Monday, 2026-05-11

-@gerrit:opendev.org- Michal Nasiadka proposed: [openstack/project-config] 988041: Switch the requirements-constraints job to py311 https://review.opendev.org/c/openstack/project-config/+/98804106:55
-@gerrit:opendev.org- Michal Nasiadka proposed: [openstack/project-config] 988041: Switch the requirements-constraints job to py311 https://review.opendev.org/c/openstack/project-config/+/98804106:58
-@gerrit:opendev.org- Michal Nasiadka proposed: [openstack/project-config] 988041: Switch the requirements-constraints job to py311 https://review.opendev.org/c/openstack/project-config/+/98804106:58
-@gerrit:opendev.org- Zuul merged on behalf of Michal Nasiadka: [openstack/project-config] 988041: Switch the requirements-constraints job to py311 https://review.opendev.org/c/openstack/project-config/+/98804107:53
-@gerrit:opendev.org- Zuul merged on behalf of Bartosz Bezak: [openstack/project-config] 986493: kayobe: Add reviewer tier with +2 rights https://review.opendev.org/c/openstack/project-config/+/98649309:31
-@gerrit:opendev.org- Stephen Finucane proposed: [openstack/project-config] 988094: Add nova-approvers to openstack/nova ACL https://review.opendev.org/c/openstack/project-config/+/98809411:10
-@gerrit:opendev.org- Stephen Finucane proposed: [openstack/project-config] 988094: Add nova-reviewers to openstack/nova ACL https://review.opendev.org/c/openstack/project-config/+/98809411:13
@jim:acmegating.comClark: it looks like there may be a problem with the temporary image storage: i see "401 Client Error: Unauthorized for url" in the zuul-launcher logs14:44
@jim:acmegating.comi'm looking into it14:44
@fungicide:matrix.orgtemporary image storage is flex swift?14:47
@jim:acmegating.comyep14:47
@clarkb:matrix.orgcorvus: the error with the previous container was 404 not found. It would show up when listing containers with openstack client but running container show also returned 40414:49
@clarkb:matrix.orgso 401 is a new one to me at least14:49
@jim:acmegating.comi *think* we may just need to flip the public access bit14:49
@jim:acmegating.comof course i can't compare to the prev container :|14:49
@clarkb:matrix.orgoh are we anonymously reading the data back when fetching it rahter than authenticating?14:50
@clarkb:matrix.orgif so then flipping that bit does make sense to me14:50
@jim:acmegating.comyep14:50
@jim:acmegating.comok i will do so14:50
@fungicide:matrix.orgmakes sense, and would explain 401 unauthorized responses14:51
@clarkb:matrix.orgfungi: for https://review.opendev.org/c/opendev/system-config/+/985834 did you check the held node? I think I'd like to proceed with that change if someone else has checked the held node.14:52
@clarkb:matrix.orgfungi: then separately there is the therpad 2.7.3 upgrade whcih I haven't checked the held node (I checked 2.7.2 but haven't made it to the 2.7.3 update yet)14:52
@jim:acmegating.comthe default in zuul-launcher is to assume public urls.  the aws driver does have a code path where it can download images via s3 using authentication.  we could add that to the openstack driver, but -- i think for opendev that public url is a benefit (anyone can just fetch our images), so i think we want to keep using public urls here.14:52
@jim:acmegating.comi flipped the switch and started anonymously downloading urls14:53
@clarkb:matrix.orgya I think in this case it should be fine. We don't treat these images as secret data or anything like that14:53
@fungicide:matrix.orgi agree, public is better. it also aligns with our usual stance on openness and transparency in operations14:55
@fungicide:matrix.orgClark: i did not check the held node at the time, but will in a moment and can then approve if you're ready for it14:55
@clarkb:matrix.orgalso as a heads up I have to do a bunch of afternoon errands today (conflicting appointments and only one car mean I get to play taxi driver). Depending on how that goes I may not get the meeting agenda out until late, But I should still be able to get it out today14:55
@clarkb:matrix.orgfungi: that would be great thank you14:56
@clarkb:matrix.orgfungi: the child change is the one that held a node14:56
@jim:acmegating.comi think the launchers may need some prodding to retry these uploads; i will restart them14:56
@fungicide:matrix.orglooks like 200.225.47.56 is the held haproxy14:57
@clarkb:matrix.orgyup that is the IP I see in the log14:58
@jim:acmegating.com#status log set public access bit on zuul image storage swift container in rax-flex and restarted launchers to fix image uploads14:58
@status:opendev.org@jim:acmegating.com: finished logging14:58
@jim:acmegating.comokay looks like image uploads are doing their thing15:01
@jim:acmegating.comit may be a while until we actually see resolute images15:02
@clarkb:matrix.orgthank you for fixing that15:06
@fungicide:matrix.org`release-volumes.py` has been running on mirror-update since some time yesterday, releasing the project.tarballs volume. this is blocking all other static volume updates. i'm lokking into it now16:21
@fungicide:matrix.organd looks like the ubuntu-ports mirror has exceeded its quota16:27
@fungicide:matrix.orgi don't see any significant change/increase in usage for project.tarballs over the past week16:30
@fungicide:matrix.org`fs listquota` doesn't show much difference between the rw and ro volume use either (not even 1k), but maybe it's almost done syncing now16:32
@clarkb:matrix.org104.239.175.4 is the held etherpad 2.7.3. I'm going to create a clarkb-test pad there now and start checking it16:43
@clarkb:matrix.orgThis is interesting 2.7.3 gives you a pad deletion token the first time you open a new pad if you are the creator of a pad 16:45
@clarkb:matrix.orgmy second browser did not get the token when opening the same pad. I suspect this is fine. It might catch some people off guard, but functioanlly I think this is probably safe enough16:46
-@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [opendev/system-config] 985834: Update Gitea to 1.26.1 https://review.opendev.org/c/opendev/system-config/+/98583416:51
@clarkb:matrix.orggitea09 has just updated16:55
@clarkb:matrix.orgmy client hits gitea10 so I'll check things momentarily once gitea10 updates16:55
@clarkb:matrix.orgbrowsing gitea10 seems to work for me16:57
@clarkb:matrix.orgall 6 backends should be upgraded now17:05
@clarkb:matrix.orghttps://zuul.opendev.org/t/openstack/buildset/491239ae82ae4f06ad48a95ec6e7ac7a reports success. I am also able to clone system-config17:06
@clarkb:matrix.orgNow to find something that has replicated after the upgrade17:06
@fungicide:matrix.orgyeah, looks like it's working to me17:07
@clarkb:matrix.orghttps://review.opendev.org/c/starlingx/app-prometheus/+/986843 has replicated to https://opendev.org/starlingx/app-prometheus/commit/839ca6b0a62691bb4d2f96d5ff7ef1aa6084141917:13
@clarkb:matrix.orgso replication also looks good17:14
@fungicide:matrix.orgagreed17:14
@clarkb:matrix.orgI've done some very early meeting agenda updates. Please let me know if there are other updates you'd like to see made. And as noted earlier I probably won't send that out until very late today, but should still get it out for some value of today (probably in my timezone at least)17:53
-@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [opendev/system-config] 987871: Patch postorius https://review.opendev.org/c/opendev/system-config/+/98787119:03
@fungicide:matrix.org#status log Increased the mirror.ubuntu-ports AFS volume quota by 100GB (from 1TB to 1.1TB) bringing utilization down to 91%19:46
@status:opendev.org@fungicide:matrix.org: finished logging19:46
@fungicide:matrix.orgi've got the lock for that volume held and am manually running the mirror update script now19:46
@fungicide:matrix.orgit's only behind by ~1.5 weeks so hopefully shouldn't take long to complete19:47
@fungicide:matrix.orgthe project.tarballs release completed roughly half an hour ago, unblocking the other volumes19:48
@fungicide:matrix.orglooks like reprepro may have made a ton of zero-length package writes when the quota was exceeded, so i'll probably have to do a ton of manual surgery to get it updating successfully again20:05
@fungicide:matrix.orgyeah, in the end it reports an incorrect size on two debs. i'll have to pick that up tomorrow, but at least the quota should be sufficient again for now20:57
@clarkb:matrix.orgthank you for digging into that20:58
@clarkb:matrix.orgI've made it through the Gerrit 3.13 release notes and edited https://etherpad.opendev.org/p/gerrit-upgrade-3.13 with info. Mostly generated a large TODO list more than anything else21:09
@clarkb:matrix.orgthe next step is going to be working through that stuff with a held node to test and then finalizing the actual upgrade process21:09
@clarkb:matrix.orgIf possible I'd like to do that testing against https://review.opendev.org/c/opendev/system-config/+/987375 built images if we can maybe plan to land that sometime this week and update 3.12 to 3.12.721:10
@clarkb:matrix.orgI don't see anything in there that super concerns me. Lots of things to double check just to amke sure they do what I expect them to. But on the whole this upgrade looks pretty straightforward. Which is good because I think that means late May/early June is still a possibility21:11
@clarkb:matrix.orgwe can discuss timing of the Gerrit image change in our meeting tomorrow (its on the agenda now)21:12

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!