Thursday, 2021-01-28

openstackgerritMartin Kopec proposed opendev/system-config master: Deploy refstack with ansible docker
clarkbfungi: ok review-test:~clarkb/gerrit-consistency-notes/preferred-email-classifications updated with notes on those other preferred email issues.00:02
clarkbalso high level notes here :
clarkbmy brain is feeling overloaded after all that. I think I'll pick this back up in the morning and start looking at the email conflicts to see if I can classify those further00:21
clarkbanything I can review or look at before I call it a day?00:21
*** tosky has quit IRC00:21
fungimaybe you should just call it a day. i don't think there's anything that can't wait00:22
ianw++ thanks for working through this!00:22
clarkbI'm beginning to think "data scientists" are super human and somehow have the ability to understand lots of characters on a screen00:23
clarkbat least for me I have a hard time writing any automation around this stuff iwthout first "seeing" it00:23
clarkbfungi: oh another thing I realized was that I bet (and I can test this) that we don't need to fix everything in one commit as long as we push a series of commits that results in a happy spot00:27
clarkbmy hunch is that gerrit is only verifying the end state result (though it could be validating each commit)00:27
clarkband even if it doesn't we can curate it that way then squash everything down if necessary00:28
fungioh, yeah worth a shot00:30
fungimight at least make them easier to review00:30
*** mlavalle has quit IRC00:57
*** DSpider has quit IRC01:05
*** priteau has quit IRC01:16
ianwwhen i was at uni, there were stories about how people hadn't left enough digits for the DJI to go above 10,000 in their various code01:27
ianwi wonder how many people are going to be caught out when GME breaks through the 10,000 mark :)01:27
*** hamalq has quit IRC02:10
kevinzianw: there?02:20
*** zbr7 has joined #opendev02:20
ianwkevinz: hey, yep02:20
kevinzHi ianw: I find the gitweb link at opendev gerrit is 404 now.02:21
kevinzDo you know what happened to it?02:21
*** portdirect has quit IRC02:21
*** mattmceuen has quit IRC02:21
kevinzFor example,, click the "Links gitweb", then redirect to 40402:22
fungiit hasn't worked since the upgrade, we need to work out how to properly link those to gitea ideally... i think mordred had a change up for that from some time back but it likely needs updating02:22
ianwyeah that is part of the gerrit upgrade, and i think we had a change somewhere to turn it into a gitea link02:22
*** mattmceuen has joined #opendev02:22
*** zbr has quit IRC02:22
*** zbr7 is now known as zbr02:22
*** portdirect has joined #opendev02:22
ianwjinx w/ fungi :)02:22
kevinzOK, thanks for that info02:23
ianw is the change in questino02:25
mordredYeah, that's the one02:52
openstackgerritIan Wienand proposed opendev/system-config master: [dnm] fail to get a gerrit for testing pluging observer changes
*** gouthamr has quit IRC04:18
*** ykarel has joined #opendev04:40
*** whoami-rajat__ has joined #opendev04:59
*** ykarel_ has joined #opendev05:49
*** ykarel has quit IRC05:52
*** ykarel_ is now known as ykarel05:53
*** marios has joined #opendev06:14
*** hemanth_n has joined #opendev06:42
*** zbr has quit IRC06:44
*** zbr has joined #opendev06:46
*** walshh__ has joined #opendev06:49
*** walshh_ has quit IRC06:49
*** walshh__ is now known as walshh_06:49
*** redrobot1 has joined #opendev07:00
*** redrobot has quit IRC07:04
*** redrobot1 is now known as redrobot07:04
*** eolivare has joined #opendev07:43
*** akahat|rover is now known as akahat|lunch07:51
*** jpena|off is now known as jpena07:55
*** slaweq has joined #opendev07:58
*** ralonsoh has joined #opendev08:01
*** jpena is now known as jpena|off08:01
*** jpena|off is now known as jpena08:01
*** andrewbonney has joined #opendev08:14
*** rpittau|afk is now known as rpittau08:27
*** sboyron has joined #opendev08:40
*** tosky has joined #opendev08:45
*** hashar has joined #opendev08:48
*** ykarel is now known as ykarel|lunch08:51
*** tkajinam has quit IRC08:58
*** klonn has joined #opendev09:22
*** ykarel|lunch is now known as ykarel10:03
*** osmanlicilegi has quit IRC10:27
openstackgerritMartin Kopec proposed opendev/system-config master: Deploy refstack with ansible docker
*** akahat|lunch is now known as akahat|rover10:44
*** osmanlicilegi has joined #opendev10:52
*** dtantsur|afk is now known as dtantsur10:55
*** osmanlicilegi has quit IRC11:01
*** calcmandan has quit IRC11:01
*** calcmandan has joined #opendev11:01
*** osmanlicilegi has joined #opendev11:06
*** ttx has quit IRC11:22
*** ttx has joined #opendev11:23
*** osmanlicilegi has quit IRC11:58
*** klonn has quit IRC12:11
*** dviroel_ is now known as dviroel12:16
*** ysandeep is now known as ysandeep|afk12:25
*** DSpider has joined #opendev12:26
*** hemanth_n has quit IRC12:26
*** jpena is now known as jpena|lunch12:31
*** priteau has joined #opendev12:41
*** klonn has joined #opendev12:43
*** osmanlicilegi has joined #opendev12:56
*** ysandeep|afk is now known as ysandeep13:21
*** DSpider has quit IRC13:24
*** jpena|lunch is now known as jpena13:24
*** redrobot5 has joined #opendev13:44
*** parallax_ has joined #opendev13:45
*** parallax has quit IRC13:45
*** redrobot has quit IRC13:45
*** tosky has quit IRC13:45
*** redrobot5 is now known as redrobot13:45
*** parallax_ is now known as parallax13:45
*** tosky_ has joined #opendev13:46
yoctozeptomorning infra; were there any answers to the tact call for help?14:02
*** DSpider has joined #opendev14:27
fungiyoctozepto: gmann offered to help out before we sent it, but no offers from anyone new yet since sending it to the ml14:29
fungii have a feeling it'll percolate in people's minds and then they may start getting slowly involved14:29
fungiperhaps not formally at first, but that's fine14:29
fungimy ulterior motive with that post was to remind everyone that we all share responsibility for these tasks14:30
yoctozeptofungi: ack, I hope you are right, fingers crossed for TACTful future :D14:31
ykarelreviews please
*** redrobot has quit IRC14:44
*** ysandeep is now known as ysandeep|away14:48
openstackgerritMarcin Juszkiewicz proposed openstack/project-config master: CentOS 8 Stream initial enablement for AArch64
*** redrobot has joined #opendev14:57
*** sshnaidm|ruck is now known as sshnaidm|afk15:01
*** marios is now known as marios|call15:02
openstackgerritMerged zuul/zuul-jobs master: Fix repo files for CentOS Stream
*** gouthamr has joined #opendev15:14
*** redrobot has quit IRC15:27
*** klonn has quit IRC15:29
*** redrobot has joined #opendev15:34
*** mlavalle has joined #opendev15:44
*** klonn has joined #opendev15:48
*** sboyron has quit IRC15:49
openstackgerritMarcin Juszkiewicz proposed openstack/project-config master: CentOS 8 Stream initial enablement for AArch64
*** tosky_ is now known as tosky16:06
*** mgagne has joined #opendev16:11
*** ykarel has quit IRC16:16
openstackgerritRiccardo Pittau proposed openstack/project-config master: Remove anachronistic job from scciclient
*** sboyron has joined #opendev16:34
openstackgerritClark Boylan proposed opendev/system-config master: Use patch instead of sed to update Gerrit's x/ route
*** sshnaidm|afk is now known as sshnaidm|ruck16:58
*** jpena is now known as jpena|off17:02
*** marios|call is now known as marios17:03
*** tosky has quit IRC17:07
*** sboyron_ has joined #opendev17:07
*** tosky has joined #opendev17:07
*** sboyron has quit IRC17:07
*** marios is now known as marios|out17:49
clarkbthe above change is looking good so far in zuul. I'll update review-test with it to double check the sed to patch conversion still does what we want17:52
clarkbjsut need jobs to finish up first17:52
*** eolivare has quit IRC17:54
*** rpittau is now known as rpittau|afk17:57
openstackgerritClark Boylan proposed opendev/system-config master: Build Gerrit 3.3 images
openstackgerritClark Boylan proposed opendev/system-config master: Update Gerrit 3.2 plugin versions on image builds
clarkbthose two changes occured to me while doing the x/ thing. I'll actually test the result of 772931 instead of 77792318:20
fungiahh, yep18:23
*** dtantsur is now known as dtantsur|afk18:25
*** eolivare has joined #opendev18:25
*** andrewbonney has quit IRC18:26
sshnaidm|ruckis there an option not to run specific jobs/queue/projects on a specific provider cloud?18:30
sshnaidm|ruckfor example if I want to avoid running jobs on vexxhost ca-ymq-118:30
fungisshnaidm|ruck: we don't have any means of expressing that, but also mnaser might be interested in addressing whatever makes your jobs not work in vexxhost ca-ymq-118:32
sshnaidm|ruckfungi, yep, we'll talk to him18:32
openstackgerritGhanshyam proposed openstack/project-config master: Remove Karbor projects from infra
*** eolivare has quit IRC18:42
clarkbok I've got builds of that image now to test19:07
clarkbproceeding with that momentarily19:07
*** diablo_rojo has joined #opendev19:11
diablo_rojoI set up an etherpad for tracking the channel renames fungi:
diablo_rojoNot sure what else to include at this point.19:11
clarkbok review-test has been updated. I am able to clone x/ansible-role-jenkins from it19:12
clarkbI think and are good to review19:13
*** marios|out has quit IRC19:13
kopecmartinclarkb: hi, it passed! \o/ , what now?19:16
*** portdirect has quit IRC19:17
clarkbkopecmartin: next up we need to actually deploy a new server with it. The best thing for that may be to add it to our weekly meeting agenda and we can find volunteers to help and/or send email about it to service-discuss@lists.opendev.org19:17
*** portdirect has joined #opendev19:17
fungidiablo_rojo: ooh! i'll open a browser tab and... something, something, profit? but seriously i'll try to take a look, thanks for driving this19:17
clarkbkopecmartin: probably what we'll do is deploy it but not really expose it first, check it looks good, then schedule a downtime to cut over production to the new instance19:17
diablo_rojofungi, no problem, maybe later today or tomorrow we can divvy up tasks or whatever.19:18
fungiyep, absotumitably19:19
kopecmartinclarkb: ok, sign me up, i'll help with testing19:20
clarkbkopecmartin: can you send the email and/or update the meeting agenda with info on the change and asking for help to get it deployed?19:21
clarkbthere is a good chance I'll end up helping but getting it to a wider audience first is nice :)19:21
kopecmartinclarkb: sure, can you please point me to the agenda?19:22
kopecmartinclarkb: thanks19:23
*** d34dh0r53 has quit IRC19:25
*** d34dh0r53 has joined #opendev19:30
openstackgerritRadosÅ‚aw Piliszek proposed opendev/system-config master: ask.o.o: Fix two issues with ro message
rm_workhey, have been on a bit of a hiatus and came back to this new gerrit interface... hoping someone here has some good greasemonkey scripts to bring back something more usable?19:53
rm_workIE, something akin to the old greasemonkey script that gave us realtime zuul statuses (which seem to now be missing altogether?!)19:54
clarkbrm_work: the zuul status are there as a static table which is how we've always shipped them19:55
clarkbthere is a openstack-discuss thread on that subject though and you may find useful info there19:55
rm_workit seems to be completely missing from the new gerrit interface19:56
clarkb"Zuul Summary"19:56
clarkb^F that string will get you to it19:56
rm_workah it's tabbed, doesn't just show up19:56
rm_worki couldn't even tell those were clickable19:56
rm_workyeah ok, can probably just make a greasemonkey script to expose that at a higher level, will check the ML first tho to see if someone already fixed it, thanks :)19:57
fungirm_work: if you're looking for the status for in-progress builds for a change, that's what's currently being discussed yeah. there's an available gerrit plugin which we may be able to use or at least adapt to work in our deployment19:58
*** whoami-rajat__ has quit IRC19:58
fungithe current zuul report tab you're seeing is just for completed buidls19:59
rm_workright, but previously it was higher level, not buried19:59
rm_workand yeah there was that great greasemonkey script before that had it fully realtime, was amazingly useful19:59
fungiwell, we were adding it with a hacky custom javascript overlay which depended on implementation details of their ui20:00
rm_workright, and it worked brilliantly :D20:00
fungiit wasn't really part of gerrit20:00
fungiand the ui uses a totally different framework now, but is also pluggable, so we can actually upstream our plugins20:00
rm_workyeah, would love to see it back at top-level20:01
rm_workof course I also wish they hadn't removed all of the lines/borders/colors such that everything all just blends into a giant mess, but i guess eventually i'll figure out how to parse this20:01
rm_workin the meantime I bet I can fix it with some custom CSS :D20:01
*** d34dh0r53 has quit IRC20:04
clarkbif you're really interested in helping with upstream gerrit ui design they have a thread on their repo-discuss mailing list soliciting ui feedback nits and such20:05
openstackgerritGuillaume Chauvel proposed opendev/system-config master: Increase autogenerated comment width to avoid line wrap
openstackgerritGuillaume Chauvel proposed opendev/system-config master: [DNM] test comment width: review without autogenerated tag
clarkbI think it is worth pointing out that a ton of effort has gone into this and while it isn't perfect it is a compromise for the better. We have asked for help pretty regularly since like september? This shouldn't be a surprise and the best way to make things better is to help20:08
rm_workright like I said, I've just come back from hiatus20:09
rm_workI think the last time I really did a review was ... october20:09
rm_workgetting up to speed20:10
clarkbrm_work: that is the upstream thread if you are interested20:11
rm_worki am very interested, thanks20:12
rm_workother than the fact that I HATE mailing lists T_T20:12
fungihopefully once you gain some familiarity with the new gerrit interface, you'll find it an improvement over the previous one, even if there are still some rough edges and missing feature parity with what we had (still trying to close the gap on some of it)20:12
clarkbya there are a lot of new great features like the single page diff viewing of all files20:13
fungithe easiest way to deal with mailing lists is to use an mua which will let you just delete everything from the list ;)20:13
fungithen read from the archives if someone points you to a message which was important20:13
clarkbits fairly low volume on that one (maybe one or two messages a day max). But I think google groups also allows you to just respond directly via the web view20:14
clarkb(I've never done that though)20:14
openstackgerritMerged opendev/system-config master: ask.o.o: Fix two issues with ro message
*** d34dh0r53 has joined #opendev20:15
rm_workit supports themes, right?20:15
rm_workbut I guess they have to be installed on the server side?20:15
fungiyup, there are a couple already20:15
fungiif you go into your preferences there's a "dark mode" theme you can switch to20:16
rm_workhonestly i think it's entirely just theming issues for me20:16
fungialso theming is pluggable, that's how we do the opendevification for our deployment. example here:
rm_workyeah, i guess for now I'll see if the existing greasemonkey script can be easily adapted20:18
*** ralonsoh has quit IRC20:48
ianwspeaking of, can some people click around on the two changes @ and make sure the summary tab is updating for them21:19
ianwthe change is rather simple; but just want to make sure it works for !me too :)21:20
corvusianw: when i click between changes, it goes back to the files tab, then i click on zuul summary again, and the contents are different for the 2 different changes.  iiuc that is expected behavior and your patch is gtg22:26
ianwcorvus: thanks, yep that's expected behaviour22:26
corvusianw: i left a V+1 based on that22:27
*** klonn has quit IRC22:47
*** sboyron_ has quit IRC23:05
*** slaweq has quit IRC23:11
*** mlavalle has quit IRC23:17
clarkbianw: do the borg things need reviews or were we still waiting on a new ps?23:19
ianwclarkb: i got a bit distracted, again, sorry23:20
clarkbk no worries23:20
* clarkb has been finding distractions too23:20
ianwi think that ask.o.o is one of the bigger users23:20
clarkbhuh it should be fairly static at this point?23:20
ianwyeah, rolling over and gzipping the db , it must have some sort of timestamps that change it just enough23:21
clarkband then gzip compounds that "makes sense"23:21
ianwi'm thinking that maybe we just take one manual backup snapshot of it and then disable further periodic backups23:21
fungii suppose with it being in read-only mode we could stop the db dumps yeah23:22
*** diablo_rojo has quit IRC23:22
fungior end backups for it entirely, sure23:22
clarkbya I think that is a reasonable next step23:22
fungiserver snapshot seems plenty sufficient to preserve it until we turn it off23:22
ianwi will get to mordred's comment on to fix up the scripts to make even better raw dump outputs23:22
ianwmaybe if you want to review that WRT having the raw streams, that would be good23:23
ianwwasn't 100% clear if we're all on board with that approach23:23
clarkbI think I'm willing to accept it :)23:23
clarkbhaving docs about it too would be good23:23
clarkbso that we don't confuse ourselves in the future when we wonder why it is split that way23:24
ianwsure, i added some stuff to the readme for the roles, but can also expand in our docs/*23:24
clarkbI think docs was the only major thing missing after my previous look at it23:25
clarkboh and do we need to squash the child into the parent?23:26
clarkbto avoid issues with pruning once that change lands if there is a gap?23:26
clarkb(also it isn't clear to me how we migrate the existing backups to the newly named backps or maybe we don/t)23:27
clarkbso ya mostly just docs :)23:27
ianwyeah, i can, i was expecting to merge them both close, but can squash if we like23:27
ianwyeah, it's going to be a bit manual swapping over, i've still got to clear space on vexxhost23:27
clarkbif we don't squash them I think we should land them together23:27
clarkbianw: also should we set a reserved space in the borg configs?23:27
ianwi feel like 1tb should be enough for our working set of backups, that's why i'm trying to get us within that23:28
clarkbthis will keep it from completely filling the disk23:28
ianwwe can, i've not played with that option.  i guess the backup fails one way or the other23:28
clarkbyup it will still fail but ensures you can do other operations I think23:29
ianwsorry i've just freaking notcied that the zuul summary fix doesn't seem to work on chrome23:34
ianwof course it gives no error or other indication, other than not being there :(23:35
clarkbmmm browsers23:35
clarkbunfortunately I'm not a ton of help in that arean23:35
*** knikolla has quit IRC23:36
*** knikolla has joined #opendev23:36
*** ildikov has quit IRC23:36
*** zaro_ has joined #opendev23:37
*** dviroel_ has joined #opendev23:37
*** aprice has quit IRC23:37
*** zaro has quit IRC23:37
*** aprice has joined #opendev23:37
*** dviroel has quit IRC23:37
*** dviroel_ is now known as dviroel23:37
*** zaro_ is now known as zaro23:37
*** ildikov has joined #opendev23:38
ianwnow it does, and i did nothing but copy the file?23:40
ianwmaybe it's a caching thing23:40
ianwif someone else with google chrome could validate that would be great23:41
ianwit seems to work from my personal laptop & phone23:41
mordredchrome won't let me load that23:42
clarkbianw: what do you want me to verify there?23:42
mordredbecause of certs23:42
ianwclarkb: just that the zuul summary tab is there and looks like it's working23:42
clarkbianw: it does on chrome for me23:42
mordredianw: fwiw:
mordredis the error I get23:43
mordredand chrome will not let me say "yeah, I don't care"23:43
ianwmordred: interesting, i got told to "click here to enable enhanced protection" but was allowed to accept the self-signed23:44
clarkbya I could accept the risk in chrome too23:44
mordredianw: I think I updated chrome a couple of days ago23:44
mordredso - maybe I'm a glimpse of the future?23:45
mordredoh wow23:45
ianwclarkb: thanks, i don't think it's anything related to javascript then.  when i re-copied the file it popped back up, so i guess it must have been something to do with serving it, being stale or something23:45
mordredit's a mac thing23:45
ianwmy phone is 88.0.4324.9323:46
mordredif I click on the warning page and type "thisisunsafe" it proceeds23:46
mordredI see the zuul summary tab23:46
clarkb"its a mac thing" <- I feel left out :P23:46
mordredand clicking through the changes the zuul summary tabs change23:47
ianwIt's a UNIX system!  I know this!23:47
ianwexcept these days i think it's less and less a unix system23:47
mordredyah. srrsly23:47
mordredof course, I could say the same of linux too ...23:48
ianwthanks for checking it.  i will presume nobody has an issue with the row highlight or left padding i added23:48
mordredianw: I think your upstream patch is ready to land, yes?23:48
ianwmordred: yeah, i stacked a couple of UI bits on too if you'd like to look23:48
clarkbit looked fine to me23:49
clarkbbut I tend to not be super particular23:49
mordredthose are all applied to the test instance, right?23:50
ianwthere was a proposal to shrink the columns, but i wanted to make sure across different CI's (which are separate tables) the columns remained a consistent width23:50
ianwi.e. all the times line up23:50
mordredianw: I landed the first one, and have +2+V'd the rest of them23:50
ianwi think the row highlight makes it easier to pick out the info, and is consistent with the file picker, which is 100% and uses a row highlight23:51
ianwthanks; i want to update the zuul job to add some more changes to the test gerrit, it will mean next time to debug similar you won't have to manually inject sample changes23:52
clarkbianw: one thing that occured to me is that it might be good to have a x/foo repo in the test and have a test that clones it (to confirm our workaround for that continues to function)23:53
clarkbianw: is there already code there to create a random repo with some content? or would I need to build that up?23:53
clarkbparticularly as the 3.3 image spinup happens I think confirming that for 3.3 too would be good23:54
ianwclarkb: yeah, i think all that needs is a x/foo project create like !
ianw@ even23:58
clarkbexcellent, I'll take a look at that tomorrow as I think that will help avoid us regressing in that spot23:59

Generated by 2.17.2 by Marius Gedminas - find it at!