Monday, 2019-06-24

*** threestrands has joined #zuul00:04
*** mrhillsman has quit IRC00:51
*** threestrands has quit IRC01:41
*** jhesketh has quit IRC02:12
*** jhesketh has joined #zuul02:13
*** smyers_ has joined #zuul02:18
*** smyers has quit IRC02:19
*** smyers_ is now known as smyers02:19
*** sgw has quit IRC02:33
*** bhavikdbavishi has joined #zuul02:45
*** bhavikdbavishi1 has joined #zuul03:34
*** bhavikdbavishi has quit IRC03:35
*** bhavikdbavishi1 is now known as bhavikdbavishi03:35
*** saneax has joined #zuul04:43
*** chandankumar has quit IRC05:06
*** chandankumar has joined #zuul05:07
*** swest has joined #zuul05:16
*** swest has quit IRC05:34
*** swest has joined #zuul05:35
*** swest has quit IRC05:37
*** swest has joined #zuul05:38
*** swest has quit IRC05:40
*** swest has joined #zuul05:41
*** AJaeger has joined #zuul05:57
*** swest has quit IRC06:03
*** swest has joined #zuul06:05
*** iamweswilson_ has joined #zuul06:07
*** daniel3 has joined #zuul06:07
*** evgenyl_ has joined #zuul06:08
*** samccann_ has joined #zuul06:08
*** aprice_ has joined #zuul06:08
*** ChrisShort_ has joined #zuul06:08
*** zer0c00l has joined #zuul06:10
*** swest has quit IRC06:13
*** swest has joined #zuul06:13
*** gouthamr_ has joined #zuul06:14
*** evgenyl has quit IRC06:15
*** samccann has quit IRC06:15
*** ChrisShort has quit IRC06:15
*** smcginnis has quit IRC06:15
*** zer0c00l_ has quit IRC06:15
*** aprice has quit IRC06:15
*** iamweswilson has quit IRC06:15
*** gouthamr has quit IRC06:15
*** daniel2 has quit IRC06:15
*** samccann_ is now known as samccann06:15
*** iamweswilson_ is now known as iamweswilson06:15
*** aprice_ is now known as aprice06:15
*** ChrisShort_ is now known as ChrisShort06:15
*** evgenyl_ is now known as evgenyl06:15
*** irclogbot_3 has quit IRC06:18
*** panda has quit IRC06:18
*** irclogbot_0 has joined #zuul06:20
*** panda has joined #zuul06:20
*** pcaruana has joined #zuul06:25
*** badboy has joined #zuul06:36
*** gtema has joined #zuul06:48
*** gtema has quit IRC07:06
*** gtema has joined #zuul07:09
*** swest has quit IRC07:32
*** hashar has joined #zuul07:34
*** flepied has joined #zuul07:38
*** sshnaidm|off has quit IRC07:48
*** sshnaidm has joined #zuul07:49
*** swest has joined #zuul07:53
*** swest has quit IRC08:06
*** themroc has joined #zuul08:12
badboyanybody can tell me what's causing this error?09:51
badboyZuul encountered an error while accessing the repo MYREPO The error was: type object 'EllipticCurvePublicKey' has no attribute 'from_encoded_point'09:52
*** bstinson has quit IRC09:52
*** bstinson has joined #zuul09:53
*** gtema has quit IRC09:54
*** gtema has joined #zuul09:54
*** jpena|off is now known as jpena10:00
*** gtema_ has joined #zuul10:06
*** gtema has quit IRC10:06
badboyso the problem is that paramiko-2.6.0 (latest) is causing this error, after downgrading to paramiko-2.4.2 everything is working10:08
*** bhavikdbavishi has quit IRC10:34
*** hashar has quit IRC10:43
*** jpena is now known as jpena|lunch11:20
*** rfolco has joined #zuul11:46
*** rlandy has joined #zuul11:55
*** hashar has joined #zuul11:59
tobiashcorvus, fungi: I just found a build leak in zuul that leads to builds that never end if they will pause at the end. And as far as I can see that would also leat to a node leak.12:01
tobiashcorvus, fungi: that might be related to the node leak you were seeing last week12:01
tobiashcorvus: fungi: see this https://opendev.org/zuul/zuul/src/branch/master/zuul/executor/client.py#L40112:01
tobiashcorvus, fungi: I've seen this also in my environment where the reason was that the build url was only set a few minutes after we wanted to cancel the job12:02
tobiashbut I don't know yet how to solve this12:03
tobiashI've also seen the message 'Build foo is not in the current build set' defined here: https://opendev.org/zuul/zuul/src/branch/master/zuul/scheduler.py#L128712:05
tobiashno, here: https://opendev.org/zuul/zuul/src/branch/master/zuul/scheduler.py#L118512:06
tobiashwe just return here and don't resume/cancel the job in this case, so I think we just could cancel the job in this state which should resolve that leak12:08
*** jangutter has joined #zuul12:12
*** hashar has quit IRC12:14
*** hashar has joined #zuul12:14
openstackgerritTobias Henkel proposed zuul/zuul master: Fix build leak with paused jobs  https://review.opendev.org/66707212:15
tobiashcorvus, fungi: I think this would solve that leak but I don't know how to test this ^12:17
*** gundalow has quit IRC12:36
*** gtema_ has quit IRC12:38
*** gundalow has joined #zuul12:39
*** AJaeger has quit IRC12:41
*** AJaeger has joined #zuul12:53
*** zbr is now known as zbr|ruck12:53
*** jpena|lunch is now known as jpena13:02
*** NBorg_ has quit IRC13:09
pabelangertobiash: does that mean you've upgraded to 3.9.0?13:12
pabelangeralso, have you had a change to do any ansible 2.8 zuul jobs testing?13:12
pabelangerwe flipped all jobs to 2.8 over the weekend, and haven't seen any zuul-jobs syntax issues so far13:13
tobiashpabelanger: we run 2.7 by default atm and I personally didn't test 2.8 yet13:13
tobiashbut we run most of the time from master + x ;)13:13
pabelangerack. Mostly trying to collect stats on 2.8 issues, to decided when to make the default change13:13
pabelangertobiash: ah, cool. So you are running the refactor of retries for github then13:14
tobiashyes13:15
pabelangerokay cool13:15
pabelangertobiash: I suspect, I'll have some time this week to help provide feedback, but need to review your parallel PR too13:15
tobiashoh not yet, after the next restart tomorrow morning I will13:15
*** jeliu_ has joined #zuul13:16
*** hashar has quit IRC13:17
*** hashar has joined #zuul13:25
pabelangerAnother thing, I know I asked before, why does github driver require depends-on to be the first comment of a PR? Can we not get it from each commit message?13:26
*** wxy-xiyuan has quit IRC13:29
tobiashthe thing is that a pr is complicated and can contain hundreds of commits and merges back and forth13:32
tobiashso the only workable place is the pr description13:33
tobiashand that's the only place where the user can see it easily (as github hides the commit messages themselves pretty good)13:34
pabelangerisn't commit message found under: https://github.com/ansible/ansible/pull/58273/commits for example13:36
tobiashyes, but you need to walk through all commit messages then and it's hidden from the main pr page13:38
tobiashand you cannot easily change a depends-on that's buried in the 345th commit of a pull request13:38
pabelangerRight, the bonus to that, is not having to update github UI, when you make a depends on change, via commit message. But, I hear what you are saying13:39
tobiashcheck out the hub tool, I create most pull requests via commandline13:40
tobiashand with that you don't need the ui either13:40
tobiashhttps://github.com/github/hub13:41
tobiashI've an alias 'pr' that creates a pr with the current remote tracking branch13:41
pabelangertobiash: yah, I use git pull-request from jd. It helps a lot, but with ansible/ansible, there are also https://github.com/ansible/ansible/blob/devel/.github/PULL_REQUEST_TEMPLATE.md13:42
pabelangerwhich, doesn't seem to update each git push that is done13:42
tobiashah ok13:42
pabelangerI wonder how depends-on works with pagure now that it is merged13:43
pabelangerit would be neat, if it work the same way across all drivers13:43
openstackgerritFabien Boucher proposed zuul/zuul master: URLTrigger driver time based  https://review.opendev.org/63556713:46
fungibadboy: it looks like that error means you need a newer release of pyca/cryptography13:51
*** bhavikdbavishi has joined #zuul14:03
*** bhavikdbavishi1 has joined #zuul14:06
*** bhavikdbavishi has quit IRC14:07
*** bhavikdbavishi1 is now known as bhavikdbavishi14:07
*** smcginnis has joined #zuul14:12
openstackgerritJames E. Blair proposed zuul/nodepool master: WIP: new devstack-based functional job  https://review.opendev.org/66502314:22
*** hashar has quit IRC14:24
*** themroc has quit IRC14:32
*** jamesmcarthur has joined #zuul14:58
*** panda has quit IRC15:09
*** bhavikdbavishi has quit IRC15:10
*** panda has joined #zuul15:11
openstackgerritTobias Henkel proposed zuul/zuul master: Split ansible output into its own sub logger  https://review.opendev.org/66713115:22
*** sgw has joined #zuul15:23
*** daniel3 is now known as daniel215:29
openstackgerritJames E. Blair proposed zuul/nodepool master: WIP: new devstack-based functional job  https://review.opendev.org/66502315:32
*** jpena is now known as jpena|off15:48
*** sgw has quit IRC15:53
*** jamesmcarthur has quit IRC15:56
*** gtema_ has joined #zuul16:00
openstackgerritJames E. Blair proposed zuul/zuul master: Allow config projects to override allowed-projects  https://review.opendev.org/66673316:05
*** sgw has joined #zuul16:05
corvusSpamapS: https://review.opendev.org/666733 should be ready16:15
SpamapScorvus: reviewed (note there are some errant spaces-only chunks in the YAML.. I forgot to comment on them.. I don't actually care.. but.. if you make another rev. :)16:23
clarkb /query fungi16:23
clarkbderp16:23
*** gtema_ has quit IRC16:24
corvusoops, yeah, i'll get those if there's another ps16:24
tobiashLgtm, do you want to clear those whitespaces before landing?16:30
*** mattw4 has joined #zuul16:32
openstackgerritJames E. Blair proposed zuul/zuul master: Allow config projects to override allowed-projects  https://review.opendev.org/66673316:32
corvussure :)16:32
*** hashar has joined #zuul16:37
*** bhavikdbavishi has joined #zuul16:38
openstackgerritJames E. Blair proposed zuul/nodepool master: Switch functional testing to a devstack consumer job  https://review.opendev.org/66502316:41
*** igordc has joined #zuul16:42
openstackgerritJames E. Blair proposed zuul/nodepool master: WIP: remove devstack plugin functional test jobs  https://review.opendev.org/66715616:45
clarkbcorvus: ^ one small suggestion for further optimizing the openstack cloud install16:47
*** mattw4 has quit IRC16:50
*** hwangbo has joined #zuul16:51
*** mattw4 has joined #zuul16:52
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Add install-devstack role  https://review.opendev.org/66715716:53
corvusclarkb: good point, thx -- i was about to move that role into zuul-jobs too, so i'll do that as part of the move and next patchset ^16:54
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Add install-devstack role  https://review.opendev.org/66715716:55
openstackgerritJames E. Blair proposed zuul/nodepool master: Switch functional testing to a devstack consumer job  https://review.opendev.org/66502316:56
openstackgerritJames E. Blair proposed zuul/nodepool master: WIP: remove devstack plugin functional test jobs  https://review.opendev.org/66715616:56
*** hashar has quit IRC17:06
openstackgerritMerged zuul/zuul-jobs master: Use password lookup for run-buildset-registry role  https://review.opendev.org/66311917:11
*** michael-beaver has joined #zuul17:18
*** panda has quit IRC17:18
*** panda has joined #zuul17:20
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Add install-devstack role  https://review.opendev.org/66715718:12
evgenylHi everyone, I'm trying to get `zuul-base-jobs` and `zuul-jobs`  to work for my Zuul installation. When the job is triggered it cannot find `add-build-sshkey` https://paste.ubuntu.com/p/wbfW6HK37J/ , I specified `zuul/zuul-jobs` in `untrusted-projects` as it is described in the docs https://paste.ubuntu.com/p/QDwwkwTnKj/ , any ideas how to debug that? Thanks!18:54
*** jangutter has quit IRC18:58
pabelangerevgenyl: did you add zuul/zuul-jobs to your base? eg: https://zuul-ci.org/docs/zuul/user/config.html#attr-job.roles19:00
evgenylpabelanger: Hmm, let me check that, have not noticed this config, thanks!19:02
pabelangerI am guessing there is an issue with: https://opendev.org/zuul/zuul-base-jobs/src/branch/master/zuul.yaml#L1419:02
pabelangerif you are consuming that directly19:02
*** themroc has joined #zuul19:04
evgenylpabelanger: Yes, I see now! What options do I have? I can probably add a repo with name `git.zuul-ci.org` and project `zuul-jobs`,  or am I expected to override it for every job?19:05
pabelangerevgenyl: you should use opendev.org/zuul/zuul-jobs19:05
pabelangeropendev.org is optional, depending on your connection settings in zuul.conf19:05
pabelangerevgenyl: for example, https://github.com/ansible-network/windmill-config/blob/master/zuul/zuul.conf.j2#L88 is how we set up the connection to opendev (via gerrit)19:06
evgenylpabelanger: Thanks, let me try that.19:11
*** bhavikdbavishi has quit IRC19:18
*** themroc has quit IRC19:29
*** gtema_ has joined #zuul19:33
corvusi guess we should update the base-jobs repo with the new url19:42
*** saneax has quit IRC19:44
corvusclarkb: i think https://review.opendev.org/665023 is ready for realz19:45
pabelangerHmm, we just had a log fail to upload to swift: http://paste.openstack.org/show/753324/20:00
pabelangerI wonder if it is safe to register task variable and loop to upload again20:01
clarkbcorvus: reviewing it now and I think the one piece I'm not seeing is what runs zookeeper?20:01
corvusclarkb: i'm guessing the bindep does it?20:03
clarkbcorvus: yup http://logs.openstack.org/23/665023/38/check/nodepool-functional-openstack/9168c9a/job-output.txt.gz#_2019-06-24_18_17_49_391929 looks like that is the case20:05
clarkbboth changes lgtm20:05
clarkbpabelanger: looks like you got an error from vexxhost20:08
clarkb504 gateway timeout20:08
clarkbthat would be on the vexxhost side that failed20:08
pabelangerclarkb: yah, about the first time that has happened. I think adding it some sort of retry option, will help to remove the POST_FAILURE zuul reports20:13
clarkbif the servers are timing out on the backend a retry may only make the problem worse though20:13
clarkbwe may want to check with mnaser to see if he thinks that would help20:13
openstackgerritJames E. Blair proposed zuul/nodepool master: WIP: remove devstack plugin functional test jobs  https://review.opendev.org/66715620:26
openstackgerritJames E. Blair proposed zuul/nodepool master: Add env var customization to openstack func job  https://review.opendev.org/66721720:26
pabelangerclarkb: good point20:29
*** pcaruana has quit IRC20:44
openstackgerritJames E. Blair proposed zuul/nodepool master: Remove devstack plugin functional test jobs  https://review.opendev.org/66715620:46
openstackgerritJames E. Blair proposed zuul/nodepool master: Remove devstack plugin functional test jobs  https://review.opendev.org/66715620:51
clarkbThe parallel github requests with one thread per installation id change lgtm https://review.opendev.org/#/c/666982/1. I did have one minor thing inline if it needs a respin20:51
clarkbtobiash: ^20:51
tobiashclarkb: thanks20:52
*** gtema_ has quit IRC20:53
*** gtema has joined #zuul20:56
*** rfolco has quit IRC20:57
corvuszuul-maint:  https://review.opendev.org/667228 Add zuul-operator project20:59
corvuszuul-maint: i think there is general consensus that we'd like to have an operator; even though the details are still being worked out in the spec, i think we can go ahead and import that repo and start working on the scaffolding.21:00
clarkbcorvus: fwiw https://review.opendev.org/#/c/641474/ exists too21:00
corvuswell that would have saved some time21:01
clarkbthat change definitely needs updating so maybe easiest to abandon and go with the new one21:01
clarkbbut one will want ot be abandoned I think21:01
tobiashYay :)21:01
corvusoh yeah21:01
corvuslet's go with mine, since i'm importing tristan's poc21:01
openstackgerritJames E. Blair proposed zuul/nodepool master: Add env var customization to openstack func job  https://review.opendev.org/66721721:05
openstackgerritJames E. Blair proposed zuul/nodepool master: Remove devstack plugin functional test jobs  https://review.opendev.org/66715621:05
*** gouthamr_ is now known as gouthamr21:16
*** ianychoi_ has joined #zuul21:19
*** ianychoi has quit IRC21:22
*** michael-beaver has quit IRC21:25
*** ianychoi_ has quit IRC21:30
*** ianychoi_ has joined #zuul21:31
*** gtema has quit IRC21:40
*** PaulSheldon has joined #zuul21:45
*** PaulSheldon has quit IRC21:49
clarkbcorvus: looking at https://review.opendev.org/#/c/666733/3 seems like that allows for a lot of new footgunning but https://review.opendev.org/#/c/666733/3/doc/source/user/config.rst isn't updated to warn people of the implications of this. Maybe we should add something like "a config project may still add this job to any project's pipeline. Apply caution when doing so as other projects may expose the source21:57
clarkbproject's secrets.21:57
clarkbor something21:57
*** panda has quit IRC21:58
*** panda has joined #zuul22:03
*** tobiash has quit IRC22:10
corvusclarkb: good point.  it's basically a third party stepping in and saying that the second party is trustworthy -- we should convey that.22:25
fungimaybe with a big bold admonition callout box22:26
corvusfungi: above or below the existing one? :)22:26
fungiheh22:26
fungiyeah, i guess inside22:27
clarkbor maybe even a separate "things to consider with secerts" section22:30
openstackgerritJames E. Blair proposed zuul/zuul master: Strengthen the caution about allowed-projects  https://review.opendev.org/66724122:31
corvusclarkb, fungi: ^ implemented as a followup so we can bikeshed with abandon22:31
clarkbthat looks good for now. I think any things to consider seciont would be a bigger followup22:33
*** jeliu_ has quit IRC22:39
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Expand documentation of test-setup role  https://review.opendev.org/66724422:54
openstackgerritJames E. Blair proposed zuul/nodepool master: Switch functional testing to a devstack consumer job  https://review.opendev.org/66502322:55
openstackgerritJames E. Blair proposed zuul/nodepool master: Add env var customization to openstack func job  https://review.opendev.org/66721722:55
openstackgerritJames E. Blair proposed zuul/nodepool master: Remove devstack plugin functional test jobs  https://review.opendev.org/66715622:55
*** ianw_pto is now known as ianw23:08
*** saneax has joined #zuul23:41
*** rlandy has quit IRC23:41

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!