openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: executor: use node python path https://review.openstack.org/637339 | 00:03 |
---|---|---|
openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool master: Add python-path option to node https://review.openstack.org/637338 | 00:12 |
tristanC | corvus: i have seen it though it's pretty low priority for me atm. should i work on it sooner? | 00:14 |
*** jamesmcarthur has quit IRC | 00:27 | |
openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: executor: enable zuul_return to update Ansible inventory https://review.openstack.org/590092 | 00:38 |
*** daniel2 has joined #zuul | 00:52 | |
*** jamesmcarthur has joined #zuul | 00:58 | |
*** jamesmcarthur has quit IRC | 01:15 | |
*** jamesmcarthur has joined #zuul | 01:16 | |
*** jamesmcarthur has quit IRC | 01:39 | |
*** jamesmcarthur has joined #zuul | 01:40 | |
*** jamesmcarthur has quit IRC | 01:44 | |
*** jamesmcarthur has joined #zuul | 02:35 | |
*** jamesmcarthur has quit IRC | 02:59 | |
*** jamesmcarthur has joined #zuul | 03:04 | |
*** jamesmcarthur has quit IRC | 03:18 | |
*** jamesmcarthur has joined #zuul | 03:19 | |
*** jamesmcarthur has quit IRC | 03:49 | |
*** bjackman has joined #zuul | 04:24 | |
openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: zuul-runner: add command line interface https://review.openstack.org/644770 | 04:37 |
tristanC | jhesketh: ^ fixes issues in the zuul-runner command line, it's now working as before | 04:38 |
*** bjackman has quit IRC | 04:39 | |
tristanC | now i'm not sure how to rebase the remaining changes (/connections endpoint, depends-on support, --list-parameters, ...). I guess we should wait for the base of the stack review | 04:40 |
openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: webtrigger: add support for parameterized trigger https://review.openstack.org/644484 | 04:54 |
jhesketh | tristanC: yeah, I wouldn't spend too long on keeping them up-to-date given how reviewing has been going. Lets get the base in and then rethink how we want to approach the rest | 04:58 |
tristanC | jhesketh: alright, well at least that last zuul-runner patch can be used to test the code and run a job locally with the cli | 05:09 |
jhesketh | +1 | 05:10 |
tristanC | and i'll wait for review before continuing to work on that topic for now | 05:11 |
*** bjackman has joined #zuul | 05:24 | |
*** pcaruana has joined #zuul | 06:11 | |
*** saneax has joined #zuul | 06:41 | |
tristanC | the tarball published at http://tarballs.openstack.org/zuul/zuul-3.7.0.tar.gz doesn't have symlinks, the zuul plugin code are copied to each ansible version directories. well it's not an issue, just an odd content. | 06:58 |
*** openstackgerrit has quit IRC | 08:17 | |
*** jpena|off is now known as jpena | 08:45 | |
*** themroc has joined #zuul | 08:56 | |
*** hashar has joined #zuul | 08:59 | |
*** gtema has joined #zuul | 09:42 | |
*** pcaruana has quit IRC | 09:45 | |
*** pcaruana has joined #zuul | 09:46 | |
*** zbr has quit IRC | 10:08 | |
*** zbr has joined #zuul | 10:09 | |
*** zbr has quit IRC | 10:16 | |
*** electrofelix has joined #zuul | 10:16 | |
*** zbr has joined #zuul | 10:21 | |
*** zbr has quit IRC | 10:22 | |
*** Corran has joined #zuul | 10:41 | |
Corran | hey does zuul work with gitlab? | 10:42 |
*** zbr has joined #zuul | 11:38 | |
*** panda is now known as panda|lunch | 11:52 | |
*** jpena is now known as jpena|lunch | 12:01 | |
mgoddard | Hello Zuul. I'm writing an upgrade job. Is it possible to list a required project twice with different branches - as the source and target of the upgrade? | 12:06 |
*** jamesmcarthur has joined #zuul | 12:11 | |
*** rf0lc0 is now known as rfolco | 12:16 | |
*** jamesmcarthur has quit IRC | 12:30 | |
*** jamesmcarthur has joined #zuul | 12:31 | |
*** gtema has quit IRC | 12:35 | |
*** jamesmcarthur has quit IRC | 12:35 | |
*** yolanda has joined #zuul | 12:46 | |
*** bjackman has quit IRC | 12:53 | |
*** bjackman has joined #zuul | 13:03 | |
*** panda|lunch is now known as panda|sick | 13:07 | |
*** irclogbot_2 has quit IRC | 13:07 | |
*** swest has quit IRC | 13:09 | |
*** irclogbot_3 has joined #zuul | 13:10 | |
*** jamesmcarthur has joined #zuul | 13:10 | |
pabelanger | Corran: currently no, there is no gitlab driver. You can use the git driver, but won't get most of the features zuul offers for gating. | 13:19 |
*** gtema has joined #zuul | 13:19 | |
*** jpena|lunch is now known as jpena | 13:20 | |
*** altlogbot_1 has quit IRC | 13:23 | |
*** altlogbot_2 has joined #zuul | 13:25 | |
*** gtema has quit IRC | 13:35 | |
*** altlogbot_2 has quit IRC | 13:39 | |
mordred | Corran: people have expressed interest in the past in writing a driver, but that hasn't materialized. also, I think the gitlab api is similar to the github api, so I imagine it would not be SUPER hard to write one copying the github driver. but so far we don't have a developer to do it | 13:40 |
*** altlogbot_3 has joined #zuul | 13:42 | |
*** irclogbot_3 has quit IRC | 13:45 | |
*** irclogbot_0 has joined #zuul | 13:47 | |
*** bjackman has quit IRC | 13:54 | |
Corran | Thanks for answer! :) | 13:55 |
*** gtema has joined #zuul | 14:08 | |
*** gtema has quit IRC | 14:15 | |
*** gtema has joined #zuul | 14:19 | |
fungi | mgoddard: i'm not sure what you're going for exactly, but zuul preps all branches for you automatically (so if there is a change ahead in a dependent pipeline or as an explicit depends-on or git parent, that will be included in the merges which make up the tips of the respective branches) | 14:21 |
*** themroc has quit IRC | 14:22 | |
fungi | mgoddard: so your upgrade job can checkout oldbranch in that repo, do some things, then checkout newbranch and do some other things, and the corresponding heads for those already include the expected references | 14:22 |
mgoddard | fungi: interesting, that sounds like just what I want, but hadn't considered checking for other branches | 14:23 |
fungi | basically zuul is already branch aware, so all branches of the repos it prepares are for the prospective future states described by the build | 14:23 |
mgoddard | fungi: so a branch listed in the required-project is only the initially checked out branch? | 14:23 |
fungi | i'd have to see an example, but i think you're talking about a branch filter (which would claim that these are the only branches for which this should be a required project) | 14:24 |
fungi | the initially checked-out branch will normally be the same name as the target branch for the change being tested, or "master" if there is no branch in a given repository by that same name | 14:25 |
mgoddard | fungi: https://github.com/openstack/kayobe/blob/master/zuul.d/jobs.yaml#L37 | 14:26 |
mgoddard | fungi: override-checkout - the naming makes sense | 14:26 |
*** jamesmcarthur has quit IRC | 14:29 | |
fungi | oh, yes in that case it overrides the branch zuul would have normally checked out initially, but doesn't prevent zuul from still prepping all relevant branches for you | 14:30 |
*** themroc has joined #zuul | 14:31 | |
*** gtema has quit IRC | 14:32 | |
*** jamesmcarthur has joined #zuul | 14:35 | |
*** altlogbot_3 has quit IRC | 14:35 | |
jamesmcarthur | corvus: apologies for the delayed reply. I found the meeting time here (https://wiki.openstack.org/wiki/Meetings/Zuul). | 14:36 |
jamesmcarthur | I'll be sure to remove that from my schedule :) | 14:36 |
jamesmcarthur | corvus: Re: meeting - discussing over email is fine as well. I hadn't gotten too much traction on my post regarding the board slide deck. | 14:37 |
*** altlogbot_2 has joined #zuul | 14:37 | |
jamesmcarthur | So mainly I wanted to see how I could help. | 14:37 |
corvus | jamesmcarthur: ah thanks! i'll clean up that page :) | 14:37 |
mgoddard | fungi: great, thanks for the help | 14:37 |
*** irclogbot_0 has quit IRC | 14:38 | |
fungi | wow, i forgot we even had that page | 14:38 |
fungi | but yeah, those meetings were primarily used to coordinate work towards the v3 release | 14:39 |
*** irclogbot_2 has joined #zuul | 14:40 | |
*** saneax has quit IRC | 14:43 | |
mgoddard | fungi: slight hurdle, I'm not sure how to change the branch on the executor. It's not a critical issue, just makes it a little harder to generate templates (I have a workaround for that already though) | 14:48 |
mgoddard | fungi: local commands are prevented | 14:48 |
mordred | I think the git module should work | 14:50 |
fungi | mgoddard: that's getting beyond my ken, but perhaps we allow use of ansible's git module? | 14:50 |
mordred | we only prevent it from touching directory paths outside of the work_root | 14:50 |
mgoddard | with delegate_to: localhost? | 14:50 |
*** panda|sick is now known as panda|drappt | 14:50 | |
fungi | i expect just setting host: loclahost would work, but i really don't know | 14:51 |
mordred | you can just do a play with hosts: localhost and tasks: - git: blah | 14:51 |
mordred | fungi: we're in sync today :) | 14:51 |
mgoddard | ok, guess I'll have to break up my mega play | 14:51 |
fungi | i shall endeavor to get out of sync, you seem to have a better handle on this ;) | 14:51 |
mordred | mgoddard: delegate_to: localhost should also work if youv'e got a larger structure you're needing to do things in the context of | 14:52 |
mgoddard | mordred: I'll try that | 14:53 |
mgoddard | thanks mordred and fungi | 14:53 |
fungi | my pleasure! | 14:53 |
*** themroc has quit IRC | 14:54 | |
*** themroc has joined #zuul | 14:58 | |
*** Corran has quit IRC | 15:03 | |
*** themroc has quit IRC | 15:04 | |
dmsimard | If I have patch A in project X that Depends-On patch B from project Y -- if I rebase or submit a new patchset on patch B and there were ongoing jobs for patch A, Zuul will abort everything on patch A and mention "This change depends on a change that failed to merge." | 15:08 |
dmsimard | The message is not very accurate of what just happened -- should this case be handled differently ? | 15:09 |
corvus | likely so :) | 15:16 |
clarkb | The message is incredibly accurate to the point of being misleading :P | 15:16 |
clarkb | at least when using Gerrit | 15:16 |
mordred | yeah. | 15:24 |
dmsimard | well, maybe we're not looking at a different message | 15:25 |
mordred | what I'd LOVE is if the A jobs were restarted on the new version of the B | 15:25 |
dmsimard | should the jobs for the dependant patches be restarted with the new ref automatically ? | 15:25 |
dmsimard | mordred: that would imply that the scheduler is able to find all depended-on patches for a single patch. Is it able to do that ? | 15:26 |
mordred | well, it cancels the relevant jobs right now, so it knows which things are relevant | 15:27 |
*** jamesmcarthur has quit IRC | 15:27 | |
mordred | I said relevant twice. my brain is mushy | 15:27 |
dmsimard | mordred: good point | 15:29 |
dmsimard | I mean, about finding the relevant things | 15:29 |
mordred | the brain mush is also a good point, I promise :) | 15:29 |
dmsimard | that's when there are ongoing jobs | 15:29 |
mordred | yeah. I don't think it should automatically do anything with jobs it wouldn't otherwise cancel | 15:30 |
*** jamesmcarthur has joined #zuul | 15:45 | |
*** jamesmcarthur has quit IRC | 15:50 | |
*** jamesmcarthur has joined #zuul | 15:54 | |
*** openstackgerrit has joined #zuul | 15:59 | |
openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Update docs for provider removal. https://review.openstack.org/645220 | 15:59 |
corvus | tobiash: http://logs.openstack.org/34/645034/1/check/tox-py35/d35ee3e/testr_results.html.gz looks like a legit failure in the uri module fix... i'm puzzled | 16:10 |
corvus | oh! we're testing the 'path' attribute in that test | 16:12 |
corvus | was that ever a thing? | 16:12 |
clarkb | path was available in 2.5's normal module | 16:14 |
corvus | right, but did the uri module actually use it? | 16:14 |
clarkb | but it was an internal var not set directly by the user | 16:14 |
corvus | okay, i think we should update the test case to use 'dest' | 16:16 |
corvus | (for all versions) | 16:16 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Fix Ansible 2.7 uri module https://review.openstack.org/645034 | 16:18 |
corvus | tobiash, clarkb: ^ | 16:18 |
*** jamesmcarthur has quit IRC | 16:29 | |
mgoddard | fungi: mordred: just to follow up, zuul says no to git module + delegate_to: localhost: http://logs.openstack.org/40/616640/20/check/kolla-ansible-centos-source-upgrade/b103cf3/job-output.txt.gz#_2019-03-21_15_22_20_437955 | 16:46 |
pabelanger | mgoddard: yah, you need to be trusted playbook for git on localhost | 16:46 |
mgoddard | pabelanger: ah, that makes sense. Sensible not to trust my garbage playbook | 16:47 |
pabelanger | mgoddard: there is a topic on ML right now about removing that restriction, if you are interested in reading up more on it | 16:48 |
mgoddard | Is there a trusted role available for checking out a branch of required projects? | 16:48 |
pabelanger | mgoddard: my first question would be why do you want to do that on the executor? The node should have a copy of everything pushed to it, that you can modify | 16:49 |
mgoddard | context: writing an upgrade job, need to be able to change branches mid-run | 16:49 |
fungi | i think you could just break that out into a separate playbook and stick it in a trusted repository, then call it from the job in your untrusted repo | 16:49 |
mgoddard | that mostly works for us, but templates must be on localhost | 16:50 |
mgoddard | I have a workaround to fetch the template to the executor | 16:50 |
mgoddard | I'll probably just stick with that for now | 16:50 |
fungi | idea being you have more oversight for playbooks added to trusted repos, and proposed changes to them won't be executed speculatively | 16:50 |
pabelanger | ah, so you want to change which branch ansible-playbook on executor is using for a run | 16:50 |
mgoddard | pabelanger: yes | 16:51 |
mgoddard | thinking about it, that could get messy | 16:52 |
fungi | out of curiosity, what is the job doing on the executor such that it needs other git states there rather than on a remote job node? | 16:52 |
mgoddard | given that the playbook being executed is in that repo | 16:52 |
mgoddard | fungi: executing the template module on a file in the repo. The template source has to be on the executor | 16:53 |
fungi | ahh, okay. i'm unfamiliar with the template module or what you might use it for in that context | 16:53 |
pabelanger | so, I think you could do nested ansible.. which kinda changes your jobs. However, I've often thought it would be neat, for deloyment jobs, if you could some how reuse the same node for a series of jobs to run. | 16:53 |
pabelanger | I can't think of a way to change brances between pre-run / run either | 16:54 |
pabelanger | unless you had a trusted playbook some how modify the work_dir | 16:54 |
*** jamesmcarthur has joined #zuul | 16:55 | |
mgoddard | we're doing nested ansible already - this is for testing kolla-ansible. Although that's an implementation detail here | 16:56 |
clarkb | if you are doing nested ansible already why not chagne the branch on the inner ansible running on the test node? | 16:57 |
mgoddard | the inner ansible is the code under test | 17:00 |
clarkb | ya I think your upgrade job on the zuul size should run from a single branch | 17:00 |
clarkb | then it manipulates the branches of code under test | 17:00 |
*** jamesmcarthur has quit IRC | 17:01 | |
clarkb | *zuul side | 17:01 |
mgoddard | I tend to agree now, having thought about it. Changing the executor code while executing it is asking for trouble | 17:01 |
mgoddard | I think I have a plan, thanks all for your input | 17:02 |
openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: wip - Elasticsearch Zuul reporter https://review.openstack.org/644927 | 17:10 |
openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: wip - Elasticsearch Zuul reporter https://review.openstack.org/644927 | 17:11 |
openstackgerrit | Luigi Toscano proposed openstack-infra/zuul-jobs master: DNM Debug stage-output https://review.openstack.org/645239 | 17:12 |
*** hashar has quit IRC | 17:20 | |
openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Update docs for provider removal. https://review.openstack.org/645220 | 17:21 |
*** gchenuet has joined #zuul | 17:26 | |
gchenuet | Hi guys! Quick question about Nodepool and `boot-from-volume` option: Is there an option to delete volume when instances are deleted ? | 17:27 |
openstackgerrit | Merged openstack-infra/zuul master: Add missing Ansible invocation debug env variables https://review.openstack.org/645035 | 17:28 |
pabelanger | gchenuet: the volume should get deleted by default | 17:29 |
gchenuet | pabelanger: thanks, i'll double check my config | 17:31 |
dmsimard | gchenuet: I double checked to make sure and nodepool is supposed to set the "delete on termination" when creating VMs with boot-from-volume: https://opendev.org/openstack-infra/nodepool/src/branch/master/nodepool/driver/openstack/provider.py#L300-L304 | 17:32 |
gchenuet | dmsimard: i'm looking this file on the same time :) | 17:34 |
*** bjackman has joined #zuul | 17:36 | |
*** gchenuet has quit IRC | 17:43 | |
pabelanger | it is possible you are leaking volumes, I have see that | 17:57 |
*** bjackman has quit IRC | 17:59 | |
pabelanger | at one point, we I know I raised the idea of having nodepool check for leaked volumes, then clean them up (like FIPs) but haven't pushed up any code for that | 18:00 |
pabelanger | I think we needed to add something into shade / openstacksdk | 18:00 |
clarkb | fwiw newer openstack seems to not leak volumes as much | 18:14 |
clarkb | we definitely saw leaking volumes on older openstack deployments but vexxhosts sjc1 which is rocky? doesn't seem to do it | 18:15 |
pabelanger | yah, so far ansible-network tenant, haven't see any leaks in last 5 months | 18:15 |
*** jpena is now known as jpena|off | 18:22 | |
openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: wip - Elasticsearch Zuul reporter https://review.openstack.org/644927 | 18:39 |
*** electrofelix has quit IRC | 18:43 | |
*** sshnaidm|rover is now known as sshnaidm|off | 19:03 | |
openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: wip - Elasticsearch Zuul reporter https://review.openstack.org/644927 | 19:13 |
openstackgerrit | Merged openstack-infra/zuul master: Fix Ansible 2.7 uri module https://review.openstack.org/645034 | 19:21 |
*** badboy has quit IRC | 20:06 | |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add fetch-sphinx-tarball role https://review.openstack.org/645346 | 20:51 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: DNM: test fetch-sphinx-tarball https://review.openstack.org/645347 | 20:51 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add fetch-sphinx-tarball role https://review.openstack.org/645346 | 20:52 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: DNM: test fetch-sphinx-tarball https://review.openstack.org/645347 | 20:52 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add fetch-sphinx-tarball role https://review.openstack.org/645346 | 21:04 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: DNM: test fetch-sphinx-tarball https://review.openstack.org/645347 | 21:04 |
*** pcaruana has quit IRC | 21:33 | |
*** irclogbot_2 has quit IRC | 22:05 | |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: DNM: test fetch-sphinx-tarball and download-artifact https://review.openstack.org/645347 | 22:06 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add download artifact role https://review.openstack.org/645384 | 22:06 |
*** mgoddard has quit IRC | 22:08 | |
*** mgoddard has joined #zuul | 22:09 | |
corvus | tobiash: yay! the uri fix fixed the docker promote jobs! https://hub.docker.com/u/zuul | 23:24 |
*** jamesmcarthur has joined #zuul | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!