Thursday, 2023-06-29

tonybI recently broke the release-tools code in openstack/project-config:roles/copy-release-tools-scripts/files/release-tools01:57
tonybdo we have any "prior art" for creating testing for that kind of emedded role/tools?01:58
Clark[m]The thing that makes it difficult is being in project-config which is trusted and therefore not able to be run pre review.02:00
Clark[m]You could make copies of jobs/roles/etc in another repo and test them that way or do something like base vs base-test02:00
*** enick_178 is now known as diablo_rojo04:03
*** ralonsoh_ is now known as ralonsoh07:11
SvenKieskeClark[m]: thanks for helping poking pypi people, seems it worked:
SvenKieskeso hopefully no more CI failures due to pypi sending stale data07:55
zigoHi there! What can I do to fix this?
fricklerzigo: might be relevant? also not sure there is anyone still working on that project. stephenfin might know more? finally I'm not sure which channel to best discuss this in. #openstack-dev maybe09:37
stephenfinsqlalchemy-migrate is dead. We can switch off the CI for it if we'd like09:39
zigofrickler: I did a +2w on it, thanks.09:39
zigostephenfin: Well, I still need to maintain zed + antelope with it... This failed:
Supriya_Kotwal_Gerrit review for -
Supriya_Kotwal_Gerrit review for - :  DSM unable to fetch WWNs which are in lowercase10:06
Supriya_Kotwal_ Gerrit review for - :  DSM unable to fetch WWNs which are in lowercase11:11
mnasiadkaSupriya_Kotwal_: I think that's better to discuss in #openstack-cinder ?11:13
Supriya_Kotwal_Thank you.11:13
*** blarnath is now known as d34dh0r5314:13
clarkbSvenKieske: well now we'll get failures when the CDN's primary backend fails. But I think that is safer for all pip users and maybe this will lead to fixes for the secondary backend (or avoiding the primary failures in the first place)15:33
clarkbprogress either way even if it may not result in happier jobs immediately15:33
SvenKieskeclarkb: yes; could you maybe empty the cache once? we still had some failures at least in the morning (european timezone, so like 7 hours ago)15:36
clarkbSvenKieske: the ttls on those files are only 10 minutes15:37
clarkbwe shouldn't need to clear the cache15:37
clarkbas  Isaid you may still get failures when the primary backend fails15:37
clarkbthis won't make the failures go away. But it does represent progress on the pypi side towards possibly correcting things15:37
SvenKieskeah good to know the ttl is that small15:38
SvenKieskeis it so small so the cache doesn't cache too much data? because a cache time of 10 minutes strikes me as rather short?15:39
clarkbSvenKieske: the cache is primarily there for the actual package data not the indexes. THe package data should be cached for our maximum of one day. The indexes have a ttl of 600 seconds last I checked though15:40
SvenKieskemhhm I see15:40
opendevreviewMerged zuul/zuul-jobs master: Use zuul_workspace_root for prepare/mirror workspace roles
TheJuliao/ are there any graphana graphs which tie cloud to overall job outcome (i.e. success/failure) ?16:48
Clark[m]No, you would need to use the open search data for that16:54
*** iurygregory_ is now known as iurygregory16:58
dpanechclarkb: could you add me to this gerrit group starlingx-app-power-metrics-core ? Its for yet another new repo that merged yesterday (
mnaserheads up: GitHub is down, so expect weird stuff for anything pulling from there17:56
mnasereven their static pages, it’s a full on outage. who knows what’s going on there17:58
clarkbdpanech: done18:00
dpanechclarkb: thank you18:02
clarkbI'm able to load some github issues at least18:03
clarkb#status log Deleted an etherpad at the request of its authors.18:41
opendevstatusclarkb: finished logging18:41
*** dtantsur_ is now known as dtantsur23:50

Generated by 2.17.3 by Marius Gedminas - find it at!