Wednesday, 2026-01-28

elodilleshmmm. it seems that https://releases.openstack.org did not get refreshed yesterday :/ it still shows the old state of victoria, wallaby, and xena09:02
elodillesany idea why that did not work? :-o09:03
opendevreviewMerged openstack/releases master: Release Swift 2.37.0  https://review.opendev.org/c/openstack/releases/+/97494312:10
opendevreviewMerged openstack/releases master: Release ironic for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97447812:19
opendevreviewMerged openstack/releases master: Release ironic-python-agent-builder for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97447712:20
opendevreviewMerged openstack/releases master: Release networking-baremetal for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97448812:26
opendevreviewMerged openstack/releases master: Release networking-generic-switch for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97449112:30
opendevreviewMerged openstack/releases master: Release sushy for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97451612:46
opendevreviewMerged openstack/releases master: Release ceilometer for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97446712:55
opendevreviewMerged openstack/releases master: Release magnum for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97448113:06
opendevreviewMerged openstack/releases master: Release python-magnumclient for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97451113:07
opendevreviewMerged openstack/releases master: Release magnum-ui for stable/2024.2  https://review.opendev.org/c/openstack/releases/+/97448013:08
elodillesokay, release.o.o has refreshed. so it just needed some time. (~24hrs?)15:12
*** gthiemon1e is now known as gthiemonge16:00
ttxMaybe it needed a release to land to trigger the refresh16:02
elodillesoh, i see. i thought it refreshes like every hr or so, if there was any change16:08
opendevreviewMerged openstack/releases master: Update 2024.2 Dalmatian's transition date  https://review.opendev.org/c/openstack/releases/+/97474016:51
fricklerfrom what I see, releases.o.o is updated by the publish-tox-docs-releases job after each merge. sadly we do not guarantee that these jobs run in the order of changes merged. so yesterday 974498 and 974814 merged in this order at almost the same time, but the publish job for the first change ran after the one for the second change, publishing the un-eoled status17:51
fricklerI'm not sure whether we can configure the release-post pipeline differently, maybe fungi or clarkb can take a look at that. otherwise the next merged change will (and did) fix things again17:52
clarkbI think that this is why there is a semaphore that we disable during the release process (the semaphore enforces the job order the rest of the year)17:59
fricklerclarkb: well it enforces serialization of the jobs, but not necessarily preserving merge order it seems?18:00
clarkbfrickler: no merge order should be preserved by git tree or depends on18:03
clarkbI think the enqueue order to the gate should also be preserved18:04
clarkbfrickler: in your statement above it sounds like the merge roder was correct. The problem was the order of the jobs running18:04
clarkb`the publish job for the first change ran after the one for the second change` the semaphore exists to prevent that aiui18:04
fungifrickler: clarkb: where's the semaphore? maybe we don't have one for the release-post jobs? (a supercedent pipeline would enforce the run order, but ref-updated pipelines are dependent on gerrit event timing)18:40
clarkbfungi: I was talking more generally . We have a semaphore that exists to address this particular problem. I don't knwo if it is applied to these specific jobs18:55

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!