| *** Defolos has quit IRC | 00:07 | |
| *** mattw4 has quit IRC | 00:33 | |
| *** shanemcd has quit IRC | 00:55 | |
| *** shanemcd has joined #zuul | 00:55 | |
| *** rfolco has quit IRC | 01:44 | |
| *** Goneri has quit IRC | 02:34 | |
| *** swest has quit IRC | 02:36 | |
| *** igordc has quit IRC | 02:45 | |
| *** openstackstatus has joined #zuul | 02:56 | |
| *** ChanServ sets mode: +v openstackstatus | 02:56 | |
| *** rlandy|bbl is now known as rlandy | 03:03 | |
| *** bhavikdbavishi has joined #zuul | 03:14 | |
| *** jamesmcarthur has joined #zuul | 03:14 | |
| *** bhavikdbavishi1 has joined #zuul | 03:17 | |
| *** bhavikdbavishi has quit IRC | 03:18 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:18 | |
| *** sgw has quit IRC | 03:25 | |
| *** evrardjp has quit IRC | 03:42 | |
| *** evrardjp has joined #zuul | 03:45 | |
| *** sgw has joined #zuul | 03:45 | |
| *** jamesmcarthur has quit IRC | 03:48 | |
| *** jamesmcarthur has joined #zuul | 03:49 | |
| *** jamesmcarthur has quit IRC | 03:54 | |
| *** jamesmcarthur has joined #zuul | 04:13 | |
| *** jamesmcarthur has quit IRC | 04:25 | |
| *** rlandy has quit IRC | 04:41 | |
| *** bhavikdbavishi has quit IRC | 04:58 | |
| *** bhavikdbavishi has joined #zuul | 04:58 | |
| *** saneax has joined #zuul | 05:01 | |
| *** bolg has joined #zuul | 05:08 | |
| *** mmedvede has quit IRC | 05:17 | |
| *** mmedvede has joined #zuul | 05:18 | |
| *** evrardjp has quit IRC | 05:34 | |
| *** evrardjp has joined #zuul | 05:34 | |
| *** reiterative has quit IRC | 05:37 | |
| *** reiterative has joined #zuul | 05:38 | |
| *** NBorg has joined #zuul | 06:01 | |
| NBorg | Hello. I have added a project to a tenant, but it does not appear under projects for that tenant. I've tried smart-reconfigure a couple of times and full-reconfigure once, but I don't see anything about that project in the scheduler log. Any ideas about where I could look for clues? | 06:15 |
|---|---|---|
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 06:17 |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Diable misfire grace time of apscheduler job https://review.opendev.org/707842 | 06:22 |
| *** yolanda has joined #zuul | 06:22 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Disable misfire grace time of apscheduler job https://review.opendev.org/707842 | 06:22 |
| *** jamesmcarthur has joined #zuul | 06:27 | |
| *** jamesmcarthur has quit IRC | 06:31 | |
| *** felixedel has joined #zuul | 06:55 | |
| *** felixedel has quit IRC | 06:59 | |
| *** felixedel has joined #zuul | 07:00 | |
| *** swest has joined #zuul | 07:26 | |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Scheduler test app factory https://review.opendev.org/708812 | 07:27 |
| *** dpawlik has joined #zuul | 07:31 | |
| *** snapiri has joined #zuul | 07:47 | |
| *** jpena|off is now known as jpena | 07:51 | |
| *** jcapitao_away is now known as jcapitao | 08:06 | |
| *** Defolos has joined #zuul | 08:07 | |
| *** bolg has quit IRC | 08:15 | |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 08:21 |
| *** tosky has joined #zuul | 08:29 | |
| mnaser | NBorg: do you see any issues in your zuul web interface? there might be a little bell in the top right corner | 08:36 |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 08:43 |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Scheduler test app factory https://review.opendev.org/708812 | 08:48 |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: ci: use bionic for functional tests https://review.opendev.org/708820 | 08:53 |
| *** felixedel has quit IRC | 08:56 | |
| *** felixedel has joined #zuul | 08:58 | |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: DNM: exercising usage of bionic for stream jobs https://review.opendev.org/708821 | 09:01 |
| *** felixedel has quit IRC | 09:02 | |
| *** carli has joined #zuul | 09:10 | |
| *** avass has joined #zuul | 09:30 | |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 09:39 |
| *** bhavikdbavishi has quit IRC | 09:56 | |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: Authorization rules: add templating https://review.opendev.org/705193 | 10:01 |
| *** sugaar has joined #zuul | 10:07 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Make most test cases work on MacOS https://review.opendev.org/707585 | 10:08 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Tests ensure-tox on all-platforms https://review.opendev.org/708642 | 10:14 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Tests bindep role on all-platforms https://review.opendev.org/708704 | 10:14 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Tests bindep role on all-platforms https://review.opendev.org/708704 | 10:15 |
| *** NBorg has quit IRC | 10:28 | |
| *** sgw has quit IRC | 10:48 | |
| *** reiterative has quit IRC | 11:04 | |
| *** reiterative has joined #zuul | 11:04 | |
| *** sshnaidm|afk is now known as sshnaidm | 11:23 | |
| *** paulalbertella has joined #zuul | 11:29 | |
| *** reiterative has quit IRC | 11:30 | |
| yoctozepto | https://storyboard.openstack.org/#!/story/2007316 I found how to pass zuul while breaking all jobs :-) | 11:38 |
| *** jcapitao is now known as jcapitao_lunch | 11:52 | |
| *** rfolco has joined #zuul | 11:52 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Make most test cases work on MacOS https://review.opendev.org/707585 | 11:59 |
| *** jpena is now known as jpena|lunch | 12:00 | |
| tobiash | yoctozepto: interesting | 12:01 |
| *** bhavikdbavishi has joined #zuul | 12:16 | |
| AJaeger | yoctozepto: Congratulations! ;) | 12:32 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Add zuul-operator-upload-image job https://review.opendev.org/708860 | 12:34 |
| mnaser | oh neat | 12:35 |
| mnaser | my little 'warning' check caught some technically invalid zuul test configs | 12:35 |
| yoctozepto | AJaeger: well, thank you ;p actually discovered by having to fix other mess in kolla | 12:35 |
| *** jcapitao_lunch has quit IRC | 12:38 | |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 12:42 |
| tobiash | zuul-maint: I'd appreciate a second review on merge handling improvements: https://review.opendev.org/#/q/topic:merger-refactor | 12:46 |
| *** rlandy has joined #zuul | 12:51 | |
| *** sshnaidm has quit IRC | 13:00 | |
| *** sshnaidm has joined #zuul | 13:01 | |
| Shrews | corvus: tobiash: I left a comment (and +2) on the scheduler scale out spec that I think is pretty important. | 13:12 |
| *** jamesmcarthur has joined #zuul | 13:14 | |
| tobiash | Shrews: should we write this into the spec (gathering metrics) or just to keep in mind? | 13:14 |
| Shrews | tobiash: just to keep in mind. I think ideally there would be some sort of workload that we could capture and replay during development to gather the metrics | 13:15 |
| *** jamesmcarthur has quit IRC | 13:15 | |
| tobiash | Shrews: we're using the zookeeper exporter and have quite decent performance data of our zookeeper | 13:15 |
| Shrews | tobiash: but i'm not sure how feasible that is | 13:15 |
| Shrews | tobiash: what is zk exporter? | 13:16 |
| tobiash | it's a metrics exporter that makes zookeeper metrics available to prometheus | 13:16 |
| Shrews | oh that's neat | 13:16 |
| openstackgerrit | Felix Edel proposed zuul/zuul master: Report retried builds via sql reporter. https://review.opendev.org/633501 | 13:17 |
| Shrews | tobiash: i think the biggest thing we want to be able to say to someone running maybe 1 zk server (or 3, etc) is if they should consider adding more nodes or not before using the new scaled out scheduler | 13:19 |
| tobiash | Shrews: zk (write) performance is highest with fewer nodes (3 to 7 is recommended) | 13:20 |
| tobiash | fwiw we're using this exporter: https://github.com/carlpett/zookeeper_exporter | 13:21 |
| openstackgerrit | Albin Vass proposed zuul/zuul-base-jobs master: Remove ssh key in base cleanup run. https://review.opendev.org/708871 | 13:21 |
| *** jpena|lunch is now known as jpena | 13:24 | |
| tobiash | Shrews: looks like for reads it scales quite well with number of nodes: https://zookeeper.apache.org/doc/r3.5.5/zookeeperOver.html#Performance | 13:25 |
| avass | ^ We changed our base job to look like this instead since we encountered some problems with removing ssh keys during post-run with static nodes. | 13:25 |
| tristanC | tobiash: Shrews: also, during the implementation phase, would it be possible to toggle the zookeeper usage to avoid affecting the performance of existing user, at least until the scaled-out feature is completed | 13:26 |
| tristanC | tobiash: hmm, is it normal there is no performance graph for writes ? :) | 13:28 |
| tobiash | tristanC: that graph contains the writes (as a percentage on the x axis) | 13:28 |
| *** jamesmcarthur has joined #zuul | 13:29 | |
| tristanC | tobiash: oh i see, iiuc it's less than 10k write per sec if you have 13 server | 13:30 |
| tobiash | tristanC: of you have only writes then yes ;) | 13:30 |
| Shrews | I just want us to keep ZK performance in mind as we add workload to it. I don't know anything about tuning it or what factors cause it to slow down, but I know it can hurt us if it doesn't work well. That's all I'm suggesting. :) | 13:30 |
| tobiash | so tldr, writes are expensive and we'll try to minimize them | 13:31 |
| *** jamesmcarthur has quit IRC | 13:31 | |
| tristanC | should we recommend to use two clusters in big setup, like one for nodepool and one for zuul? | 13:32 |
| *** jamesmcarthur has joined #zuul | 13:32 | |
| tristanC | iirc, nodepool is already doing quite a few rw request | 13:32 |
| tobiash | tristanC: that cannot be said atm, we need to see how it performs | 13:33 |
| Shrews | yep, nodepool is constantly iterating through the nodes. lots of zk traffic from it already | 13:33 |
| tristanC | tobiash: i meant, will it even be possible to use two clusters? | 13:33 |
| tobiash | tristanC: theoretically that should be no problem, we'd need to refactor the config for this | 13:34 |
| tobiash | but I think we should avoid it if possible | 13:34 |
| *** avass has quit IRC | 13:34 | |
| tristanC | tobiash: i agree it seems over complicated, but as the scaled-out implementation is a breaking change, we might as well do that config refactor if it's needed | 13:35 |
| tobiash | but we have a quite decent deployment, so opendev or we will likely be the first ones to notice performance problems in this regard ;) | 13:36 |
| tobiash | I'm not sure if it will be that breaking once it's working | 13:36 |
| tristanC | tobiash: yeah, hence my question about toggling the feature until it's ready for prime-time. | 13:37 |
| tristanC | tobiash: it is breaking in the sense that executor and schedulers will requires access to zookeeper | 13:37 |
| tobiash | scheduler already has access to zookeeper | 13:37 |
| tobiash | what's breaking is executor's access to zookeeper | 13:38 |
| tristanC | executor access to zookeeper is a significant change, especially without zookeeper-auth ... | 13:39 |
| tobiash | do you have executos on different locations? | 13:39 |
| tristanC | we need that to avoid using floating-ip yes | 13:41 |
| tobiash | then I think we should consider reviving your zk auth changes | 13:42 |
| tristanC | i'm on it yes :) | 13:42 |
| *** jamesmcarthur has quit IRC | 13:44 | |
| tobiash | tristanC: an alternative could be ssl.clientAuth (https://zookeeper.apache.org/doc/r3.5.5/zookeeperAdmin.html) on zookeeper | 13:45 |
| openstackgerrit | Tristan Cacqueray proposed zuul/nodepool master: Implement zookeeper-auth https://review.opendev.org/619155 | 13:45 |
| tristanC | tobiash: okay... seems like a very different alternative. should we do that instead? | 13:47 |
| tobiash | https://cwiki.apache.org/confluence/display/ZOOKEEPER/ZooKeeper+SSL+User+Guide | 13:47 |
| tobiash | that link is better | 13:47 |
| tobiash | sounds like if you're doing ssl you're automatically auth'ed | 13:47 |
| tobiash | maybe that's even better than handling it in zuul | 13:47 |
| tobiash | (needs at least zookeeper 3.5.5) | 13:48 |
| tristanC | tobiash: and what about management. sasl auth is what is currently supported by the bitnami zookeeper image, is there operator or recipe for managing a ssl zookeeper service? | 13:49 |
| tobiash | ah I guess we still need to set the ACLs (which can refer then the client id) | 13:50 |
| tristanC | hmm, https://github.com/pravega/zookeeper-operator doesn't support either sasl or ssl | 13:51 |
| tobiash | maybe start first with your change (it'll be neccessary anyway as it seems) and add tls if needed | 13:51 |
| tristanC | why would we need acl with tls? | 13:52 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Tests ensure-tox on all-platforms https://review.opendev.org/708642 | 13:52 |
| tobiash | tristanC: reading the docs it reads like anyone can connect but you can set acls referring to client certs | 13:52 |
| tobiash | so pretty much like sasl | 13:52 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Implement zookeeper-auth https://review.opendev.org/619156 | 13:53 |
| *** jamesmcarthur has joined #zuul | 13:54 | |
| *** Goneri has joined #zuul | 13:55 | |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 14:05 |
| *** phildawson has joined #zuul | 14:08 | |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: ci: use bionic for functional tests https://review.opendev.org/708820 | 14:12 |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: DNM: exercising usage of bionic for stream jobs https://review.opendev.org/708821 | 14:12 |
| *** jamesmcarthur has quit IRC | 14:26 | |
| mordred | pabelanger: heya - I have a task to start publishing the openstack collection to galaxy. the role we have in zuul-jobs is only for doing galaxy import vs galaxy publish ... before I go and do any work here, I know you're doing a lot with collections - do you have any collections content that's not in zuul-jobs that I might want to use or steal? | 14:29 |
| *** sshnaidm is now known as sshnaidm|afk | 14:34 | |
| openstackgerrit | Felix Edel proposed zuul/zuul master: Report retried builds via sql reporter. https://review.opendev.org/633501 | 14:39 |
| pabelanger | mordred: I've been using https://github.com/ansible/ansible-zuul-jobs/tree/master/roles/upload-ansible-collection-fork but hasn't pushed upstream yet | 14:39 |
| pabelanger | I think we could now | 14:39 |
| mordred | pabelanger: oh neat- that uses twine? | 14:40 |
| pabelanger | https://dashboard.zuul.ansible.com/t/ansible/build/abd816bcc33d45b3abcaa9b450e8a620 is an example of our release-ansible-collection-galaxy | 14:40 |
| pabelanger | mordred: no, readme is wrong | 14:40 |
| pabelanger | copypasta | 14:40 |
| mordred | :) | 14:40 |
| pabelanger | we use ansible-galaxy | 14:40 |
| mordred | cool | 14:40 |
| pabelanger | https://github.com/ansible-network/releases/blob/master/ansible_releases/cmd/generate_ansible_collection.py | 14:41 |
| pabelanger | is how we generate version numbers for galaxy | 14:41 |
| pabelanger | galaxy.yml | 14:41 |
| mordred | pabelanger: I think the ansible-galaxy setup is identical to ansible-galaxy-upload - so maybe we should split out an ansible-galaxy-login role and include_role it | 14:41 |
| pabelanger | I want to add that into ansible-galaxy CLI by default | 14:41 |
| mordred | YES | 14:41 |
| mordred | but I will be stealing that in the meantime | 14:41 |
| pabelanger | mordred: yah, there is some overlap. But, you need an ansible.cfg file now, roles didn't | 14:42 |
| *** jamesmcarthur has joined #zuul | 14:42 | |
| mordred | nod. maybe we'll just leave it for now | 14:42 |
| mordred | pabelanger: do you have time to submit that to zuul-jobs? or would you like me to? | 14:42 |
| pabelanger | sure, I can add PR | 14:43 |
| pabelanger | err, change | 14:43 |
| mordred | pabelanger: you're being infected ;) | 14:43 |
| mnaser | based on what i'm seeing, it seems like if a job is aborted inside zuul, it doesnt actually get logged in the sql reporter? | 14:44 |
| mordred | pabelanger: and thanks! sshnaidm|afk will be happy | 14:44 |
| pabelanger | ++ | 14:44 |
| mordred | tobiash: weren't you working on better reporting for aborted jobs? | 14:44 |
| mnaser | my scenario is if a 2 hour job has been running and then someone submits another change | 14:44 |
| mordred | tobiash: or am I just hallucinating that? | 14:44 |
| pabelanger | I guess I only pushed up https://review.opendev.org/702502/ | 14:44 |
| pabelanger | for github actions | 14:44 |
| mnaser | just for context | 14:45 |
| mnaser | using the zuul builds db for billing purposes | 14:45 |
| mordred | pabelanger: oy | 14:45 |
| mnaser | so needing accuracy for every single confused build minute for example | 14:45 |
| mnaser | confused?? | 14:45 |
| mnaser | what am i even writing | 14:45 |
| mordred | :) | 14:45 |
| mnaser | counted is what i assume i was going for | 14:45 |
| *** snapiri has quit IRC | 14:46 | |
| tobiash | mordred, mnaser: check out https://review.opendev.org/633501, https://review.opendev.org/632727 and https://review.opendev.org/696670 | 14:46 |
| pabelanger | mnaser: yah, openstack and rdo track that via ELK, when logs are uploaded | 14:46 |
| tobiash | mnaser: there are also statsd-based usage metrics that report job-hours based on labels | 14:47 |
| mordred | tobiash: that last one seems to match mnaser's use case - gate resets would include jobs aborted due to new patch yeah? | 14:47 |
| mnaser | yeah i come from the prometheus land so i assume we can use that too, but i dont trust prometheus for billing stuff (yet) | 14:48 |
| tobiash | mordred: only in the gate, check would still not land in the db with that | 14:48 |
| tobiash | (would need to be added as well then) | 14:49 |
| mordred | tobiash: gotcha. so we'd still need a followup patch | 14:49 |
| mordred | seems to be a reasonable thing to add with adding the others | 14:49 |
| mnaser | yeah if this lands i think i can add that onto it | 14:49 |
| mnaser | as this will help establish a pattern i can reuse | 14:50 |
| mordred | ++ | 14:50 |
| tobiash | mordred: but I just today discussed with Felix the possible addition of a dequeue-reporter that would be useful when working with checks api | 14:50 |
| tobiash | this also could be used to report any items that got dequeued without error or success | 14:50 |
| mnaser | something i've been mulling over is decoupling of pipeline and zuul config... | 14:50 |
| mordred | ooh yeah | 14:50 |
| tobiash | I think this would solve the check use case as well as a side effect | 14:51 |
| mordred | yeah, I think so | 14:51 |
| mordred | mnaser: whatcha mean? | 14:51 |
| mnaser | right now for things like reporters/triggers/etc, you need to *know* what the zuul.conf connections are defined as | 14:51 |
| mnaser | say, if i changed the connection name in my zuul.conf i could potentially break a lot of pipelines | 14:52 |
| mnaser | and i guess we operate a lot with the idea of person defining pipelines = person running zuul | 14:52 |
| tobiash | mnaser: with the ha scheduler patches we'll make sql reporting independent of pipelines | 14:52 |
| mnaser | right, that takes care of that aspect but i was generally just thinking from a UI pov when you're defining pipelines with triggers/etc, you need to know what the name of the connections are in zuul.conf | 14:53 |
| tobiash | mnaser: maybe this could be changed to 'all connections of type x' | 14:54 |
| tobiash | but you'd still need different settings for say gerrit and github | 14:55 |
| mnaser | yeah, or something where you alias it inside the tenant config | 14:55 |
| mnaser | like i totally dont have a smart solution for this | 14:55 |
| mnaser | but its something i've been trying to think about solving in a clean way.. | 14:55 |
| mordred | tobiash: I haven't thought it through yet - but I think I like the idea of being able to define "all connections of type x" - because that way pipeline definitions could be shared a bit more easily by people looking for similar workflows without getting caught up in the connection specifics | 14:58 |
| mordred | I'm guessing if you have both public and private github, you're still likely to want a "gate" pipeline to behave the same on both | 14:59 |
| tobiash | yeah, most people anyway add all connections there | 14:59 |
| mordred | yeah | 14:59 |
| tobiash | but there is still the need to specify a specific connection | 14:59 |
| tobiash | so the syntax would get tricky there | 14:59 |
| corvus | opendev has very different pipeline configurations for the same types of connections in different tenants | 15:00 |
| mordred | corvus: yah - I definitely don't think it's a config option that could supllant the current one - even in the same tenant gerrit-review and opendev gerrit behave differently | 15:01 |
| corvus | yeah while it's certainly the case that saying "i want check to behave this way for all gerrits" is common, the counter is not at all abnormal... | 15:01 |
| mordred | yah | 15:01 |
| *** avass has joined #zuul | 15:02 | |
| corvus | mordred, mnaser: i think that would be possible to express (ie, i don't think there's a logical problem with the idea), and the implementation wouldn't be too bad, but is also not exactly a one-liner :) | 15:08 |
| mnaser | i mean i was thinking/considering some sort of pipeline inheritance thingamajig | 15:09 |
| corvus | (i think you'd have to iterate over all the matching connections when constructing the pipeline event filters and add one for each) | 15:09 |
| mnaser | which would be quite similar to how we handle jobs and wouldn't be far off reach of usual consumers for example | 15:09 |
| corvus | mnaser: btw, the list of available connection names is available to the end user on this page: https://zuul.opendev.org/t/zuul/projects | 15:10 |
| mnaser | ah yes | 15:10 |
| mnaser | thats true | 15:10 |
| corvus | (and set aside sql for the moment -- we're solving that by getting it out of pipeline configs) | 15:10 |
| mnaser | perhaps if we added the type of connection too, that would be helpful for users in building their pipelines (obviously some are self-explanitory with names) | 15:10 |
| corvus | yeah that'd be no problem | 15:11 |
| corvus | mnaser: i'm going to mull over pipeline inheritance. i've never thought about that. :) | 15:11 |
| mnaser | :) | 15:11 |
| mnaser | also, i iterated over the 'your job is borked' stuff and it surfaced some other failures in some of our fixtures that we use -- https://review.opendev.org/#/c/708731/ | 15:12 |
| mnaser | if we agree that this is something we want to move forward with then i can go ahead and 'fix' those fixtures to be more sane, but didnt want to go all over them in case the failures actually uncovered a real possible scenario that we missed | 15:12 |
| mnaser | (this is a job with no playbooks + secrets without pass-to-parent:true) | 15:13 |
| corvus | mnaser: cool! i'll look into those and verify | 15:13 |
| mnaser | thank you, corvus | 15:14 |
| *** sgw has joined #zuul | 15:14 | |
| corvus | mnaser: those are all okay to "fix". i believe they all merely omit playbooks to keep the test data simpler. | 15:22 |
| mnaser | corvus: ok cool, i'll iterate and push up a fix shortly | 15:25 |
| mnaser | thanks | 15:25 |
| openstackgerrit | Merged zuul/zuul-jobs master: set jobs for installing openvswitch https://review.opendev.org/708727 | 15:31 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add ensure-bazelisk role https://review.opendev.org/708899 | 15:34 |
| corvus | mordred: ^ 1/2 | 15:34 |
| mordred | corvus: \o/ | 15:35 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add ensure-bazelisk role https://review.opendev.org/708899 | 15:36 |
| corvus | forgot to add the job :) | 15:36 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add ensure-java role https://review.opendev.org/708901 | 15:43 |
| corvus | mordred: ^ 2/2 | 15:43 |
| *** sshnaidm|afk is now known as sshnaidm | 15:43 | |
| mordred | corvus: awesome. I'll work on getting our gerrit jobs changed as soon as I'm done with gitea this morning | 15:44 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Tests ensure-tox on all-platforms https://review.opendev.org/708642 | 15:47 |
| sshnaidm | pabelanger, for building collection role, where do you install ansible? in tox venv or other place? | 15:51 |
| pabelanger | sshnaidm: in virtuelenv | 15:52 |
| sshnaidm | pabelanger, it has configurable path, right? | 15:52 |
| sshnaidm | pabelanger, and ansible version | 15:52 |
| pabelanger | yes, build-collection role we had exposed that | 15:52 |
| pabelanger | let me share | 15:52 |
| sshnaidm | pabelanger, I'd like to use ansible-galaxy from devel now, it has some critical features | 15:53 |
| pabelanger | yes, you can use it | 15:53 |
| sshnaidm | nice | 15:53 |
| pabelanger | we today test with ansible-base, which is migrated content from devel | 15:53 |
| pabelanger | https://github.com/ansible/ansible-zuul-jobs/tree/master/roles/build-ansible-collection | 15:54 |
| pabelanger | that is our role | 15:54 |
| sshnaidm | pabelanger, yeah, I stole it for us :) https://review.opendev.org/#/c/708372/3/ci/playbooks/roles/build-ansible-collection/tasks/main.yml | 15:54 |
| sshnaidm | pabelanger, but didn't see there deps isntall | 15:55 |
| sshnaidm | pabelanger, do you have it in separate role? | 15:55 |
| pabelanger | yah, we do it outside of that role | 15:56 |
| pabelanger | 1 sec | 15:56 |
| pabelanger | https://dashboard.zuul.ansible.com/t/ansible/build/3d8c01cfbfaf4b82ba6d982d2cbcbdf5 | 15:57 |
| pabelanger | is an example job, build-ansible-collection | 15:57 |
| pabelanger | it is a little different, as today it builds _all_ collection dependencies, not just one | 15:57 |
| pabelanger | but, pre-run playbooks setup ansible venv | 15:57 |
| *** rfolco is now known as rfolco|doctor | 16:04 | |
| *** jpena is now known as jpena|off | 16:07 | |
| *** mattw4 has joined #zuul | 16:08 | |
| sshnaidm | pabelanger, so you actually use ansible from tox venv, ok | 16:31 |
| sshnaidm | pabelanger, also wanted to use it for installing various ansible releases.. | 16:32 |
| *** Defolos has quit IRC | 17:08 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Tests ensure-tox on all-platforms https://review.opendev.org/708642 | 17:13 |
| *** jpena|off is now known as jpena | 17:27 | |
| *** evrardjp has quit IRC | 17:34 | |
| *** carli has quit IRC | 17:34 | |
| *** evrardjp has joined #zuul | 17:35 | |
| *** chandankumar is now known as raukadah | 18:14 | |
| *** jpena is now known as jpena|off | 18:34 | |
| *** igordc has joined #zuul | 18:34 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add ensure-bazelisk role https://review.opendev.org/708899 | 18:35 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add ensure-java role https://review.opendev.org/708901 | 18:36 |
| mnaser | how much people would i piss off if i split off tests_v3 :) | 18:43 |
| corvus | mnaser: if you can figure out a good organizational structure, it may be welcome. | 18:45 |
| fungi | since i'm not sure what tests_v3 is, not me at least! ;) | 18:45 |
| mnaser | i just worry that it will merge conflict the heck of everyone | 18:45 |
| corvus | fungi: most of zuul's tests are in 2 files. as clarkb describes it, "test_scheduler" is "test all the things" and "test_v3" is "test all the rest of the things" | 18:45 |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: configloader: validate playbook-less + secret jobs https://review.opendev.org/708731 | 18:46 |
| mnaser | my vim runs pylint post-save but its doing it non asynchronously so it destroys my machine on every save :p | 18:47 |
| clarkb | testv3 was where the v3 new behavior stuff got tested and test_scheduler is where the bulk of testing has lived and was retrofitted. Generally those files don't tend to merge conflict too much though as people work in specific areas of them (that said organizing a bit more would probably be worthwhile) | 18:47 |
| fungi | i'm not getting any hits out of `git ls-files '*tests_v3*'` in zuul/zuul | 18:47 |
| clarkb | fungi: test_v3 iirc no extra 's' | 18:47 |
| fungi | aha, tests/unit/test_v3.py thanks! | 18:47 |
| *** saneax has quit IRC | 18:55 | |
| tobiash | mnaser: it will conflict with all I have, but I'd still appreciate that ;) | 18:56 |
| mnaser | i think i might have a reaosnable step #1 | 18:56 |
| tobiash | actually I'd like to split some more files (e.g. scheduler, executor server) as I find it still hard to navigate through those files with >2500 lines | 18:58 |
| tobiash | but I don't know what the general feeling in this regard is so I didn't so far | 18:59 |
| tobiash | e.g. AnsibleJob could be moved into its own file | 19:00 |
| openstackgerrit | Mohammed Naser proposed zuul/zuul master: cleanup: move AnsibleZuulTestCase to test_ansible https://review.opendev.org/708952 | 19:06 |
| mnaser | corvus, tobiash: here is a small thing we can do (passed pep8 locally) | 19:07 |
| *** plaurin has joined #zuul | 19:10 | |
| tobiash | mnaser: I think we shouldn't judge on the base class but more on the use case/component that get's tested | 19:10 |
| mnaser | tobiash: i agree, but this seemed like a small first iteration of slowly splitting things | 19:10 |
| tobiash | ok, I think it makes sense of a first step to make the file smaller, however we probably should think about a general structure | 19:13 |
| *** rlandy is now known as rlandy|mtg | 19:16 | |
| mnaser | i agree | 19:16 |
| mnaser | hmm | 19:26 |
| mnaser | corvus: there seems to be a behaviour change in my change here https://review.opendev.org/#/c/708731/ in the unit tests - https://ad793cc4ef661bc84354-a22e2178400a1d74c0dfc0d0570ba9cf.ssl.cf5.rackcdn.com/708731/8/check/tox-py37/e3057dc/testr_results.html | 19:27 |
| mnaser | maybe i did something wrong :\ but im not sure why adding 'run' into the playbook made the secret come from parent instead of the job itself | 19:28 |
| mnaser | even tho that wasnt a pass-to-parent, so i'm a bit confused | 19:28 |
| *** rlandy|mtg is now known as rlandy | 19:34 | |
| *** Defolos has joined #zuul | 19:35 | |
| plaurin | Hello irc people ;-) | 19:36 |
| plaurin | Hey about the discussions we had the other days about kubernetes related srdout logging, hum, I'm not used to participate that much in open source communities (very limited experience...) Should I be reporting somehting like a bug or feature request somewhere? | 19:37 |
| plaurin | Also then I can maybe follow the progress on this an help in any way I can, by testing fixes or else | 19:38 |
| *** igordc has quit IRC | 19:38 | |
| clarkb | a story on storyboard would probably help. (https://storyboard.openstack.org/#!/story/list then click create story) you'll need an ubuntu one account for that | 19:40 |
| corvus | plaurin: if you do that, you could add this link to the story: http://eavesdrop.openstack.org/irclogs/%23zuul/%23zuul.2020-02-03.log.html#t2020-02-03T19:30:52 | 19:43 |
| corvus | then we'll remember what we need to do :) | 19:43 |
| plaurin | okay will do, yeah I have my ubuntu one account somewhere...... :P | 19:44 |
| plaurin | thx | 19:44 |
| *** jamesmcarthur has quit IRC | 19:45 | |
| plaurin | What would you recommend as a story name? | 19:45 |
| plaurin | Kubernetes logging in zuul? | 19:45 |
| corvus | plaurin: "Kubernetes log streaming" | 19:46 |
| corvus | plaurin: the 'streaming' keyword helps narrow down what kind of logging we're talking about | 19:46 |
| corvus | though i guess we also discovered that the stdout isn't working correctly | 19:47 |
| corvus | so maybe "kubernetes command output" | 19:47 |
| plaurin | shouldn't zuul be a keyword as well? | 19:47 |
| corvus | plaurin: the story will be against the zuul project, so it's implied | 19:47 |
| corvus | (or, well, the task will be, but that's enough) | 19:47 |
| plaurin | okay, hmm.. how is it implied that this is for zuul? (sorry that may be dumb from your perspective... :P ) | 19:48 |
| clarkb | plaurin: when you create the story one of the fields is the project which you should enter 'zuul/zuul' for | 19:49 |
| plaurin | oh, got it | 19:49 |
| plaurin | private? | 19:57 |
| fungi | public | 19:57 |
| fungi | private is mainly for reporting suspected security vulnerabilities | 19:57 |
| plaurin | makes sense | 19:57 |
| fungi | just so that we might have a chance of fixing them before more folks find out and can exploit them | 19:58 |
| plaurin | yes, that's a good idea | 19:58 |
| plaurin | okay I submitted the story! | 19:58 |
| fungi | thanks plaurin! looks like it's https://storyboard.openstack.org/#!/story/2007321 | 20:00 |
| plaurin | yep! | 20:00 |
| plaurin | ah, I should add the link above to the story | 20:01 |
| plaurin | the eavesdrop url, done | 20:03 |
| fungi | perfect, thanks again! | 20:06 |
| plaurin | My first report on story board hehe \o/ thx to you | 20:08 |
| mordred | plaurin: thanks you! | 20:31 |
| plaurin | anytime | 20:32 |
| *** bhavikdbavishi has quit IRC | 20:39 | |
| *** armstrongs has joined #zuul | 21:07 | |
| *** armstrongs has quit IRC | 21:11 | |
| *** Goneri has quit IRC | 21:25 | |
| *** igordc has joined #zuul | 21:28 | |
| *** plaurin has quit IRC | 21:40 | |
| *** dpawlik has quit IRC | 22:33 | |
| *** sshnaidm is now known as sshnaidm|afk | 23:02 | |
| *** Defolos has quit IRC | 23:38 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!