on the held node, then | 09:20 |
| elodilles | ACK, i might also disappear for some minutes, so i uploaded a workaround patch: https://review.opendev.org/c/openstack/project-config/+/926885 | 09:29 |
| frickler | so I can reproduce the issue and your fix seems to be fine as a workaround. the real fix should happen in pbr though. it is using distutils.log to output that line, which is vendored in setuptools for py312 and has a note "Retained for compatibility and should not be used." | 11:02 |
| frickler | going to approve your change now | 11:02 |
| elodilles | ACK, thanks | 11:05 |
| frickler | for later reference: https://opendev.org/openstack/pbr/blame/branch/master/pbr/git.py#L233 | 11:08 |
| frickler | hmm, weird, the last run of the docs build worked https://zuul.opendev.org/t/openstack/build/f9516d140b9b4dfb9d748edc8f7702a8 it also did install the pcre lib via bindep | 11:16 |
| frickler | not sure why this was failing earlier | 11:16 |
| frickler | thinking about it, likely because I re-enqueued the earlier hash from the releases repo and in that scenario it doesn't get rebased onto master like for the check pipeline | 11:17 |
| elodilles | at least it works now :) | 11:25 |
| elodilles | is there anything else that needs to be fixed? | 11:26 |
| elodilles | and of course: re-enqueue? | 11:26 |
| frickler | I don't think we can re-enqueue the release buildsets where the propose-constraints failed, since that would try to republish to pypi. so we'll need to generate the three missing u-c bumps manually | 11:32 |
| frickler | everything else should be fine by now, afaict | 11:33 |
| elodilles | ACK, then i'll approve a release patch, to double-check (and because there is one that can be merged :)) | 11:34 |
| elodilles | (oslo.utils) | 11:35 |
| opendevreview | Merged openstack/releases master: Release oslo.utils 7.3.0 for 2024.2 dalmatian. https://review.opendev.org/c/openstack/releases/+/926622 | 11:43 |
| elodilles | there it is: https://review.opendev.org/c/openstack/requirements/+/926904 | 12:17 |
| elodilles | \o/ | 12:17 |
| frickler | cool, now I'm curious to see whether the delay helped | 12:23 |
| elodilles | well, i checked and pypi had oslo.utils at the time when the pause started o:) | 12:30 |
| elodilles | at least, according to pypi's rss ( https://pypi.org/rss/project/oslo.utils/releases.xml ) | 12:32 |
| frickler | iiuc the delay is needed due to our proxy, which caches the index fro 10m | 12:33 |
| fungi | yeah, the apache mod_proxy/mod_cache we use for pypi in our nodepool providers caches responses from pypi for up to 10 minutes. we might be able to work out a way to signal cache invalidation from jobs which are more timing-sensitive to new uploads, but that would need some research | 12:46 |
| *** bauzas_ is now known as bauzas | 12:48 | |
| elodilles | i see. then ignore my above lines :X | 12:52 |
| ykarel | hi can some one check https://review.opendev.org/c/openstack/releases/+/926724 | 12:59 |
| elodilles | ykarel: +2'd | 13:07 |
| frickler | +3 | 13:09 |
| ttx | OK I'll do some reviews now if things are back to normal | 13:09 |
| fungi | seems to be so far | 13:12 |
| elodilles | yepp | 13:13 |
| opendevreview | Merged openstack/releases master: Release oslo.serialization 5.4.1 for 2024.1 (Caracal) https://review.opendev.org/c/openstack/releases/+/926724 | 13:22 |
| ykarel | thx elodilles frickler | 13:26 |
| opendevreview | Merged openstack/releases master: Release final mistral-extra for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926762 | 13:29 |
| opendevreview | Merged openstack/releases master: Release final ironic-lib for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926758 | 13:29 |
| opendevreview | Merged openstack/releases master: Release final sushy for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926770 | 13:31 |
| opendevreview | Merged openstack/releases master: Release final os-ken for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926767 | 13:34 |
| opendevreview | Merged openstack/releases master: Release final stevedore for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926769 | 13:34 |
| opendevreview | Merged openstack/releases master: Release final taskflow for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926771 | 13:37 |
| opendevreview | Merged openstack/releases master: Release oslo.log 6.1.2 for 2024.2 dalmatian. https://review.opendev.org/c/openstack/releases/+/926712 | 13:37 |
| dansmith | who can approve requirements changes? frickler? need this landed to proceed with the inspector stuff: https://review.opendev.org/c/openstack/requirements/+/926904 | 13:45 |
| frickler | dansmith: I was hoping that the reqs team could handle those. also the change hasn't even passed CI yet | 13:53 |
| dansmith | frickler: I went looking at the requirements-core group and it seems like not a lot of still-active people | 13:54 |
| dansmith | I know it still needs CI, I was just hoping to line up some people | 13:55 |
| frickler | yes, one can't say the reqs team is overstaffed | 13:58 |
| dansmith | looks like they're merging with only one core so I guess that's helpful | 13:58 |
| Clark[m] | fungi: the easiest and possibly most correct thing would be to talk to PyPI directly. However, PyPI may have cached things for 10 minutes at their cdn as well | 14:04 |
| dansmith | prometheanfire: there's a lot of work to do stacked up behind this.. if you could prioritize approving it I'd sure appreciate it: https://review.opendev.org/c/openstack/requirements/+/926904 | 14:19 |
| dansmith | prometheanfire: almost entirely new yet-unused code added to oslo.utils, so hopefully pretty uncontroversial | 14:20 |
| fungi | dansmith: also, no idea if it helps, but there's an #openstack-requirements channel for that team too | 14:26 |
| dansmith | okay sorry | 14:27 |
| elodilles | well, Release Managers group is added to requirements-core, so for example I could review things, too, but i never felt the authority to review things there o:) | 14:40 |
| elodilles | though new-release patches could be the ones that relmgt team members could approve anytime i guess | 14:41 |
| dansmith | tbh, I'm not sure what the big deal is with these.. they bump a minor version and all the tests pass, so what is anyone else going to really do other than react if we merge and it breaks something? | 14:43 |
| elodilles | that's true :) | 14:44 |
| elodilles | i fast-approved the oslo.utils patch a couple of sec ago anyway :D | 14:45 |
| dansmith | elodilles: many thanks! | 14:46 |
| elodilles | np | 14:46 |
| elodilles | i've also +2+W'd the rest of the new-release library u-c bumps that have Verified+1 | 14:51 |
| prometheanfire | elodilles: thanks for getting to it :D | 15:15 |
| elodilles | prometheanfire: np o:) | 15:22 |
| *** bauzas_ is now known as bauzas | 18:26 | |
| *** melwitt is now known as jgwentworth | 19:38 | |
| *** jgwentworth is now known as melwitt | 19:38 | |
| opendevreview | Brian Rosmaita proposed openstack/releases master: Release final os-brick for 2024.2 (Dalmatian) https://review.opendev.org/c/openstack/releases/+/926766 | 22:02 |
| *** bauzas_ is now known as bauzas | 22:59 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!