| *** diablo_rojo is now known as Guest3022 | 00:55 | |
| *** Guest3022 is now known as diablo_rojo | 01:03 | |
| *** amoralej|off is now known as amoralej | 06:34 | |
| opendevreview | Mark Goddard proposed openstack/releases master: Add final Victoria releases for Kolla projects https://review.opendev.org/c/openstack/releases/+/839703 | 08:21 |
|---|---|---|
| opendevreview | Mark Goddard proposed openstack/releases master: [kolla] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837933 | 08:23 |
| opendevreview | Tobias Urdin proposed openstack/releases master: [xena] Release Horizon 20.1.2 https://review.opendev.org/c/openstack/releases/+/839707 | 08:57 |
| *** diablo_rojo is now known as Guest3062 | 09:09 | |
| elodilles | mgoddard: hi, sorry, i added a question to the final victoria release patch: https://review.opendev.org/c/openstack/releases/+/839703/1#message-f9c229a9044e724ecb3c82bf6b91f90f5660da08 | 09:16 |
| mgoddard | elodilles: thanks, replied | 09:21 |
| elodilles | mgoddard: ack, thx | 09:23 |
| elodilles | mgoddard: just for the record, other option would be to not release from Kolla (and mark previous releases as victoria-em) and let downstream consumers decide whether they want to consume the fixes and build an own package from that that solves their issues. the question is really just for those who have already installed kolla stable/victoria and now would upgrade, could it cause some issues | 09:29 |
| elodilles | for them? | 09:29 |
| elodilles | mgoddard: anyway, i understand that the team decided to release, just wanted to mention this ^^^ o:) | 09:30 |
| elodilles | sorry for the extra round o:) | 09:30 |
| elodilles | (i've +2'd the patch) | 09:31 |
| mgoddard | I think we took the pragmatic approach of keeping things working in a difficult situation | 09:32 |
| mgoddard | thanks for the input & review | 09:32 |
| elodilles | ack :] thanks, too! | 09:33 |
| opendevreview | Merged openstack/releases master: [cinder] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837980 | 10:04 |
| opendevreview | Merged openstack/releases master: Add final Victoria releases for Kolla projects https://review.opendev.org/c/openstack/releases/+/839703 | 10:05 |
| *** dviroel|rover|out is now known as dviroel|rover | 11:23 | |
| elodilles | hberaud: i was about to merge the 'victoria-em patches without any respone' (thanks for their reviews!), but then i thought that tox-docs job would cause release job failures... but thinking it through again i think if there are any errors with a patch's tox-docs job then that patch won't even merge | 11:34 |
| elodilles | hberaud: long story short, i'll start to merge now the patches | 11:34 |
| elodilles | * without any response from PTL or release liaisons | 11:34 |
| hberaud | ack | 11:44 |
| hberaud | elodilles: Concerning the tox-docs failure you meant about the oslo.policy story? or not related? | 11:46 |
| elodilles | hberaud: i meant the oslo.policy issue, yes o:) | 11:50 |
| elodilles | hberaud: i guess either another release needs to be made from oslo.policy or restrict its 3.12.0 release from requirements, right? | 11:51 |
| hberaud | hm since these patches are for victoria I don't think this issue will impact those patches as the pulled version of oslo.policy is not the same for victoria | 11:52 |
| hberaud | but maybe I missed some point | 11:53 |
| hberaud | elodilles: ^ | 11:53 |
| hberaud | Else I asked to my oslo fellows to validate the oslo.policy fix so we'd be able to release it ASAP | 11:59 |
| hberaud | the fix is now validated and will be merged soon | 12:03 |
| elodilles | hberaud: ack, thanks | 12:03 |
| elodilles | hberaud: hmmm, i'm a bit confused now about what is impacted and what isn't o:) | 12:04 |
| elodilles | rel-mgmt's tox-docs was passing today, so that's probably not an issue | 12:05 |
| opendevreview | Merged openstack/releases master: [keystone] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837935 | 12:07 |
| elodilles | hberaud: i've seen actually the same errors on stable branches as well, like on victoria in nova: https://zuul.opendev.org/t/openstack/build/c9c0c08c987a486bb04c17e6d33ef4ee | 12:07 |
| opendevreview | Merged openstack/releases master: [freezer] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837977 | 12:08 |
| opendevreview | Merged openstack/releases master: [watcher] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837955 | 12:08 |
| opendevreview | Merged openstack/releases master: [mistral] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837945 | 12:08 |
| opendevreview | Merged openstack/releases master: [heat] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837940 | 12:08 |
| opendevreview | Merged openstack/releases master: [sahara] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837949 | 12:08 |
| opendevreview | Merged openstack/releases master: [barbican] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837974 | 12:08 |
| opendevreview | Merged openstack/releases master: [adjutant] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837985 | 12:08 |
| elodilles | hberaud: actually it is strange because oslo.policy is upper constrained to 3.5.0 in stable/victoria, but still 3.12.0 is being installed. i'll try to figure this out... | 12:13 |
| *** amoralej is now known as amoralej|lunch | 12:25 | |
| opendevreview | Merged openstack/releases master: [kolla] Transition Victoria to EM https://review.opendev.org/c/openstack/releases/+/837933 | 12:46 |
| *** amoralej|lunch is now known as amoralej | 13:11 | |
| opendevreview | Merged openstack/releases master: Release Oslo's Yoga deliverables https://review.opendev.org/c/openstack/releases/+/839517 | 13:40 |
| ykarel | elodilles, /me also trying to find the same, was looking into neutron failures in stable branches | 13:54 |
| ykarel | they using u-c from master and failing | 13:54 |
| elodilles | ykarel: strange, because in neutron's tox.ini victoria link is used: https://opendev.org/openstack/neutron/src/branch/stable/victoria/tox.ini#L137 | 13:56 |
| ykarel | elodilles, actually ^ is not used from tox.ini version | 13:57 |
| ykarel | but tox_constraints_file: '{{ ansible_user_dir }}/src/opendev.org/openstack/requirements/upper-constraints.txt' | 13:57 |
| ykarel | and from what i see in logs correct requirements branch is cloned | 13:57 |
| ykarel | may be can hold a node and see | 13:58 |
| elodilles | i also saw that the correct victoria branch is checked out from openstack/requirements | 14:01 |
| elodilles | so it is strange :S | 14:01 |
| ykarel | frickler, fungi can you help in getting node held for ^ | 14:05 |
| ykarel | or may be you even already know what can cause this behavior | 14:06 |
| ykarel | https://review.opendev.org/c/openstack/neutron/+/839772 is the patch, job is openstack-tox-docs | 14:12 |
| elodilles | ykarel: maybe that is not needed. i can reproduce it locally :-o | 14:18 |
| fungi | os tthere a summary of the problem? | 14:18 |
| fungi | er, is there | 14:18 |
| ykarel | elodilles, with stable requirements checked out, it's still using master constraints ? | 14:18 |
| elodilles | ykarel: nope, it uses the correct upper-constraints file but still installs the unconstrained version | 14:19 |
| fungi | where's an example build result which demonstrates the problem? | 14:19 |
| ykarel | fungi, https://zuul.openstack.org/build/bef65c74ee9e462ebcaa290ea2ef0b8f | 14:20 |
| ykarel | elodilles, ahh :( | 14:20 |
| elodilles | fungi: oslo.policy was released yesterday, for zed, which causes openstack-tox-docs job to fail in multiple repos, and the strange thing is we see broken jobs even in stable branches, | 14:21 |
| elodilles | fungi: it seems that despite oslo.policy is upper constrained, pip installs the unconstrained oslo.policy | 14:21 |
| elodilles | fungi: and it can be reproduced locally, so it seems something wrong with pip (?) or tox (?) or something other? | 14:23 |
| fungi | docs develop-inst: /home/zuul/src/opendev.org/openstack/neutron | 14:24 |
| fungi | i think we've seen this before, yes, trying to remember what the issue is now but i seem to recall it was related to things being included in test-requirements and/or editable installs | 14:25 |
| fungi | docs builds are special because they include their own separate requirements file | 14:26 |
| ykarel | elodilles, seems you are right, i can also reproduce locally, constraints are not entertained | 14:30 |
| fungi | try adding usedevelop=false in testenv:docs and see if it still happens | 14:30 |
| ykarel | sorry ignore me, it not reproduced on my side | 14:31 |
| fungi | i think the problem is that tox is performing an editable install of neutron as a separate step, and because the constraints option is being passed in the deps setting, it's not applied when the develop inst happens | 14:32 |
| opendevreview | Hervé Beraud proposed openstack/releases master: Release oslo.policy to fix gates https://review.opendev.org/c/openstack/releases/+/839775 | 14:34 |
| fungi | the installation of oslo.policy is happening as a result of tox's develop-inst according to the log anyway | 14:34 |
| hberaud | elodilles: ^ | 14:34 |
| fungi | another option would be to include requirements.txt in the deps list for testenv:docs | 14:35 |
| fungi | then we would preinstall it with the other dependencies | 14:35 |
| elodilles | yepp, fungi is right. i fixed some similar cases in the past.... :S | 14:35 |
| hberaud | elodilles: this new version should goes away the problem and then our gates should recover once landed | 14:35 |
| elodilles | fungi: yepp, i usually did that: added -rrequirements.txt to docs target's deps section as well | 14:36 |
| elodilles | hberaud: cool, that solves our problem on master :) and it seems some project need some update in their 'docs' target :) | 14:37 |
| elodilles | hberaud: thanks! | 14:37 |
| elodilles | fungi: thanks, too! | 14:37 |
| hberaud | np | 14:37 |
| fungi | yw | 14:37 |
| ykarel | elodilles, hberaud fungi Thanks now i got it and can reproduce locally | 14:45 |
| ykarel | will fix it in neutron side | 14:45 |
| elodilles | fungi: interestingly usedevelop=False does not work. but adding -rrequirements does work, at least | 14:52 |
| ykarel | pushed https://review.opendev.org/c/openstack/neutron/+/839777 for neutron | 14:53 |
| ykarel | elodilles, i think with usedevelop project itself will not be installed, and only what defined in deps will be installed | 14:54 |
| fungi | right, usedevelop=false avoids the additional (unconstrained) project install step, but if docs builds need the project to be installed then that still needs to happen some way | 14:58 |
| fungi | e.g. by adding toxinidir or . to the deps | 14:58 |
| fungi | (toxinidir is preferable to . since the latter won't work correctly if you're running tox with cwd as a different directory | 14:59 |
| *** amoralej is now known as amoralej|off | 15:36 | |
| elodilles | hberaud: is there any reason for the oslo.policy release to be MINOR version bump instead of a PATCH version bump? :-o | 15:39 |
| hberaud | elodilles: ah nope sorry my bad | 15:42 |
| hberaud | let me fix that | 15:43 |
| opendevreview | Hervé Beraud proposed openstack/releases master: Release oslo.policy to fix gates https://review.opendev.org/c/openstack/releases/+/839775 | 15:44 |
| elodilles | hberaud: no problem, i was just wondering whether it is necessary o:) | 15:45 |
| elodilles | thanks for the update! | 15:47 |
| hberaud | np | 15:58 |
| *** dviroel|rover is now known as dviroel|rover|lunch | 16:02 | |
| *** dviroel|rover|lunch is now known as dviroel|rover | 16:41 | |
| *** dviroel|rover is now known as dviroel|rover|brb | 20:31 | |
| *** dviroel|rover|brb is now known as dviroel|rover | 23:52 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!