| *** jamesmcarthur has joined #zuul | 01:06 | |
| jamesmcarthur | corvus: and mordred - a quick question on availability and slide deck for the Wednesday meeting. If either of you have a chance to ping me ahead of Wednesday.... | 01:12 |
|---|---|---|
| *** jamesmcarthur has quit IRC | 01:16 | |
| *** swest has quit IRC | 01:21 | |
| *** swest has joined #zuul | 01:37 | |
| *** jamesmcarthur has joined #zuul | 01:42 | |
| *** bhavikdbavishi has joined #zuul | 01:47 | |
| *** jamesmcarthur has quit IRC | 01:54 | |
| *** bhavikdbavishi1 has joined #zuul | 01:59 | |
| *** bhavikdbavishi has quit IRC | 01:59 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 01:59 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: Add ansible-lint job https://review.opendev.org/532083 | 03:02 |
| *** sajauddin has joined #zuul | 05:09 | |
| sajauddin | Can anyone help me configuring ZuulV3? I have installed ZuulV3 and trying to configure with various dependent modules like nodepool, ansible playbooks etc | 05:12 |
| sajauddin | I am following Zuul official documentation, which is a bit confusing for me. If anyone has other docs or links, please share | 05:13 |
| *** quiquell has joined #zuul | 06:09 | |
| *** quiquell is now known as quiquell|rover | 06:09 | |
| SpamapS | sajauddin:we'll need more specific questions before we can answer. What is confusing? Where are you stuck? | 06:10 |
| sajauddin | I am trying to setup a test/sample pipeline and job. I could not figure out from where to start. From where should I pull base ansible playbooks (zuul-jobs/zuul-openstack-jobs) and where these should be kept. | 06:14 |
| sajauddin | Looking for any sample reference pipeline/job definition, which has end to end implementation. | 06:15 |
| SpamapS | sajauddin:Are you following the quick start? | 06:17 |
| SpamapS | https://zuul-ci.org/docs/zuul/admin/quick-start.html | 06:18 |
| sajauddin | Spamaps: Refering https://zuul-ci.org/docs/zuul/admin/zuul-from-scratch.html and https://zuul-ci.org/docs/zuul/admin/installation.html | 06:22 |
| sajauddin | We already have gerrit and nodepool already available in our environment | 06:22 |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: Pagure driver - https://pagure.io/pagure/ https://review.opendev.org/604404 | 06:33 |
| *** quiquell|rover is now known as quique|rover|brb | 06:46 | |
| *** saneax has joined #zuul | 06:58 | |
| *** sshnaidm|off is now known as sshnaidm | 07:00 | |
| *** quique|rover|brb is now known as quiquell|rover | 07:02 | |
| *** sanjayu_ has joined #zuul | 07:05 | |
| *** saneax has quit IRC | 07:07 | |
| *** yolanda_ has joined #zuul | 07:18 | |
| *** pcaruana has joined #zuul | 07:19 | |
| *** yolanda_ has quit IRC | 07:52 | |
| *** pcaruana has quit IRC | 08:45 | |
| *** zxiiro-away is now known as zxiiro | 08:48 | |
| *** sanjayu__ has joined #zuul | 08:58 | |
| *** sanjayu_ has quit IRC | 08:59 | |
| *** sanjayu__ has quit IRC | 09:00 | |
| *** sanjayu_ has joined #zuul | 09:00 | |
| *** quiquell|rover is now known as quiquell|ruck | 10:21 | |
| quiquell|ruck | hello | 10:29 |
| quiquell|ruck | Is there any way to pin to one of the docker containers here https://hub.docker.com/u/zuul ? | 10:30 |
| SpamapS | quiquell|ruck:those are really just for learning purposes. If you want to get serious, you will probably want to build your own from the git repo. | 10:48 |
| quiquell|ruck | SpamapS: it's no longer pbrx ? | 10:49 |
| *** pcaruana has joined #zuul | 10:58 | |
| quiquell|ruck | SpamapS: I see multi-stage Dockerfile | 11:29 |
| mordred | quiquell|ruck: that's right - we're no longer using pbrx | 11:49 |
| mordred | quiquell|ruck: although those are not just for learning purposes - we fully intend to start deploying from those in opendev | 11:49 |
| mordred | quiquell|ruck: that said, we haven't started tagging them when we tag the git repo - so there is currently no way to pin to a released version | 11:50 |
| mordred | quiquell|ruck: so for the time being, if you want to do that, you should follow SpamapS advice and build your own from git | 11:50 |
| quiquell|ruck | mordred: yep doing it, doesn't take too much, thanks to both | 11:51 |
| mordred | sweet! | 11:51 |
| quiquell|ruck | mordred: Do you know if this is a limitation of container build jobs ? I mean missing tags | 11:51 |
| mordred | quiquell|ruck: the container build jobs know how to tag things generally, but we haven't set one up on the release pipeline yet ... and it's possible when we dig in to that we might discover we need to do something with the job to improve it | 11:53 |
| mordred | quiquell|ruck: which is to say - I'm not sure :) | 11:53 |
| quiquell|ruck | mordred: maybe we try to use those jobs but pushing to our registry and using the taggin mechanism | 11:54 |
| quiquell|ruck | mordred: will let you know | 11:55 |
| mordred | quiquell|ruck: please do! | 11:55 |
| quiquell|ruck | humm stupid question | 11:55 |
| quiquell|ruck | do you know if there is a limit for tags at docker hub ? | 11:55 |
| quiquell|ruck | I suppose it will be a limit for layres | 11:56 |
| mordred | I do not know | 12:03 |
| *** rlandy has joined #zuul | 12:06 | |
| *** rlandy is now known as rlandy|ruck | 12:06 | |
| *** quiquell|ruck is now known as quiquell|rover | 12:12 | |
| AJaeger | anybody wants to review this pbrx OpenDev change, please? | 12:21 |
| *** quiquell|rover is now known as quique|rover|eat | 12:22 | |
| mordred | AJaeger: done | 12:29 |
| AJaeger | thanks | 12:30 |
| *** bhavikdbavishi1 has joined #zuul | 12:33 | |
| *** bhavikdbavishi has quit IRC | 12:34 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 12:34 | |
| *** rlandy|ruck is now known as rlandy|ruck|mtg | 12:36 | |
| *** sajauddin has quit IRC | 12:38 | |
| openstackgerrit | Merged x/pbrx master: Fix playbooks/container-images/pre.yaml after OpenDev https://review.opendev.org/654260 | 12:38 |
| *** rlandy|ruck|mtg is now known as rlandy|ruck | 12:42 | |
| *** jamesmcarthur has joined #zuul | 12:42 | |
| *** quique|rover|eat is now known as quiquell|rover | 12:52 | |
| *** rf0lc0 is now known as rfolco | 13:02 | |
| *** rlandy|ruck is now known as rlandy|ruck|mtg | 13:08 | |
| quiquell|rover | mordred: Ansible is not install at the Dockerfile from zuul repos ? | 13:09 |
| openstackgerrit | Merged zuul/zuul-preview master: Use splice instead of erase/push_front https://review.opendev.org/652148 | 13:09 |
| *** jamesmcarthur has quit IRC | 13:14 | |
| *** tflink has quit IRC | 13:34 | |
| *** tflink has joined #zuul | 13:34 | |
| *** jvv_ has joined #zuul | 13:56 | |
| jvv_ | Hello, I am trying to setup my first Zuul instance. I've been able to setup zookeeper and nodepool, however I am experiencing problems while starting zuul-scheduler. It is complainig due to lack of /var/lib/zuul/scheduler.socket. Does anyone know what should create this socket? Many thanks. | 13:58 |
| *** rlandy|ruck|mtg is now known as rlandy|ruck | 13:58 | |
| fungi | jvv_: the scheduler should create that... perhaps the user as which it's running lacks write access to /var/lib/zuul ? | 14:00 |
| *** jamesmcarthur has joined #zuul | 14:03 | |
| corvus | quiquell|rover: ansible is not installed at the root level in the docker image; it's installed in the virtualenvs in the executor docker image | 14:06 |
| quiquell|rover | corvus: yep just found out ack | 14:10 |
| jvv_ | fungi: The error message received from scheduler says "No such file or directory" rather than "Permission denied". Its complaining on connect() that socket not there. | 14:11 |
| fungi | jvv_: do you get a python traceback from that? if so, mind pasting it to http://paste.openstack.org/ and letting us know the url? helps to have some context as to what file/function is raising that exception | 14:12 |
| corvus | jvv_: the output of 'ls -la /var/lib/zuul' might help too | 14:13 |
| jvv_ | fungi: sure, here you go: http://paste.openstack.org/show/749596 | 14:15 |
| quiquell|rover | Humm after the opendev migration I have a zuul connecting to it and fails here https://opendev.org/zuul/zuul/src/branch/master/zuul/configloader.py#L1672 | 14:21 |
| quiquell|rover | with https://paste.fedoraproject.org/paste/UokZRGe3tbt3lQ8ZUwLXLQ | 14:21 |
| corvus | quiquell|rover: if you look in the logs on your mergers (or executors) you may find a better error message | 14:26 |
| quiquell|rover | corvus: ack | 14:26 |
| *** openstackgerrit has quit IRC | 14:28 | |
| fungi | jvv_: is there anything about the way or platform on which you've deployed the scheduler which could be silently preventing the creation of named pipes? the zuul.scheduler.Scheduler.__init__() function should be creating that fifo on instantiating the class | 14:29 |
| corvus | jvv_: it looks like 'zuul-scheduler' is being invoked with a command argument, so that it is trying to act as a client to send a command to a running server, but there is no running server. | 14:29 |
| fungi | oh, hah, i didn't pick up on that | 14:30 |
| corvus | jvv_: for example, running "zuul-scheduler full-reconfigure" will tell a running scheduler to reconfigure itself | 14:30 |
| fungi | but yes, seems it's calling into zuul.cmd.__init__() so that's the rpc client | 14:30 |
| corvus | jvv_: so maybe look at the startup file (i guess this is systemd, so the systemd unit file) and see what the command it's using is | 14:31 |
| jvv_ | I just copied systemd and zuul.conf files from zuul-ci.org docs :( | 14:31 |
| fungi | it should have dawned on me that the scheduler doesn't call send_command() to its own socket | 14:31 |
| jvv_ | full-reconfigure throws the same error | 14:32 |
| corvus | jvv_: https://opendev.org/zuul/zuul/src/branch/master/etc/zuul-scheduler.service.d/centos.conf this file? | 14:32 |
| corvus | jvv_: or this ? https://opendev.org/zuul/zuul/src/branch/master/etc/zuul-scheduler.service | 14:32 |
| jvv_ | the first one, however it was not working properly, so I ended up with similar version that the second one you gave. let me just try to repaste and try again. | 14:34 |
| fungi | jvv_: you can find out details of the invoking systemd unit with the `systemctl show zuul-scheduler` command; the ExecStart= line is probably relevant... is it calling /etc/init.d/zuul-scheduler or something else? | 14:34 |
| jvv_ | it looks like this: http://paste.openstack.org/show/749598/ | 14:38 |
| corvus | jvv_: i like fungi's idea of using `systemctl show zuul-scheduler` so if systemd did something unexpected (what with all the overlay files, etc), hopefully it will show us | 14:38 |
| jvv_ | /zuul-venv is a python venv I've installed it in... what bugs me is that /var/log/zuul/scheduler.log is juz empty file :( | 14:40 |
| jvv_ | systemctl show zuul-scheduler shows same thing I've pasted here http://paste.openstack.org/show/749596/ | 14:41 |
| quiquell|rover | corvus: executor/merger is waiting for scheduler but scheduler goes down so the issue is not at merger or executor | 14:41 |
| jvv_ | I dont know how to debug this :( | 14:41 |
| jvv_ | (selinux it tured off) | 14:42 |
| fungi | `systemctl show zuul-scheduler` shows a python traceback? | 14:42 |
| *** quiquell|rover is now known as quiquell|off | 14:42 | |
| jvv_ | oh, sorry... you wrote 'show' I read 'status'... let me check | 14:43 |
| fungi | jvv_: well, the ExecStart= line presumably looks something like what's in that unit file too | 14:43 |
| corvus | this is what our 'show' looks like: http://paste.openstack.org/show/749599/ | 14:43 |
| fungi | yeah, in our case we're using systemctl's sysvinit compat layer to call an initscript | 14:44 |
| corvus | quiquell|off: the error happens when the scheduler asks a merger (executors are mergers) to run a job but doesn't get anything back. there isn't enough history in the log you pasted, but if that error happened 5 minutes after it submitted the job, and there are no logs about the job on the mergers, then it means that it timed out. that may mean that the mergers are unable to talk to the gearman server. | 14:45 |
| jvv_ | in my case it looks like this http://paste.openstack.org/show/749600 | 14:47 |
| jvv_ | I dont see anything wonky | 14:47 |
| quiquell|off | corvus: there were no jobs just tenant starting up, at the moment were is loading config from gerrits | 14:49 |
| corvus | jvv_: the only thing i can think of is that somehow that's the result of a stop command and not start.... | 14:49 |
| corvus | jvv_: can you run 'systemctl stop zuul-scheduler' (i expect that to fail) then 'systemctl start zuul-scheduler' and see if the output is any different? | 14:49 |
| corvus | (hopefully that will convince systemd it is really stopped, and we can start over) | 14:50 |
| corvus | quiquell|off: i didn't mean a zuul job, i meant a gearman job. zuul's scheduler communicates with mergers via gearman jobs. | 14:50 |
| quiquell|off | ahh ok | 14:50 |
| quiquell|off | Ok will check tomorroiws | 14:51 |
| quiquell|off | corvus: thanks again | 14:51 |
| jvv_ | corvus: looks the same | 14:53 |
| jvv_ | ok, but why is it trying to connect() when it didnt create the pipe? | 14:53 |
| corvus | jvv_: because the server hasn't started yet because, for some unknown reason, it is being invoked with a command argument rather than no arguments. | 14:55 |
| corvus | jvv_: are you installing zuul from the git master branch, or a release? | 15:01 |
| jvv_ | corvus: sorry, I am lost now... what 'command argument' do you mean? | 15:02 |
| jvv_ | corvus: I've installed it from PyPi | 15:03 |
| jvv_ | 3.8.0 version | 15:03 |
| corvus | jvv_: i have no idea, sorry. systemd doesn't show one, but the only way zuul-scheduler would behave in the way you describe is if there were one. there is something very strange happening, and i don't know what it is. if you have time for more debugging, i can make some suggestions. | 15:04 |
| fungi | one thing which may be worth trying is to su to the zuul user (assuming it has a viable shell set) and then try invoking /zuul-venv/bin/zuul-scheduler on the command line | 15:04 |
| corvus | that's a good idea | 15:05 |
| fungi | that will at least cut systemd out of the picture temporarily so it can either be ruled out or identified as the cause | 15:05 |
| corvus | hrm... here's a weird thing about that output | 15:07 |
| jvv_ | yeah, tried /zuul-venv/bin/zuul-scheduler by hand / without systemd ... but it just dies, no log trail, no console output, nothing | 15:07 |
| corvus | it says that the execstop command ran after the exectstart | 15:07 |
| corvus | that output *is* consistent with running 'zuul-executor stop' after zuul failing to start | 15:07 |
| fungi | aha, so maybe what we're looking at is zuul silently failing to start | 15:08 |
| fungi | and the traceback is the stop rpc subcommand breaking because the scheduler wasn't running | 15:08 |
| corvus | jvv_: try running "/zuul-venv/bin/zuul-scheduler -d" | 15:08 |
| corvus | fungi: yeah, that's what i'm starting to think | 15:09 |
| fungi | possible as a future improvement we could have zuul-scheduler stop print something useful when there is no command socket, rather than just falling over | 15:09 |
| corvus | certainly :) | 15:11 |
| *** bhavikdbavishi has quit IRC | 15:11 | |
| jvv_ | /zuul-venv/bin/zuul-scheduler -c /etc/zuul/zuul.conf -d complains about zookeeper hosts missing | 15:12 |
| fungi | aha, yes, as of a recent release you now must specify the zk hosts in configuration. there should have been a release note about that | 15:13 |
| fungi | i wonder if we missed updating an example config? | 15:13 |
| jvv_ | ha! | 15:14 |
| jvv_ | works! | 15:14 |
| fungi | seems that was as of zuul 3.7.0: https://zuul-ci.org/docs/zuul/releasenotes.html#relnotes-3-7-0 | 15:14 |
| fungi | jvv_: which example config did you use? we need to correct it | 15:15 |
| jvv_ | so this 'connect thing' was confusing me... the problem was missing zookeeper stanza | 15:15 |
| fungi | yes, i think better output from invoking `zuul-scheduler stop` when there's no socket would have significantly shortened the time to debug | 15:16 |
| jvv_ | https://zuul-ci.org/docs/zuul/admin/zuul-from-scratch.html | 15:16 |
| fungi | thanks! | 15:16 |
| fungi | and you set it to localhost:2181 presumably to get it to work? | 15:16 |
| jvv_ | didnt give the port even... just [zookeeper]\nhosts=localhost | 15:17 |
| jvv_ | thats all | 15:17 |
| fungi | got it | 15:17 |
| fungi | i'll try to work up quick patches for those two items | 15:18 |
| jvv_ | many thanks | 15:19 |
| fungi | more thanks to you for helping us spot two things we need to fix! ;) | 15:20 |
| AJaeger | here's an OpenDev related change for zuul-website repo, another core review is welcome: https://review.opendev.org/#/c/654223/ | 15:20 |
| *** openstackgerrit has joined #zuul | 15:22 | |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul master: Include zookeeper.hosts in ZFS example zuul.conf https://review.opendev.org/654483 | 15:22 |
| fungi | jvv_: there's one ^ | 15:22 |
| openstackgerrit | Merged zuul/zuul-website master: Update review.openstack.org URLs https://review.opendev.org/654223 | 15:32 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul master: Log an error when command socket is missing https://review.opendev.org/654488 | 15:33 |
| fungi | jvv_: and there's the second ^ | 15:33 |
| fungi | i'm unsure whether the logger is viable in that case, may be i need to augment the exception text instead? at any rate, our esteemed reviewers will know | 15:35 |
| *** jamesmcarthur has quit IRC | 15:50 | |
| *** openstackgerrit has quit IRC | 15:58 | |
| *** openstackgerrit has joined #zuul | 16:08 | |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Only alias docker registries on test nodes https://review.opendev.org/654555 | 16:08 |
| *** jamesmcarthur has joined #zuul | 16:21 | |
| *** sshnaidm is now known as sshnaidm|afk | 16:22 | |
| *** jamesmcarthur_ has joined #zuul | 16:47 | |
| *** jamesmcarthur has quit IRC | 16:50 | |
| *** chandankumar is now known as raukadah | 16:56 | |
| *** ericbarrett has joined #zuul | 17:05 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Only alias docker registries on test nodes https://review.opendev.org/654555 | 17:15 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Only alias docker registries on test nodes https://review.opendev.org/654555 | 17:18 |
| *** nickx-intel has joined #zuul | 17:25 | |
| nickx-intel | morning | 17:26 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Only alias docker registries on test nodes https://review.opendev.org/654555 | 17:37 |
| *** mhu has quit IRC | 17:48 | |
| *** mhu has joined #zuul | 17:48 | |
| *** mattw4 has joined #zuul | 17:51 | |
| SpamapS | mordred:FYI, I didn't mean to say the dockerhub images aren't of production quality, but only that without tags, they're not really *useful* for production. | 17:51 |
| SpamapS | which I think you said too, but just wanted to be clear, they're great images.. just.. I'm not sure anybody should use them beyond learning. | 17:52 |
| clarkb | SpamapS: tags for releases or? | 17:52 |
| SpamapS | If you can't say FROM zuul:3.8.0 ... you're *screwed* whenever zuul 3.9.0 is tagged and you haven't tested it and your next config change pulls in new software without your permission. | 17:53 |
| SpamapS | The change_ tags are ok.. but not for mere mortals. | 17:54 |
| clarkb | I think that should be a simple addition once we've got the jobs running reliably with ipv6 (for which we hope we've got a fix for now) | 17:55 |
| SpamapS | Actually I may flip to the change_ tags .. been meaning to try that out but I'm also not entirely sure they're going to be there forever. | 17:55 |
| clarkb | and can possibly even go back and tag some more recent releases? | 17:55 |
| *** zbr is now known as zbr|pto | 17:56 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Only alias docker registries on test nodes https://review.opendev.org/654555 | 18:10 |
| jvv_ | btw, is there some comparison article zuul vs. jenkins? I make a case in my organization for zuul. | 18:39 |
| jvv_ | I would like to make...* | 18:39 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Don't use underscore in buildset_registry_alias https://review.opendev.org/654603 | 18:40 |
| clarkb | jvv_: I don't thnik we've written such a document. I think there are/were palns to write a "things that make zuul stand out" document without direct comparison to any other CI tools. But I don't know that that has been done either | 18:42 |
| clarkb | corvus: ^ do you recall? | 18:42 |
| corvus | SpamapS: the change_ tags are not supposed to be there forever, they get auto-pruned after 24 hours, possible sooner | 18:43 |
| *** jamesmcarthur_ has quit IRC | 18:44 | |
| corvus | jvv_, clarkb: there's this: https://zuul-ci.org/media/zuul_solution_brief.pdf | 18:44 |
| clarkb | ah yup the why zuul bullet list in particular is probably a good start | 18:44 |
| *** jamesmcarthur has joined #zuul | 18:48 | |
| *** jamesmcarthur has quit IRC | 18:50 | |
| *** jamesmcarthur has joined #zuul | 18:50 | |
| jvv_ | ok, many thanks | 18:53 |
| openstackgerrit | Merged zuul/zuul-jobs master: Don't use underscore in buildset_registry_alias https://review.opendev.org/654603 | 18:53 |
| *** dkehn has joined #zuul | 19:59 | |
| *** sshnaidm|afk has quit IRC | 20:16 | |
| *** rlandy|ruck is now known as rlandy|ruck|biab | 20:26 | |
| *** pcaruana has quit IRC | 20:46 | |
| SpamapS | jvv_:I've run in a high-throughput hybrid Jenkins + Zuul environment. They're not entirely in competition. | 21:00 |
| SpamapS | Certainly in some areas, but not all. | 21:00 |
| jvv_ | SpamapS: does Zuul have all the features of Jenkins? | 21:02 |
| SpamapS | jvv_:no, and Jenkins doesn't have all of the features of Zuul. | 21:03 |
| fungi | zuul and jenkins attempt to solve different sorts of problems, so there is some overlap but also each has features the other does not | 21:03 |
| SpamapS | Which is one reason a side-by-side isn't really a thing. | 21:03 |
| *** zbr|pto has quit IRC | 21:03 | |
| SpamapS | Jenkins is more of a fancy integration middle-man. Ansible, which Zuul leverages, has as much overlap with Jenkins as Zuul does. | 21:04 |
| fungi | and depending on how you look at it, the choice not to implement certain features can also be seen as a feature | 21:04 |
| SpamapS | fungi: ^^^ so much this jvv_ | 21:04 |
| *** mnaser has quit IRC | 21:04 | |
| SpamapS | Some things that Jenkins lets you do are actually detrimental to the mission of Zuul. | 21:04 |
| jvv_ | can you give an example, please? | 21:05 |
| fungi | graphical job configuration interface with a non-human-readable (effectively, unless you *really* like xml) storage backend for them | 21:05 |
| *** tima has quit IRC | 21:05 | |
| *** tima has joined #zuul | 21:07 | |
| fungi | another would be its choice for prioritizing extensibility over security, to the point where they recommend you don't expose it to the internet for safety reasons | 21:07 |
| *** mnaser has joined #zuul | 21:08 | |
| jvv_ | ok | 21:11 |
| jvv_ | for now I need to understand Zuul really. It has significant amount of moving parts which I need to get familiar with | 21:12 |
| jvv_ | now, gerrit time. | 21:13 |
| fungi | but generally speaking it's that zuul's design focuses on solving different problems, as i said. it's very tied to code review workflows (gerrit, github, ...) and revision control (git) systems, leverages existing remote execution mechanisms (ansible) rather than providing its own, has a focus on testing and deploying software rather than general-purpose automation, is designed to be horizontally scalable, | 21:13 |
| fungi | provides built-in resource management for job execution environments... | 21:13 |
| fungi | SpamapS: what config change did you make, exactly, to end up with errors finding the zuul-jobs repo? | 21:15 |
| fungi | updated line 15 of zuul-base-jobs zuul.yaml from git.zuul-ci.org/zuul-jobs to opendev.org/zuul/zuul-jobs ? | 21:16 |
| fungi | oh, you're saying you changed your zuul scheduler config to opendev.org | 21:17 |
| fungi | so a [connection "something"] section in your zuul.conf? | 21:18 |
| fungi | what was the original server line, and what did you try replacing it with? | 21:18 |
| SpamapS | fungi: correct | 21:18 |
| SpamapS | Loading diff ... | 21:19 |
| SpamapS | .... | 21:19 |
| fungi | but yeah, i can see where you would presumably need a connection matching the domain name in the base job | 21:19 |
| fungi | again our deployment didn't hit this because we don't actually consume the zuul-base-jobs base job | 21:20 |
| SpamapS | http://paste.openstack.org/show/749615/ | 21:20 |
| SpamapS | fungi:always happy to be the vanguard for users :) | 21:21 |
| SpamapS | Well, not happy. But willing. | 21:21 |
| fungi | you're so good at it! | 21:22 |
| SpamapS | I didn't try making a new connection, and listing them both in the tenant. | 21:23 |
| SpamapS | Which was going to be my next experiment before I ran out of time. | 21:23 |
| fungi | and yeah, i wonder whether the base job should be a relative reference there, though that then becomes a problem for projects which want baseurl as just https://opendev.org/ with no namespace restriction | 21:24 |
| fungi | i guess if a project is consuming the zuul-base-jobs and zuul-jobs repos from the same connection, then the reference from the former to the latter can be relative? any guess on possible down-sides? | 21:25 |
| SpamapS | I don't know how you can make it relative. | 21:35 |
| SpamapS | So I'm not sure I follow. | 21:35 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Fix localhost cert location for push-to-intermediate-registry https://review.opendev.org/654632 | 21:36 |
| SpamapS | fungi:Perhaps an alias setting for connections. | 21:37 |
| SpamapS | so that during a transition period folks can have one connection that satisfies both namespaces. | 21:38 |
| fungi | SpamapS: meaning setting the roles to include as just zuul-jobs instead of git.zuul-ci.org/zuul-jobs in the base job... that will cause it to look for zuul-jobs via the same connection, right? | 21:39 |
| fungi | (same connection as the zuul-base-jobs repo) | 21:39 |
| SpamapS | fungi:ah right, ok. That would work. I'm sure there's a reason it's been done the other way though. | 21:39 |
| fungi | the reason very well may be "because fungi didn't know any better when he made that repo" | 21:41 |
| fungi | i put the base job together with some fairly naive assumptions | 21:41 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Fix localhost cert location for push-to-intermediate-registry https://review.opendev.org/654632 | 21:43 |
| fungi | looks like i switched it from openstack-infra/zuul-jobs to git.zuul-ci.org/zuul-jobs in https://review.opendev.org/599607 with convention as the only indicated argument | 21:44 |
| fungi | i suppose if we make it relative, how we do that will depend on what baseurl we recommend for the connection | 21:48 |
| fungi | whether it should be a baseurl of opendev.org with roles at zuul/zuul-jobs or a baseurl of opendev.org/zuul with roles in zuul-jobs | 21:49 |
| fungi | the former has the benefit of basically mirroring what you would need to set in a gerrit connection if you wanted to consume it from review.opendev.org instead | 21:50 |
| fungi | i don't think the latter could be made to work with a gerrit connection? | 21:51 |
| *** manjeets_ has joined #zuul | 21:58 | |
| *** johanssone has quit IRC | 21:58 | |
| *** mgoddard has quit IRC | 21:58 | |
| *** panda has quit IRC | 21:58 | |
| *** odyssey4me has quit IRC | 21:58 | |
| *** mgoddard has joined #zuul | 21:59 | |
| *** manjeets has quit IRC | 21:59 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Fix localhost cert location for push-to-intermediate-registry https://review.opendev.org/654632 | 21:59 |
| *** irclogbot_2 has quit IRC | 21:59 | |
| *** panda has joined #zuul | 22:00 | |
| *** odyssey4me has joined #zuul | 22:00 | |
| *** rlandy|ruck|biab is now known as rlandy|ruck | 22:00 | |
| *** johanssone has joined #zuul | 22:02 | |
| *** irclogbot_0 has joined #zuul | 22:03 | |
| *** tima has quit IRC | 22:08 | |
| *** tima has joined #zuul | 22:09 | |
| *** tjgresha_nope has joined #zuul | 22:10 | |
| *** irclogbot_0 has quit IRC | 22:11 | |
| *** smyers_ has joined #zuul | 22:12 | |
| *** kinrui has joined #zuul | 22:14 | |
| *** smyers has quit IRC | 22:16 | |
| *** nhicher has quit IRC | 22:16 | |
| *** fbo has quit IRC | 22:16 | |
| *** fungi has quit IRC | 22:16 | |
| *** tjgresha has quit IRC | 22:16 | |
| *** SpamapS has quit IRC | 22:16 | |
| *** smyers_ is now known as smyers | 22:16 | |
| *** kinrui is now known as fungi | 22:19 | |
| *** SpamapS has joined #zuul | 22:19 | |
| *** irclogbot_3 has joined #zuul | 22:21 | |
| *** sshnaidm has joined #zuul | 22:24 | |
| *** sshnaidm is now known as sshnaidm|afk | 22:24 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Fix another opendev rename https://review.opendev.org/654643 | 22:33 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Don't log setting up creds for the intermediate registry https://review.opendev.org/654644 | 22:55 |
| *** threestrands has joined #zuul | 23:04 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Update references for opendev https://review.opendev.org/654238 | 23:15 |
| corvus | we believe we've fixed the issue with the docker jobs when they ran on an ipv6-only provider -- so hopefully image bauild and quick-start jobs should be more reliable now | 23:16 |
| clarkb | with the power of socat | 23:16 |
| corvus | as long as some kind of cat is involved | 23:17 |
| openstackgerrit | Merged zuul/zuul-jobs master: Don't log setting up creds for the intermediate registry https://review.opendev.org/654644 | 23:18 |
| fungi | socat watches you connect | 23:38 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!