Tuesday, 2019-09-24

« Monday, 2019-09-23 Index Wednesday, 2019-09-25 »
*** sgw has quit IRC00:37
*** armstrongs has joined #zuul00:53
*** jamesmcarthur has joined #zuul01:11
*** jamesmcarthur has quit IRC01:14
*** sgw has joined #zuul01:15
*** jamesmcarthur has joined #zuul01:21
*** mauved has joined #zuul01:56
*** saneax has joined #zuul02:08
openstackgerritIan Wienand proposed zuul/zuul-jobs master: Remove invalid argument from format_exc() calls  https://review.opendev.org/68416702:17
*** jamesmcarthur has quit IRC02:22
*** jamesmcarthur has joined #zuul02:23
*** jamesmcarthur_ has joined #zuul02:27
*** rfolco has quit IRC02:27
*** jamesmcarthur has quit IRC02:28
*** bhavikdbavishi has joined #zuul02:29
*** rlandy has quit IRC02:33
*** bhavikdbavishi1 has joined #zuul02:36
*** bhavikdbavishi has quit IRC02:38
*** bhavikdbavishi1 is now known as bhavikdbavishi02:38
*** saneax has quit IRC02:40
openstackgerritMerged zuul/zuul-jobs master: Add more languages for build-releasenotes  https://review.opendev.org/68392302:43
*** bhavikdbavishi has quit IRC03:25
*** bhavikdbavishi has joined #zuul04:42
*** bhavikdbavishi1 has joined #zuul04:45
*** bhavikdbavishi has quit IRC04:47
*** bhavikdbavishi1 is now known as bhavikdbavishi04:47
*** bhavikdbavishi has quit IRC04:48
*** igordc has quit IRC05:12
openstackgerritJan Kubovy proposed zuul/zuul master: Report retried builds in a build set via mqtt.  https://review.opendev.org/63272705:59
openstackgerritJan Kubovy proposed zuul/zuul master: Report retried builds via sql reporter.  https://review.opendev.org/63350105:59
*** saneax has joined #zuul06:10
*** yolanda has quit IRC06:34
*** tosky has joined #zuul06:53
*** pcaruana has joined #zuul06:56
*** avass has joined #zuul06:57
*** hashar has joined #zuul06:57
*** webknjaz has quit IRC07:12
*** webknjaz has joined #zuul07:13
flaper87Is there a way to allow to a set of trusted users to manually trigger a periodic job? (I don't even know if trigger periodic jobs manually is posisble)07:24
flaper87possible*07:24
*** tobiash has quit IRC07:27
*** tobiash has joined #zuul07:31
tobiashflaper87: with access to the scheduler you can use the zuul enqueue command07:36
tobiashflaper87: and there is work going on adding auth support for such things to the web ui07:37
*** recheck has quit IRC07:41
*** recheck has joined #zuul07:41
openstackgerritTobias Henkel proposed zuul/zuul master: Evaluate CODEOWNERS settings during canMerge check  https://review.opendev.org/64455708:01
openstackgerritMerged zuul/zuul-jobs master: Remove invalid argument from format_exc() calls  https://review.opendev.org/68416708:29
*** jpena|off is now known as jpena08:30
*** olaph has quit IRC08:45
*** saneax is now known as saneax|AFK09:34
*** saneax|AFK has quit IRC09:40
*** hashar has quit IRC09:56
*** openstackstatus has quit IRC10:12
*** openstack has joined #zuul10:17
*** ChanServ sets mode: +o openstack10:17
*** pcaruana has quit IRC10:24
*** pcaruana has joined #zuul10:48
*** avass has quit IRC10:50
*** jpena is now known as jpena|lunch11:34
*** jamesmcarthur has joined #zuul11:39
*** jamesmcarthur has quit IRC11:49
*** fdegir has quit IRC11:57
*** fdegir has joined #zuul11:58
*** jamesmcarthur has joined #zuul12:02
*** jamesmcarthur_ has joined #zuul12:03
*** jamesmcarthur has quit IRC12:07
*** rfolco has joined #zuul12:10
openstackgerritMerged zuul/zuul master: Unify Github/Github Enterpise read timeout  https://review.opendev.org/68336612:13
*** jamesmcarthur_ has quit IRC12:20
*** hashar has joined #zuul12:22
*** sgw has quit IRC12:34
*** jpena|lunch is now known as jpena12:35
*** rlandy has joined #zuul12:39
*** saneax|AFK has joined #zuul13:11
fungirobyn is wearing a zuul pin on stage13:13
fungi(ansiblefest keynotes)13:14
clarkbya she told me she stole it from our booth this morning :)13:14
fungihah!13:14
fungigood job13:14
clarkbgood thing we set up yesterday13:15
AJaeger\o/13:16
fungiwell, we do still need to get thursday added to our sessions slide, now that we have details for what's happening13:16
fungibut otherwise ready to go, yes ;)13:16
*** hashar has quit IRC13:33
flaper87Does `soft` apply to failed jobs? https://zuul-ci.org/docs/zuul/user/config.html?highlight=dependencies#attr-job.dependencies.soft13:34
flaper87If not, is there a way to have a job run after a set of jobs regardless of their exit status ?13:35
clarkbif you dont care about the exit status can you run them concurrently?13:37
clarkbthere is a zuul_return method for controlling child jobs that may work for you13:38
openstackgerritTobias Henkel proposed zuul/zuul master: Support emitting warnings via zuul_return  https://review.opendev.org/65152613:38
clarkbhttps://zuul-ci.org/docs/zuul/user/jobs.html#skipping-child-jobs and invert the checks?13:39
openstackgerritTobias Henkel proposed zuul/zuul master: Support emitting warnings via zuul_return  https://review.opendev.org/65152613:40
flaper87clarkb: the workflow I'm trying to replicate is: Job A deploys a k8s cluster -> Job B, C, D use that cluster and do a bunch of things -> Job E cleans everything13:44
tobiashflaper87: no, it only applies to jobs that didn't run13:45
flaper87I'm sure there's a better way to do this in Zuul, tho13:45
tobiashflaper87: a way to solve this is pause A after deploying the cluster and clean up in the post or cleanup playbook13:46
tobiashthen you don't need job E for this use case13:46
clarkbyup I would pause job A. We do that to run a docker registry for child jobs13:46
flaper87tobiash: clarkb ah, nice! Will give that a try13:47
openstackgerritPaul Belanger proposed zuul/zuul-jobs master: DNM - ansiblefest 2019 demo!  https://review.opendev.org/68432213:49
*** hashar has joined #zuul13:51
*** sgw has joined #zuul14:10
*** lennyb has quit IRC14:11
pabelangerMS just gave a keynote at ansiblefest, how they used ansible to manage their network infra. Level 0 to Level 1. From the sounds of it, they are ready for level 2 based on this slide: https://pbs.twimg.com/media/EFPA_5EXYAIN4VP?format=jpg&name=4096x4096 If only there was some sort of Ansible CI, that is also good at network integration14:11
*** lennyb has joined #zuul14:23
*** hashar has quit IRC14:31
openstackgerritDavid Shrewsbury proposed zuul/nodepool master: Do not overwrite image upload ZK data on delete  https://review.opendev.org/68185714:39
openstackgerritDavid Shrewsbury proposed zuul/nodepool master: Do not overwrite image upload ZK data on delete  https://review.opendev.org/68185714:43
tobiashcorvus, clarkb: is it possible to withdraw a gear job that is submitted to gearman but not yet in process?14:45
tobiashI have the situation that when all executors are busy and take some time to fetch new jobs canceling the job doesn't work14:46
tobiashhttp://paste.openstack.org/show/779190/14:47
tobiashzuul knows that the build has not started yet, but it only waits one second and tries again14:48
tobiashthis fails and after that the next free executor takes the build and fully executes it14:50
tobiashoh, it actually tries that, but it didn't seem to work: https://opendev.org/zuul/zuul/src/branch/master/zuul/executor/client.py#L39714:53
tobiashfound this in the logs: Response to cancel build request: b'ERR UNKNOWN_JOB'14:58
*** tosky has quit IRC15:00
clarkbtobiash: you have to kill the gearman clirnt15:00
tobiashyou mean on the executor?15:01
*** pcaruana has quit IRC15:01
*** armstrongs has quit IRC15:06
openstackgerritTobias Henkel proposed zuul/zuul master: Annotate cancelJobInQueue with build id  https://review.opendev.org/68437315:09
*** hashar has joined #zuul15:10
*** saneax|AFK has quit IRC15:14
pabelangerso, apparently running ansible under python2, ssh-keyscan can return connection refused on a port, and ansible is happy to continue on... with python3, not so much. Ansible shell tasks errors.15:21
*** lennyb has quit IRC15:25
*** lennyb has joined #zuul15:30
*** noorul has joined #zuul15:48
noorulIs there a python nosetest result HTML publisher role in zuul?15:48
*** hashar has quit IRC15:52
SpamapSnoorul: if there were one, it would be in the zuul-jobs repo. https://zuul-ci.org/docs/zuul-jobs/15:56
SpamapSnoorul: Looks like the focus is on tox and subunit.15:57
SpamapSpabelanger: that seems unintentional. ;)15:57
*** noorul has quit IRC15:59
*** noorul has joined #zuul16:01
*** panda is now known as panda|bbl16:02
*** sgw has quit IRC16:05
*** hashar has joined #zuul16:09
*** noorul has quit IRC16:14
*** sgw has joined #zuul16:22
*** noorul has joined #zuul16:31
*** yoctozepto has quit IRC16:35
noorulSay, in run phase I am running nose tests with some failures16:35
noorulWill zuul run the post phase?16:35
*** yoctozepto has joined #zuul16:40
clarkbnoorul: yes16:45
SpamapSnoorul: yes16:45
SpamapS:)16:45
noorulhttps://opendev.org/zuul/zuul-jobs/src/branch/master/zuul.d/python-jobs.yaml#L116:46
noorultox has a parent16:46
noorulunittest16:46
nooruldoes tox's post-run overrides unittests?16:47
noorulOr both of the gets invoked in a order?16:47
nooruls/the/them16:48
noorulAny idea which job is using https://opendev.org/zuul/zuul-jobs/src/branch/master/zuul.d/python-jobs.yaml#L46 ?16:49
AJaegernoorul: https://zuul-ci.org/docs/zuul/user/config.html?highlight=post%20run#job explains what is overridden and what not16:51
tristanCnoorul: parent's pre and post run are executed before and after the childs'16:51
AJaegernoorul: use codesearch.openstack.org to figure out where variables are used16:51
AJaegernoorul: some might just passed on to tox without change and can be used there16:52
*** panda|bbl is now known as panda16:56
noorulAJaeger: Thank you!16:56
*** jamesmcarthur has joined #zuul17:25
*** jamesmcarthur has quit IRC17:31
*** jpena is now known as jpena|off17:32
*** igordc has joined #zuul17:48
*** jamesmcarthur has joined #zuul17:54
noorulofosos: hi18:01
ofososnoorul: hi18:03
noorulofosos: Do you have date in mind to get the driver merged?18:04
ofososLet's be optimistic: I want it to merge next week18:04
noorulofosos: Great!18:04
ofososI've some free time coming up18:04
ofososI hope to address the bug issues this week and then go take care of the nits next week18:05
ofososBig18:05
noorulHmm18:06
ofososThen the next big thing to come up is eventing18:07
ofososSo, support for Zuul to listen to Bitbucket events18:07
noorulweb hook?18:07
ofososYes18:08
noorulAre you going to work on webhooks also?18:08
noorulOne of my colleagues said he will work on it18:08
ofososAfter the initial merge18:08
ofososThen let's pool resources18:08
noorulHmm18:08
ofososI don't think it's that much work18:09
noorulWe did some work as part of a hackathon18:09
noorulI think we just need to wrap it with some test cases18:09
noorulHe said he will invest some time this week18:09
noorulLet me check with him18:09
ofososI think the only thing I want to make sure is that we're able to merge next week18:09
noorulThat will be a great first step18:10
noorulIt will help to contribute on top of that18:10
*** Goneri has joined #zuul18:10
noorulAnother key thing is dependent PR18:11
ofososDependent Pipeline should be good. Dependent pr I will fix this week18:15
noorulofosos: cool, that will be great18:15
noorulofosos: In your deployment have you notice the issue, where a job is already queued for PR and user pushes another change to the PR and a new job is queued in parallel18:16
noorul?18:16
mnaserhttps://twitter.com/_mnaser/status/117655813282848768018:22
mnasera lot of people are excited by zuul :)18:22
mnaserthe room is overflowing!18:23
noorulIs there a live stream?18:26
noorulmnaser: Is there a live stream?18:28
ofososNoorul: check or gate?18:28
ofososI think there should be a pipeline parameter to handle this18:29
mnasernoorul: no i dont think so :(18:29
mnaserbut afaik the talks are recorded18:29
noorulofosos: check18:30
ofososParallel jobs should be fine for check, no?18:30
noorulActual behavior is that the old one should get dequeued18:34
noorulThere is no point in running the old one18:34
*** adam_g has joined #zuul18:35
*** hashar has quit IRC18:35
*** Goneri has quit IRC18:37
openstackgerritTobias Henkel proposed zuul/zuul master: DNM: Try to reproduce cancel failure  https://review.opendev.org/68441418:38
tobiashcorvus, clarkb: this is the scenario I'm expecting to cause problems in our production system, but I coudn't get the test case to reproduce. It successfully removed the build from the gearman queue. ^18:39
*** jamesmcarthur has quit IRC18:42
*** jamesmcarthur has joined #zuul18:44
noorulIs there a way to determine how much time each task in a playbook took?18:44
tobiashnoorul: the job-output.json contains this information18:45
noorultobiash: I see start and end18:46
noorultobiash: I am looking for elapsed18:47
*** jamesmcarthur has quit IRC18:47
tobiashelapsed can be calculated from those two18:47
noorulIt looks like ARA report has it18:48
noorultobiash: Thanks18:48
daniel2Anyone seen this before: /usr/local/lib/python3.7/site-packages/paramiko/client.py:837: UserWarning: Unknown ssh-ed25519 host key for [gerrit]:29418: b'ead8f054e8261e8ac9235061f0d0ff9f'18:50
tobiashdaniel2: this happens normally on first startup if the hostkey of gerrit is not yet known18:51
tobiashbut it's only a warning18:51
daniel2oh it is this that crashed it: sqlalchemy.exc.InternalError: (pymysql.err.InternalError) Packet sequence number wrong - got 1 expected 018:52
daniel2Zuul executor apparently died from this error18:53
*** hashar has joined #zuul18:53
*** noorul has quit IRC18:55
daniel2Does gearman come in the docker compose nodepool setup?18:56
daniel2Weird, so it's in the scheduler, however executor is giving a host not found even though I can exec into the container and telnet the port and get a connection.19:03
*** noorul has joined #zuul19:24
*** noorul has quit IRC19:29
openstackgerritTobias Henkel proposed zuul/zuul master: DNM: Try to reproduce cancel failure  https://review.opendev.org/68441419:33
tobiashcorvus, clarkb: this now contains a theory and an idea how to fix this ^19:34
tobiashcorvus, clarkb: it would be great if you could grep your scheduler logs for "Unable to cancel build" to check if this occurs in opendev as well19:35
tobiashlooking at the code I'd expect that to happen there as well (many of our projects use fail fast so this might be more dramatically for us)19:37
clarkbtobiash: I can try to remember to check when the laptop is out next19:38
tobiashthanks19:38
tobiashno hurries, enjoy ansiblefest :)19:38
tobiashI'm going to eod now19:38
tobiashI just wanted to dump all info I have here before going to sleep ;)19:39
*** jamesmcarthur has joined #zuul19:39
tobiashwe had 750 "unable to cancel build" occurrences today19:42
tobiashand I guess we should raise the log level at least to warning19:43
openstackgerritTobias Henkel proposed zuul/zuul master: DNM: Try to reproduce cancel failure  https://review.opendev.org/68441419:44
*** rlandy has quit IRC19:50
*** gtema has joined #zuul19:53
*** jamesmcarthur has quit IRC19:54
*** rlandy has joined #zuul19:54
*** jamesmcarthur has joined #zuul19:55
*** noorul has joined #zuul19:57
*** noorul has quit IRC20:02
*** jamesmcarthur has quit IRC21:06
*** rfolco has quit IRC21:11
*** rfolco has joined #zuul21:11
*** hashar has quit IRC21:33
*** sgw has quit IRC21:55
*** sgw has joined #zuul21:57
*** jamesmcarthur has joined #zuul22:06
*** jamesmcarthur has quit IRC22:13
*** gtema has quit IRC22:13
*** rfolco has quit IRC22:29
*** rfolco has joined #zuul22:29
*** noorul has joined #zuul23:00
*** noorul has quit IRC23:05
*** rfolco has quit IRC23:13
*** rfolco has joined #zuul23:14
*** jamesmcarthur has joined #zuul23:50
*** jamesmcarthur has quit IRC23:55
*** armstrongs has joined #zuul23:55
« Monday, 2019-09-23 Index Wednesday, 2019-09-25 »

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