Wednesday, 2020-07-01

« Tuesday, 2020-06-30 Index Thursday, 2020-07-02 »
*** shanemcd has quit IRC00:02
*** shanemcd has joined #zuul00:04
ianwfungi / tristanC: i'm just reading backwards -- if we're dropping nodes, we should certainly setup a netconsole.  i've had that catch stuff that doesn't come out on the console00:55
ianwthe question is, did i port all the setup stuff out of devstack-gate00:56
ianwyes i did; https://zuul-ci.org/docs/zuul-jobs/general-roles.html#role-enable-netconsole ... i thought so00:57
*** rlandy|afk has quit IRC01:37
*** swest has quit IRC01:56
fungiianw: though in this case we were able to get the nova console after the node became unreachable02:06
fungiunfortunately it was not remotely enlightening02:07
*** swest has joined #zuul02:10
ianwfungi: yeah, iirc with the last time i used that, we didn't get anything on the serial console even, but did with that capturing the oops02:15
*** bhavikdbavishi has joined #zuul02:41
*** bhavikdbavishi1 has joined #zuul02:46
*** bhavikdbavishi has quit IRC02:48
*** bhavikdbavishi1 is now known as bhavikdbavishi02:48
*** sgw1 has quit IRC02:49
*** wuchunyang has joined #zuul02:53
*** wuchunyang has quit IRC02:58
*** rfolco has quit IRC03:20
*** sgw1 has joined #zuul03:27
*** bhavikdbavishi has quit IRC03:44
*** bhavikdbavishi has joined #zuul03:45
*** iurygregory has quit IRC04:31
*** evrardjp has quit IRC04:33
*** evrardjp has joined #zuul04:33
*** wuchunyang has joined #zuul04:37
*** sgw1 has quit IRC04:40
*** mugsie has quit IRC04:53
*** wuchunyang has quit IRC04:55
*** sgw has quit IRC04:55
*** wuchunyang has joined #zuul04:56
*** wuchunyang has quit IRC04:57
*** mugsie has joined #zuul04:57
*** ysandeep|away is now known as ysandeep05:09
*** sgw has joined #zuul05:14
*** marios has joined #zuul05:16
*** vishalmanchanda has joined #zuul05:23
*** wuchunyang has joined #zuul05:28
*** wuchunyang has quit IRC05:35
openstackgerritSimon Westphahl proposed zuul/zuul master: Ensure refs for recent branches are not GCed  https://review.opendev.org/73845405:36
*** bhavikdbavishi has quit IRC05:49
*** bhavikdbavishi has joined #zuul06:05
*** bhavikdbavishi1 has joined #zuul06:08
*** bhavikdbavishi has quit IRC06:10
*** bhavikdbavishi1 is now known as bhavikdbavishi06:10
*** bhavikdbavishi1 has joined #zuul06:35
*** bhavikdbavishi has quit IRC06:37
*** bhavikdbavishi1 is now known as bhavikdbavishi06:37
*** bhagyashris is now known as bhagyashris|brb06:59
*** hashar has joined #zuul07:09
*** iurygregory has joined #zuul07:09
*** bhavikdbavishi has quit IRC07:09
*** jcapitao has joined #zuul07:13
*** bhagyashris|brb is now known as bhagyashris07:15
swestianw: could you please have another look at https://review.opendev.org/#/c/728824/07:22
swestzuul-maint: anyone up for a second review of https://review.opendev.org/#/c/725810/ ?07:25
*** bhavikdbavishi has joined #zuul07:29
*** saneax has quit IRC07:32
*** tosky has joined #zuul07:40
*** jpena|off is now known as jpena07:58
*** nils has joined #zuul08:01
*** saneax has joined #zuul08:16
*** hashar has quit IRC08:17
*** hashar has joined #zuul08:18
*** ysandeep is now known as ysandeep|lunch08:21
*** fbo has joined #zuul08:35
*** newbie2020 has joined #zuul08:49
newbie2020Hi guys,08:49
newbie2020I have a couple of questions08:49
openstackgerritMatthieu Huin proposed zuul/zuul master: Web UI: add i18n support, french translation  https://review.opendev.org/73729008:50
newbie20201. Is there a way to make zuul merge multiple PRs that have a Depends-On relationship?08:50
*** bhavikdbavishi has quit IRC08:51
newbie2020PR#1 has changes affecting a job running in the post (in this tenant all CI stuff is in the same repo)08:51
newbie2020PR#2 contains some code changes and again is affecting the result of the job in the post I mentioned earlier08:52
newbie2020PR#2 Depends-On PR#108:52
newbie2020Essentially the post jobs passes only if both PRs are merged08:52
*** bhavikdbavishi has joined #zuul08:56
*** ysandeep|lunch is now known as ysandeep09:21
*** hashar has quit IRC09:25
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877109:34
*** hashar has joined #zuul09:42
openstackgerritSimon Westphahl proposed zuul/zuul master: Expire Github installation key 3 minutes before  https://review.opendev.org/73877209:42
*** hashar is now known as hasharAway09:47
*** newbie2020 has quit IRC09:52
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877109:53
*** hasharAway has quit IRC09:57
*** hashar has joined #zuul09:58
*** bhavikdbavishi has quit IRC10:07
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877110:13
*** bhavikdbavishi has joined #zuul10:15
*** bhavikdbavishi has quit IRC10:26
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877110:26
*** ysandeep is now known as ysandeep|brb10:27
*** holser has quit IRC10:36
*** ysandeep|brb is now known as ysandeep10:37
*** holser has joined #zuul10:37
*** bhavikdbavishi has joined #zuul10:51
*** bhavikdbavishi1 has joined #zuul10:54
*** bhavikdbavishi has quit IRC10:56
*** bhavikdbavishi1 is now known as bhavikdbavishi10:56
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877111:00
*** jcapitao is now known as jcapitao_lunch11:10
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877111:12
*** sshnaidm|afk is now known as sshnaidm|ruck11:19
openstackgerritAlbin Vass proposed zuul/zuul-jobs master: Allow deleting workspace after running terraform destroy  https://review.opendev.org/73877111:25
*** jpena is now known as jpena|lunch11:29
*** dcastellani_ has joined #zuul11:57
*** dcastellani has quit IRC11:58
*** dcastellani_ is now known as dcastellani11:58
*** jcapitao_lunch is now known as jcapitao12:07
*** irclogbot_1 has quit IRC12:16
*** irclogbot_0 has joined #zuul12:18
*** Miouge- has quit IRC12:20
*** Miouge has joined #zuul12:23
*** jpena|lunch is now known as jpena12:23
*** rlandy has joined #zuul12:33
*** bhavikdbavishi has quit IRC12:36
*** bhavikdbavishi has joined #zuul12:36
*** bhavikdbavishi has quit IRC12:41
*** ysandeep is now known as ysandeep|afk12:45
*** sgw1 has joined #zuul12:55
*** rlandy is now known as rlandy|training12:57
clarkbnewbie2020 is gone but I assume "post" refers to a post merge pipeline. If both repos share a pipeline queue then they'll be enqueued together and landed in sequence for you. Then the post jobs should run in sequence (the first one may fail)12:59
clarkb* share a gate pipeline queue13:00
clarkbif they don't share that gate pipeline queue then you can approve them one after another and land them in sequence that way and again post would fail on the first one likely13:00
*** rfolco has joined #zuul13:00
clarkbif the goal here is to run only the second post job or somehow prevent the first from failing you may need the circular deps setup along with some variation of a supercedent post pipeline that will delay before deciding what the best candidate to run is?13:01
*** jhesketh has quit IRC13:22
*** jhesketh has joined #zuul13:24
*** ysandeep|afk is now known as ysandeep13:31
*** bhagyashris is now known as bhagyashris|afk13:42
*** newbie2020 has joined #zuul14:03
*** hashar has quit IRC14:06
zbrhow do I run a tox job what does not run revoke-sudo role?14:09
zbri have some molecule jobs that are implemented using tox, but they need to be able to sudo during run.14:10
fungizbr: there's a job for that (run-tox-without-sudo i think?)14:11
zbri guess the only way to do it is to override run.yaml14:11
fungis/without/with/14:12
fungitrying to find it now14:13
AJaegerfungi, zbr, openstack-tox-with-sudo14:16
zbrthat is why i did not found it in zuul-jobs, anyway, the run of tox is very basic and easy to override.14:17
fungiAJaeger: yep, just found it myself too. looks like that's something we're doing downstream in opendev for the openstack tenant14:18
fungihttps://opendev.org/openstack/openstack-zuul-jobs/src/branch/master/zuul.d/jobs.yaml#L84-L9014:18
fungibasically just substitute your own run.yaml like https://opendev.org/openstack/openstack-zuul-jobs/src/branch/master/playbooks/tox-with-sudo/run.yaml14:21
fungicompare with https://opendev.org/zuul/zuul-jobs/src/branch/master/playbooks/tox/run.yaml14:21
zbryep, exacly like mine now, thanks.14:30
*** ysandeep is now known as ysandeep|away15:15
*** bhavikdbavishi has joined #zuul15:31
*** sshnaidm|ruck is now known as sshnaidm|afk15:32
*** bhavikdbavishi1 has joined #zuul15:35
*** bhavikdbavishi has quit IRC15:36
*** bhavikdbavishi1 is now known as bhavikdbavishi15:36
*** newbie2020 has quit IRC15:40
*** hamalq has joined #zuul15:57
*** hamalq_ has joined #zuul15:58
*** sshnaidm|afk is now known as sshnaidm|ruck16:01
*** hamalq has quit IRC16:02
*** marios is now known as marios|out16:07
*** etp has quit IRC16:16
*** bhavikdbavishi has quit IRC16:16
*** etp has joined #zuul16:19
*** sshnaidm|ruck is now known as sshnaidm|afk16:20
*** jcapitao has quit IRC16:23
*** etp has quit IRC16:24
*** hashar has joined #zuul16:29
corvusavass: how does https://review.opendev.org/738517 look?  do you like the solution?  should we squash it with your change?16:37
mordredcorvus: wow, "name: "ensure-{{ container_command }}"" is a thing I'd never considered doing but which obviously makes sense16:41
corvusmordred:  :)  i wish task names were templated :/16:41
corvusi keep wanting to do "task: name: Intalle {{ something }}" and i'm sad every time i try16:42
mordredyeah16:42
*** etp has joined #zuul16:44
mordredcorvus: I think that approach looks great16:49
*** jpena is now known as jpena|off16:53
*** hashar has quit IRC16:58
*** nils has quit IRC17:04
*** marios|out has quit IRC17:05
avasscorvus: can't you template task names?17:14
*** samccann has quit IRC17:15
*** corvus has quit IRC17:15
*** bolg has quit IRC17:15
*** guillaumec has quit IRC17:15
*** andreykurilin has quit IRC17:15
*** freefood has quit IRC17:15
*** mmedvede has quit IRC17:15
*** gothicmindfood has quit IRC17:15
avassI'm absolutely sure you can do that, but I don't really like it since it gets harder to grep for it17:16
*** hashar has joined #zuul17:17
*** samccann has joined #zuul17:18
*** corvus has joined #zuul17:18
*** guillaumec has joined #zuul17:18
*** mmedvede has joined #zuul17:18
*** freefood has joined #zuul17:18
*** Miouge has quit IRC17:18
*** andreykurilin has joined #zuul17:19
*** holser has quit IRC17:19
*** harrymichal has joined #zuul17:20
*** Miouge has joined #zuul17:22
*** holser has joined #zuul17:23
avasscorvus: I think that looks good17:23
*** johanssone has quit IRC17:26
*** weshay_ruck has quit IRC17:26
*** cloudnull has quit IRC17:26
*** AJaeger has quit IRC17:26
*** irclogbot_0 has quit IRC17:28
*** johanssone has joined #zuul17:29
*** weshay_ruck has joined #zuul17:29
*** cloudnull has joined #zuul17:29
*** AJaeger has joined #zuul17:29
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Test multiarch release builds and use temp registry with buildx  https://review.opendev.org/73731517:30
*** irclogbot_1 has joined #zuul17:31
corvusavass, mordred: that's the patch squashed into the parent17:31
corvusthat way we don't have to worry about hitting the ipv6 issue in gate17:31
avassI believe there could still be a problem with buildx not having the certificate installed if someone uses a private publication registry with buildx, but I guess that could be solved later.17:33
avassjust pointing that out17:34
avasslgtm otherwise17:35
avassI'll be back in about ~1h if there's anything else17:36
*** bolg has joined #zuul17:40
corvusavass: yes, but i think we should expect them to set up what they need to access their private registry outside of that role17:45
*** hashar is now known as hasharAway17:47
openstackgerritMerged zuul/zuul-jobs master: Test multiarch release builds and use temp registry with buildx  https://review.opendev.org/73731518:19
*** hasharAway is now known as hashar18:25
*** chandankumar is now known as raukadah18:39
openstackgerritMerged zuul/zuul-jobs master: ensure-pip debian: update package lists  https://review.opendev.org/73752918:39
corvusi've re-enqueued the nodepool tag18:43
avasscorvus: yeah but the problem was that the certificate needs to be installed inside the buildx container18:52
avassotherwise I would agree with you18:52
corvusavass: that's fair; i do think it's a significant scope expansion for the role, but a reasonable one18:55
corvusshouldn't be too hard to do if someone has a need18:55
avassyep that's why I think it's too important18:55
avass*not too important18:55
avassbut might be good to know about18:55
corvusavass, mordred: !!!!! https://hub.docker.com/r/zuul/nodepool-launcher/tags  !!!!18:56
clarkbdid we figure out what the arch confusion in the manfiest was?18:56
*** harrymichal has quit IRC18:57
corvusclarkb: no idea.  istr someone thought the rework that avass and i were doing on it might just solve the problem.18:57
clarkbcool18:57
corvusif that issue is written down anywhere, or someone knows how to reproduce it, i'd be happy to check / look into it18:57
corvusi'm curious why nodepool's 'latest' image only has amd6418:58
avasscorvus, clarkb: my guess is that it has something to do with the docker cache not being able to handle manifest lists18:58
clarkbcorvus: I think reproducing it was on mordred's todo list. If you're able to pull nodepool 3.13.0 and run it on your x86 machine and it isn't an arm64 image its probably good enough for now18:58
avassoh wait, but it's just pushed from buildx, so that's probably not it18:59
corvusclarkb: i just ran zuul/nodepool-launcher:3 on my x86 and it runs and there are no qemu processes running19:00
clarkbnice19:00
corvusclarkb: i know that we tested that a lot on the build node (pulling from the temp registry) and always got the right one19:00
clarkbya and we saw it break on the test nodes iirc19:00
clarkbso if we didn't see that their either chances are we're good19:01
corvusso yeah, if the issue was the pull/push to docker image cache, then that's completely reworked, so fix makes sense19:01
corvusanyone know why latest is only one arch?19:01
avasscorvus: updated 2 days ago19:02
corvusavass: i guess i'll go dig for a build log19:03
openstackgerritJames E. Blair proposed zuul/nodepool master: Restore release jobs  https://review.opendev.org/73891419:03
avasscorvus: maybe buildx needs to tag 'latest' as well?19:03
avasscorvus: but the tag should not update the latest build anyway should it?19:04
*** saneax has quit IRC19:33
*** yoctozepto7 has joined #zuul19:37
*** yoctozepto has quit IRC19:45
*** yoctozepto7 is now known as yoctozepto19:45
*** vishalmanchanda has quit IRC19:47
*** rlandy|training is now known as rlandy19:59
corvusthe upload job looks like it should have worked; i wonder if promote can't handle multi arch20:29
corvusclarkb, fungi: can you go ahead and push https://review.opendev.org/738914 through?  i'd like to have a more current run to look at20:31
clarkblooking20:31
*** sgw1 has quit IRC20:35
corvusi'm gonna go ahead and +w that20:43
*** sgw has quit IRC20:57
clarkbya its a simple change and return to situation normal21:00
*** sgw has joined #zuul21:35
openstackgerritMerged zuul/nodepool master: Restore release jobs  https://review.opendev.org/73891421:37
corvusokay same result21:42
corvusi'll see what i can do locally21:42
*** hashar has quit IRC21:49
fungicorvus: sorry, got sidetracked by dinner prep, but lgtm21:51
corvusreplicating the upload command, this looks right: https://hub.docker.com/r/jeblair/test/tags22:15
corvusnow on to promote22:15
corvusand yep, that dropped it down to just amd6422:21
*** rlandy has quit IRC22:27
corvusapparently a manifest list is also called a "fat manifest" and has a different content-type22:34
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Handle multi-arch docker manifests in promote  https://review.opendev.org/73894522:39
corvusavass, mordred, clarkb, fungi, ianw: ^ i think that's the last step for multi-arch image publication22:40
corvusi have manually tested that locally22:40
clarkbcorvus: that is http spec breaking of them I think. If the client doesnt accept one of the things you can do you're supposed to do identity iirc22:43
corvusclarkb: well, previously we said we only supported "v2" and not "list.v2".  i guess since the server *can* supply a v2 then that was still probably the right thing to do22:47
clarkbah22:51
openstackgerritMerged zuul/zuul-jobs master: Handle multi-arch docker manifests in promote  https://review.opendev.org/73894522:53
corvusi approved https://review.opendev.org/734774 which should trigger a build and cause nodepool:latest to be multi-arch once it lasts23:01
*** saneax has joined #zuul23:14
*** tosky has quit IRC23:16
*** hamalq_ has quit IRC23:41
openstackgerritMerged zuul/nodepool master: aws: add support for attaching instance profiles  https://review.opendev.org/73477423:43
*** rfolco has quit IRC23:48
*** armstrongs has joined #zuul23:58
« Tuesday, 2020-06-30 Index Thursday, 2020-07-02 »

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!