Monday, 2026-06-08

-@gerrit:opendev.org- Zuul merged on behalf of Elod Illes: [openstack/project-config] 989179: [release-tool] check-approval use only governance data https://review.opendev.org/c/openstack/project-config/+/98917909:49
@harbott.osism.tech:regio.chatnew gerrit is spamming my tab list even more than before, the "Your / Changes" action now always spawns a new tab instead of reusing the current one :-/06:07
@harbott.osism.tech:regio.chatit also looks like gerrit is showing false positives for merge-conflicts now, like https://review.opendev.org/c/openstack/requirements/+/987877 and https://review.opendev.org/c/openstack/requirements/+/987879 are two patches that are part of a single stack, still gerrit shows each in the merge-conflicts list for the other one, which doesn't make sense to me at all06:26
@mordred:waterwanders.comJens Harbott: oh - I hadn't noticed the Your / Changes in a new tab. that's ... annoying10:21
@tkajinam:matrix.orgI was about to report that tab behavior :-P14:08
@tkajinam:matrix.orgit's wiered that behavior is specific to items under "your" and changes / open does not open a new tabe for example.14:09
@clarkb:matrix.orgJens Harbott: I wonder if it is checking them directly against each other full combinatorial mode rather than taking the tree into account14:25
@clarkb:matrix.orgBut that may explain the conflict. They would be in conflict if they didn't have a built in resolution via the tree relationship?14:26
-@gerrit:opendev.org- Roja Eswaran proposed: [openstack/diskimage-builder] 987717: Replace debootstrap with mmdebstrap https://review.opendev.org/c/openstack/diskimage-builder/+/98771714:48
@clarkb:matrix.orgfor the new tab opening behavior what link are you selecting? There is no Your -> Changes option14:50
@clarkb:matrix.organd clicking the links under the your menu does not appear to open them in new tabs for me in firefox14:51
@fungicide:matrix.orgi'm going to disappear shortly to grab lunch and run several errands, should be back around 17:30 unless something takes longer than planned15:11
-@gerrit:opendev.org- Clark Boylan proposed: [opendev/system-config] 991946: Add Gerrit 3.14 image builds and testing https://review.opendev.org/c/opendev/system-config/+/99194615:33
@mordred:waterwanders.comClark: ^^ that's what opens in a new tab for me16:55
@clarkb:matrix.orghuh I literally do not have that link. I wonder is it possible you're still using the 3.12 web client?16:58
@clarkb:matrix.orgI would take a screenshot but opening the screenshot tool seems to close the drop down :/16:59
@jim:acmegating.comi may be in the same state as Clark17:00
@clarkb:matrix.orgmaybe I've got the old client? But I've hard refreshed many times. I have Your -> Dashboard / Draft Comments / Edits / Watched Changes / Starred Changes / All Visible Changes/ Groups17:01
@mordred:waterwanders.comI just tried hard-refreshing, menu is the same17:04
@jim:acmegating.comClark: shift-printscr closes the menu for me, but just hitting printscr and then opening in the gimp and cropping works17:04
@mordred:waterwanders.comOH - well lookie there ...17:05
@clarkb:matrix.orgI just checked the two held nodes from testing in separate/new private browsing windows to avoid caching. Logged in as Zuul and both 3.12 and 3.13 have what corvus and I are seeing17:05
@clarkb:matrix.orghttps://213.32.78.192 <- 3.12 https://158.69.71.235 <- 3.1317:06
@mordred:waterwanders.comJens Harbott: ^^ check your preferences17:06
@jim:acmegating.commordred: did you create the "Unfinished" menu item?17:06
@clarkb:matrix.orgmordred: oh interesting my config has new tab all unchecked and I guess the names are user specific defined through user config17:06
@mordred:waterwanders.comcorvus: I think so - but like, a _loooong_ time ago?17:07
@clarkb:matrix.orgso maybe those of us that haven't edited those entries get defaults that work the way we expect but for those of you that have edited them the new tab is added? Or vice versa17:07
@clarkb:matrix.orgI guess that solves the mystery around the behavior but not why we have different configs. But also that probably isn't a big deal17:09
@mordred:waterwanders.comnah. totally solvable. also - I hadn't noticed it (because I normally have a billion tabs anyway) until Jens pointed it out. so it might not be a deep widespread issue17:11
@clarkb:matrix.org991946 built and tested our first Gerrit 3.14 image and reports success on that subset of jobs (other jobs are still running). I have an appointment in a few minuets i need to pop out to but then I'll look at a followup change to test the 3.13 to 3.14 upgrade path17:17
@clarkb:matrix.organd then I'll work on getting the meeting agenda out. Please let me know if there is anything to add to the agenda or feel free to add it yourself17:18
@ashinclouds:matrix.orgQuestion, are there any known issues post upgrade, I keep getting an error on changes about how robot comments are not allowed17:21
@mordred:waterwanders.comwell - most relevant, robot comments aren't a thing anymore, so robots have to use normal comment api17:24
@ashinclouds:matrix.orgI'm just using the webui and clicked into a change recently updated and got a popup error which then cleared like a minute later 17:25
@clarkb:matrix.orgashinclouds: yes this was noted in teh announcement email17:26
@clarkb:matrix.orghttps://lists.opendev.org/archives/list/service-announce@lists.opendev.org/thread/UIFRQR34WYMF3KDMNQFMO222DIK3HTUD/ basically the api doesn't exist anymore so you may need to hard refresh your client to get updated client code that stops fetching the api17:27
@clarkb:matrix.org`You can safely dismiss these error messages then perform a hard refresh of your browser client to fetch the new 3.13 client implementation which should prevent it from happening again.`17:27
@ashinclouds:matrix.orgHmm, interesting. Okay, thanks!17:27
@clarkb:matrix.orggerrit does version specific client bundles (browser specific too iirc), but for whatever reason browsers don't always see the current code as invalid when there is new code17:28
@clarkb:matrix.orgso sometimes you need to force it via a hard refresh. This is a noticeable scenario for that because the api is simply gone and the prior version uses it17:28
@fungicide:matrix.orgit's been that way for so many releases now, you'd think they'd have implemented client versioning with a feature to pop up a messagebox or something at least telling you that your browser window needs a hard refresh to update to the latest client (even if it's only via the oob notifications channel, which admittedly i set my browser to ignore/deny)17:49
@ashinclouds:matrix.orgIt didn't do that, just said there was an error. FWIW. I guess even with a few refreshes, its likely picking up old content from the local cache or something funky18:17
@clarkb:matrix.orgI think you need to do a hard refresh not a normal one18:55
@clarkb:matrix.orgctrl + shift + r or whatever it is18:55
@fungicide:matrix.orgyes, the gerrit webui is a monolithic javascript application implemented entirely client-side in the browser, and merely performs callouts to gerrit's rest api, so a simple reload doesn't replace the local application18:58
@ashinclouds:matrix.orgYeah, short of auditing the headers placed on all original respondant files, no way to really know beyond just trying again. On a plus side, it hasn't happened again today so \o/19:10
@clarkb:matrix.orginfra-root I've got a change to upgrade grafana here: https://review.opendev.org/c/opendev/system-config/+/991155 with a held node at https://23.253.20.165/ I've spot checked some dashboards and they seem to function. Should we go ahead and upgrade?19:21
@mordred:waterwanders.comClark: my spotchecks look ok. they also underscore some opportunities for cleaning up old unused dashboards19:24
@clarkb:matrix.orgya grafana has been pretty append only since deleting dashboards is a bit manual iirc19:25
@clarkb:matrix.orgbut agreed we could trim down some of them19:26
@clarkb:matrix.organd thank you for checking19:26
@mordred:waterwanders.comI haven't found any that have data in prod that don't look pretty much the same on your held node19:26
-@gerrit:opendev.org- Clark Boylan proposed:19:28
- [opendev/system-config] 991946: Add Gerrit 3.14 image builds and testing https://review.opendev.org/c/opendev/system-config/+/991946
- [opendev/system-config] 992342: Test Gerrit 3.13 to 3.14 upgrades https://review.opendev.org/c/opendev/system-config/+/992342
@fungicide:matrix.orgi'm also unsure how to go about clearing out old volumes from the afs dashboard on grafana, after deleting the references in our config didn't do it19:35
@clarkb:matrix.orgfungi: iirc there is an api token like etheropad used to use on disk somewhere. Then we have to make an explicit api call to delete the dashboard using that api token/key19:38
@fungicide:matrix.orgwell, in the case i'm talking about, the dashboard itself is still useful it just contains some graphs that are no longer in its configuration afaict19:42
@clarkb:matrix.orgoh I thought grafyaml will update definitions19:44
@clarkb:matrix.orgbut maybe only to graphs that you want to keep and it won't remove old graphs. In theory deleting the whole dashboard and letting grafyaml recreate it is safe19:44
@fungicide:matrix.orgentirely possible it does and i've missed something in the cleanup19:44
@fungicide:matrix.orgbut i've been hesitant to try to remove anything else until i figure out why the removed volumes are still being displayed19:45
@clarkb:matrix.orgif the graphs are autogenerated from graphite contents then simply removing them from afs itself may not clean up graphite so the graphs stay around too19:47
@clarkb:matrix.orgI don't know if those graphs work that way but some of them do in grafyaml. Where you query for a set of options out of graphite then graph them from there19:47
@fungicide:matrix.orghttps://grafana.opendev.org/d/9871b26303/afs shows a graph for mirror.openeuler and i don't know where that's coming from19:50
@fungicide:matrix.orgmirror.opensuse as well19:51
@clarkb:matrix.orgthose two seem to be explicitly defined. Here https://opendev.org/openstack/project-config/src/branch/master/grafana/afs.yaml#L513-L528 and here https://opendev.org/openstack/project-config/src/branch/master/grafana/afs.yaml#L462-L47719:53
@clarkb:matrix.orgSo in theory we can drop them from there and grafyaml will update the existing dashboard19:53
@fungicide:matrix.orgoh, yes i just thought to codesearch for mirror.openeuler and i guess it's not actually been deleted from the config19:53
@fungicide:matrix.orgthe others are mirror.centos and mirror.opensuse which are also still there19:54
@fungicide:matrix.orgokay, i'll try to get that trimmed up19:54
@fungicide:matrix.orgfor some reason i expected our grafana configuration to be in opendev/system-config rather than openstack/project-config but i guess we never got around to moving it19:55
@clarkb:matrix.orgya it probably could/should be moved, but hasn't19:56
@clarkb:matrix.orgThis is interesting the gerrit 3.14 relaese notes indicate you need to reindex gerrit changes offline when upgrading to 3.14. But I decided to test it without that step and the logs indicate online reindexing is performed here: https://zuul.opendev.org/t/openstack/build/b21ec122793244c1878129073ef95c2a/log/review99.opendev.org/logs/error_log#14420:42
@clarkb:matrix.orgI guess I need to followup on that in discord20:42
@fungicide:matrix.orgyeah, i wonder if there's something the online reindexing misses?20:45
@clarkb:matrix.orgya or maybe the server isn't fully compatible with version 87 and 88 at the same time?20:46
@clarkb:matrix.orgI've just sent a query on discord and will report back if I learn anything useful20:46
@clarkb:matrix.orguntil we hear back I think we can proceed using the online reindex since the job passes. That means this stack: https://review.opendev.org/c/opendev/system-config/+/992342 should be ready for review and is ready to merge as soon as we're happy to drop 3.12 image builds20:48
@clarkb:matrix.orgI'll be sure to put the grafana upgrade and gerrit upgrade followup stuff on the meeting agenda for tomorow20:48
@clarkb:matrix.orgI'll put that together as soon as the school run is done20:48
@clarkb:matrix.orgThe meeting agenda has my first pass edits in it now: https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Agenda_for_next_meeting please let me know if I'm overlooking anything21:25
@fungicide:matrix.orgthanks! lgtm21:28
@diablo_rojo:matrix.orgHello! Question! So, the FLOSS Mentoring group has reached out to me and asked to retire their ML because they have moved to google forum stuff (*cries*). How is best to do that? Just throw out a patch to delete the ML? or is there a nice way of archiving it? 21:34
@clarkb:matrix.orgI don't know that we've done one of those under mailman3 before. But I think we would pull the list out of our configuration management so that we don't try to manage it anymore then a list admin or a server admin can essentially disable the list and turn it off and put a note that it is there for historical purposes21:42
@clarkb:matrix.orgfungi: probably has thoughts as he has done most of the list management stuff like that in hte past21:42
@diablo_rojo:matrix.orgOkay then I can push up a patch to remove it. 21:44
@fungicide:matrix.orgyes, take it out of our config, then i'll set it to disallow all further posts and hide it from the list index21:45
@fungicide:matrix.orgwe'll keep the archives around for posterity, of course21:45
@fungicide:matrix.orgyou'll still be able to access them from direct links, but they won't be as easily discoverable21:45
@fungicide:matrix.orgalso i'll adjust the list description to note that it's retired, in case anyone directly references the list info page21:46
-@gerrit:opendev.org- Kendall Nelson proposed: [opendev/system-config] 992370: Remove FLOSS Mentoring ML https://review.opendev.org/c/opendev/system-config/+/99237021:47
@diablo_rojo:matrix.orgThat is perfect. 21:47
@diablo_rojo:matrix.orgThank you fungi . Is there anything else I need to do on my end aside from making sure the patch lands? 21:48
@mnasiadka:matrix.orgHello22:00
@mnasiadka:matrix.orgClark: is there anything to do from my side to land the Prometheus patches? After that probably we could try provisioning the VM in some region and try to make it work? I might have some time in the remaining of this week to address some comments if they arise22:02
@clarkb:matrix.orgmnasiadka: I think its mostly about reviews at this point. Then ya we can boot the node and add the deployment jobs (the current chagnes don't have deployment jobs yet which is fine we can add those when we update the inventory)22:05
@clarkb:matrix.orgI'll rereview them now as it looks like there are updates since I last looked22:06
@mnasiadka:matrix.orgThanks22:06
@mnasiadka:matrix.org11pm here, but happy to address anything in my morning22:07
@fungicide:matrix.orgdiablo_rojo: no, nothing else for you, just remind me i said all these things since i have a mind like a steel sieve22:11
@fungicide:matrix.orgjust ping me when you push the patch and i'll review/approve, then take care of the additional steps myself22:12
@fungicide:matrix.orgit's not a thing we do often, so kinda off-the-cuff and i'll figure it out as i go22:12
@clarkb:matrix.orgmnasiadka: ok I +2'd the changes but left some thoughts on how some additional small improvements can be made. I don't think those need to be done in these changes if we prefer using followup changes. I'll defer that decision to you and then do my best to review things however they get pushed22:44
@clarkb:matrix.orgI got a response about Gerrit online vs offline reindexing for the 3.14 upgrade. Sounds like the main consideration there is gerrit performance. It wasn't explicitly stated but seems implied this particular reindexing is computationally expensive so you may not want it to be running while the server is also running. Given that we probably should consider doing the offline reindexing so that it can be completed and over with before starting gerrit back up. But I'll try to learn more about this23:04
-@gerrit:opendev.org- Zuul merged on behalf of Kendall Nelson: [opendev/system-config] 992370: Remove FLOSS Mentoring ML https://review.opendev.org/c/opendev/system-config/+/99237023:04

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