Friday, 2023-01-20

*** bhagyashris is now known as bhagyashris|ruck04:53
marios\o06:37
ysandeephappy friday all o/06:41
*** amoralej|off is now known as amoralej07:52
*** Tengu is now known as Tengu|rover08:12
* Tengu|rover imagines now a small rover trying to make his path on rocky paths08:12
Tengu|roverEarth, here's TripleO Rover - roger08:12
marios:D08:13
Tengu|roverah, we already get the hackmd for this period. thanks to whoever created it :)08:16
Tengu|roverbhagyashris|ruck: we're all good for now? just coming online.08:17
mariosbhagyashris|ruck: Tengu|rover: o/ joining08:30
bhagyashris|ruckjoining08:35
mariosTengu|rover: https://review.rdoproject.org/zuul/buildset/e307ac532f9e43c9a5bd6805a85be0aa train buildset for chasing/checking fails08:40
mariosexample test Tengu|rover https://review.rdoproject.org/r/c/testproject/+/44462 08:40
mariosTengu|rover: promoter logs http://promoter.rdoproject.org/promoter_logs/centos8_train.log08:42
*** ysandeep is now known as ysandeep|afk08:46
*** jpena|off is now known as jpena08:47
mariosTengu|rover: bhagyashris|ruck: https://code.engineering.redhat.com/gerrit/c/tripleo-environments/+/43886508:47
Tengu|rovermarios, bhagyashris|ruck https://review.rdoproject.org/r/c/testproject/+/46686  train re-run08:54
Tengu|roverand now, breakfast.08:55
mariosthx Tengu|rover 08:57
Tengu|roverlet's see how it goes :)09:07
Tengu|roverah, thanks for pushing that directly in the hackmd09:08
*** ysandeep|afk is now known as ysandeep10:05
*** dviroel|out is now known as dviroel10:41
Tengu|roversounds pretty calm today... isn't it?10:48
Tengu|roverjm1: what's the status of your galaxy investigations? did you find any blackholes?10:48
Tengu|roverjm1: also, for the env var: it did work here because I didn't remove the "--no-cache" option... my bad.10:49
jm1Tengu|rover: abandoned my patches because they simply did not work for our use case10:50
Tengu|roverapparently ansible-config doesn't expose anything related to the no_cache. though, maybe, we can try adding it to the configuration file. Still, imho, they are doing it wrong© with the whole API answers.10:50
Tengu|roverjm1: the fact ansible-galaxy servers pushes the FQDN in the answer is so wrong... it breaks the actual use of the "-s <server>" imho.10:50
jm1Tengu|rover: ansible 2.9 does not support installing from git so i dropped that workaround10:51
Tengu|rovererf10:51
Tengu|roverso the "only" solution would be for the proxy to inject the correct name in the JSONs.10:51
jm1Tengu|rover: that sounds like too much work for too little benefit10:52
Tengu|roverapparently not really, from an infra point of view.10:52
Tengu|roverI'll follow the topic with fungi and his team, pretty sure they'll get something at some point.10:52
jm1Tengu|rover: ok cool! fyi only one job in 1 out of 10 patches failed last night. pretty good stats this time :D10:53
Tengu|roverhehj10:53
Tengu|roverso, pretty sure the ansible galaxy cache IS helping10:54
Tengu|roverbut still. imho, the whole thing is broken design.10:54
Tengu|roverthey should get a v3 asap, with proper answers that will NOT expose the fqdn, and make the client re-build the URI using whatever server is configured on their side.10:54
jm1Tengu|rover: the probably have their reasons.... hopefully.. i do not want to give up hope :D10:54
Tengu|roverthat would make the whole thing stronger, better and, well, more proxy-friendly.10:55
Tengu|rovernow, the proxy config itself is a bit flaky, but it's less "broken" than the API design.10:55
jm1Tengu|rover: "you cant fix it all" said some wise guy >> sshnaidm11:00
Tengu|roverjm1: yeah - but we can help nudging things in the right track :)11:01
jm1Tengu|rover: i am trying hard to falsify him but so far i have not been very succesful :D11:01
Tengu|roverjm1: same here. Though.... I tend to push really hard on things that resist :)11:02
Tengu|roverfor the better.... or the worst.11:02
Tengu|roveralready backfired.11:02
Tengu|rover:)11:02
Tengu|roverbut heh. can't change what I am.11:02
*** rlandy|out is now known as rlandy11:04
rlandyTengu|rover: hey - how are you holding up?11:05
rlandybhagyashris|ruck: hello - did you reach out to attila yet?11:06
bhagyashris|ruckrlandy, yes11:06
rlandywhat channel?11:07
rlandyslack?11:07
bhagyashris|ruckon rhos-ops and slack as well no response 11:07
bhagyashris|ruckon slack pinged personally 11:07
rlandybhagyashris|ruck: reached out to him on openstack-pcci11:09
rlandypls join review time11:09
rlandywe can chat afterwards11:10
rlandyotherwise we need to remove these jobs from criteria and promote11:10
rlandyTengu|rover needs the update11:10
bhagyashris|ruckrlandy, sure 11:10
rlandybhagyashris|ruck: can you check components11:12
bhagyashris|ruckAttila haven't joined rhos-ops on slack so pinged him personally on slack11:12
rlandya few were a out yesterday11:12
bhagyashris|ruckyes cheking those11:12
rlandyty11:12
rlandybhagyashris|ruck: and you can add 18 to the list of lines people should watch at this point11:12
bhagyashris|ruckrlandy, you mean in the rr tool11:13
rlandybhagyashris|ruck: in the hack md to start11:13
rlandyyeah and then the rr tool pls11:13
rlandy^^ but that can be its own card11:13
Tengu|roverrlandy: things seem to be stable enough. just launched a train promotion thing, there were 3 failed jobs.11:13
Tengu|roverit's running in testproject. otherwise, things seem pretty green.11:14
rlandyTengu|rover: did you get the alt criteria explanation?11:14
Tengu|roveroh, right, I missed that topic with marios...11:15
Tengu|roverthis morning was a bit.... weird.11:15
Tengu|roverrlandy: so nope11:15
marioso/11:16
Tengu|rovermarios: so far, so good for the jobs. seems the FS issue was indeed a temporary one.11:16
mariosTengu|rover: yes as discussed earlier, we often will run the test to confirm the issue (seen 1x/1x buildset is usually not enough for bug. we need at least 2x for verification -> file bug)11:17
Tengu|roveryup11:17
mariosotherwise we go crazy (er?)11:17
Tengu|roverI guess those don't have the "retry" thingy?11:17
Tengu|roverfor instance I've seen jobs retrying up to 3 times before really failing.11:18
Tengu|rovermaybe something to consider, since it would probably avoid those false-flags?11:18
mariosTengu|rover: that is if they are failing in pre plays usually 11:18
Tengu|roveroh, ok.11:18
ysandeepreviewbot, please add to review list: https://review.opendev.org/c/openstack/tripleo-heat-templates/+/87122311:24
reviewbotI have added your review to the Review list11:24
* soniya29 will be travelling for 1 hr11:26
*** soniya29 is now known as soniya29|afk11:26
Tengu|roverok, #lunch.11:27
ysandeeprlandy, https://review.rdoproject.org/r/c/config/+/4654911:29
rlandyhttps://sf.hosted.upshift.rdu2.redhat.com/zuul/t/tripleo-ci-internal/builds?job_name=periodic-tripleo-ci-centos-8-ovb-3ctlr_1comp-featureset001-internal-train&skip=011:30
* pojadhav stepping out bit11:33
rlandyTengu|rover; alt criteria for train is not working :(11:40
rlandyysandeep and bhagyashris|ruck are looking into that11:40
rlandyyou should have been spared the rerun11:40
rlandybut ok11:40
rlandyTengu|rover: also - asked bhagyashris|ruck to put in a patch to skip the jenkins criteria11:40
ysandeeprlandy, maybe dns/network config issue, holding a c8 train node to check further: https://code.engineering.redhat.com/gerrit/c/testproject/+/211643/178/.zuul.yaml11:41
Tengu|roverrlandy: humpf. seems infra issue (if it's the linked you posted earlier)12:19
Tengu|roverrlandy: is the "alt criteria" related to the jobs running on RH infra when it fails upstream?12:21
rlandyysandeep: ty12:29
rlandyTengu|rover: ack12:29
rlandywill explain at meeting12:29
Tengu|roversure12:29
Tengu|rovermarios: we should get train promotion,12:30
Tengu|roverlast job is finishing12:30
mariosTengu|rover: yeah i think that last might may be failed12:43
mariosi mean from https://review.rdoproject.org/zuul/stream/dc07d6fbb9364e849224479ce3de2798?logfile=console.log 12:44
mariosTengu|rover: can just try with that one if it is yet another not seen before thing12:50
mariosTengu|rover: but got the other 2 which is great12:51
Tengu|rovermarios: yeah, tempest just failed, unreachable instance.13:01
Tengu|rovermarios: so I'll amend my patch and try to get that last one in13:01
Tengu|roverrlandy: oh, damn, we missed the "alt criteria" again....13:01
Tengu|roverpfff.13:01
Tengu|rovermarios: edited. Let's see.13:02
ysandeeprlandy, bhagyashris|ruck I looked at the train job failure using the node which I put on hold, I was NOT able to manually reproduce the issue in the affected node,  but if I comment DNS entries I see the same issue - https://privatebin.corp.redhat.com/?038a3af8af44cb69#5cU5RNMwe37iHjobaV6YJHpZ2NkGnQS1PZur4wZfXHQY , need to check further why dns resolution is not properly working during job run.13:08
*** amoralej is now known as amoralej|lunch13:24
frenzy_friday0/ pls add to your review lists https://review.rdoproject.org/r/c/rdo-infra/ci-config/+/4668713:30
reviewbotI have added your review to the Review list13:30
chandankumarI will be out on monday & tuesday14:04
* pojadhav afk14:04
* bhagyashris|ruck afk14:08
Tengu|roverhumpf.... I think I'll take the opportunity to ride a bit, there's a fine sun outside. -0.5°C, but still, it's doable with proper equipment.14:10
Tengu|rovermarios, rlandy may I let you check on the things for the rest of the day? Things seem pretty calm on the upstream side.14:10
marioso/ Tengu|rover 14:11
Tengu|roverperfect :). Hopefully things won't blow up over the weekend ;).14:11
Tengu|roversee you next week folks! Take care all14:11
*** amoralej|lunch is now known as amoralej14:14
rlandyTengu|rover: sure - have a good weekend14:16
ysandeepTengu|rover, happy weekend o/ 14:20
* marios biab14:38
dasm|offo/14:52
*** dasm|off is now known as dasm14:52
*** dviroel is now known as dviroel|lunch15:49
*** ysandeep is now known as ysandeep|out16:15
* marios off in few 16:31
*** marios is now known as marios|out16:43
*** dviroel|lunch is now known as dviroel16:46
*** amoralej is now known as amoralej|off17:29
*** jpena is now known as jpena|off17:42
* dasm => break20:32
dasmvack21:07
dasmback21:07
* dviroel is having fun, but it is time to start the weekend21:20
dviroelhave a great weekend team21:20
dviroelo/21:20
*** dviroel is now known as dviroel|out21:22
rlandysee you all on monday21:27
dasmsee you!21:28

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!