| *** rlandy has quit IRC | 00:35 | |
| SpamapS | derp | 00:41 |
|---|---|---|
| SpamapS | derp a derp | 00:41 |
| SpamapS | I firewalled off the zuul port | 00:42 |
| SpamapS | ;) | 00:42 |
| *** swest has joined #zuul | 04:49 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Log connection port in static driver on timeout https://review.openstack.org/569334 | 05:59 |
| *** sshnaidm|bbl is now known as sshnaidm|off | 06:23 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Correctly use connection-port in static driver https://review.openstack.org/569339 | 06:23 |
| *** johanssone has joined #zuul | 06:28 | |
| *** threestrands has quit IRC | 07:17 | |
| *** toabctl has quit IRC | 07:26 | |
| *** gtema has joined #zuul | 07:28 | |
| *** jpena|off is now known as jpena | 07:58 | |
| *** gtema has quit IRC | 07:59 | |
| *** gtema has joined #zuul | 08:00 | |
| *** sshnaidm|off has quit IRC | 10:41 | |
| *** toabctl has joined #zuul | 11:06 | |
| *** sshnaidm|off has joined #zuul | 11:12 | |
| *** jpena is now known as jpena|lunch | 11:49 | |
| *** EmilienM is now known as EvilienM | 12:21 | |
| *** rlandy has joined #zuul | 12:35 | |
| *** jpena|lunch is now known as jpena | 12:44 | |
| *** dkranz has joined #zuul | 13:02 | |
| *** myoung|ruck|off is now known as myoung|ruck | 13:37 | |
| mordred | SpamapS: those are great stats | 13:42 |
| *** acozine1 has joined #zuul | 13:57 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Replace use of aiohttp with cherrypy https://review.openstack.org/567959 | 14:30 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Convert streaming unit test to ws4py and remove aiohttp https://review.openstack.org/568335 | 14:30 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: DNM: disable test log capture https://review.openstack.org/569429 | 14:30 |
| openstackgerrit | Merged openstack-infra/zuul master: Fix race in test_reconfigure_window_fixed https://review.openstack.org/569129 | 14:58 |
| openstackgerrit | Merged openstack-infra/zuul master: Revert "Revert "Switch to stestr"" https://review.openstack.org/568949 | 15:00 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Replace use of aiohttp with cherrypy https://review.openstack.org/567959 | 15:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Convert streaming unit test to ws4py and remove aiohttp https://review.openstack.org/568335 | 15:39 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Simplify driver API https://review.openstack.org/568704 | 15:43 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Add a "Get Started" button https://review.openstack.org/569450 | 15:46 |
| corvus | fdegir: i was just looking at 565075 and i see the additional bits for centos7 services, but i don't see any updates to the introductory text indicating that the instructions work for centos7 as well as fedora | 16:00 |
| corvus | fdegir: is there another change out there, or do we need to add that still? | 16:00 |
| corvus | fdegir: ah! it's https://review.openstack.org/564948 i think | 16:01 |
| fdegir | corvus: yes, there are 3 changes, each updating different sections | 16:01 |
| fdegir | https://review.openstack.org/#/c/564948/ | 16:02 |
| fdegir | https://review.openstack.org/#/c/564950/ | 16:02 |
| fdegir | https://review.openstack.org/#/c/565075/ | 16:02 |
| corvus | fdegir: cool, i'm caught up now :) in the future, it might make sense to stack changes like that on each other. if you make 3 commits in a row, then run 'git-review' you'll get 3 changes in gerrit, but with a dependency relationship. that makes it easy to see the final results of all three in the documentation preview, and helps reviewers find the changes. | 16:03 |
| fdegir | corvus: yes - the next ones for ubuntu will be like that | 16:04 |
| fdegir | corvus: for this one, i did it this way on purpose since the structure was heavily impacted and didn't want to have conflicts | 16:04 |
| corvus | ok i've approved all 3 | 16:04 |
| fdegir | thx a lot | 16:05 |
| fdegir | corvus: this one is a bit tricky | 16:06 |
| fdegir | corvus: https://review.openstack.org/#/c/565078/ | 16:06 |
| fdegir | corvus: i wanted to fix it properly as Tobias suggested but lack time recently | 16:06 |
| corvus | yeah, we should clean that up soon :) | 16:07 |
| fdegir | corvus: but it would be good to fix service fileto reduce the confusion until the proper fix is there | 16:07 |
| corvus | fdegir: i worry a bit about that since we can't make sure service files are updated; it'd be better to try to fix that properly real quick | 16:08 |
| openstackgerrit | Merged openstack-infra/zuul master: Add additional steps for configuring Zuul services on CentOS 7 https://review.openstack.org/565075 | 16:08 |
| fdegir | corvus: similar thing is there for nodepool as well - the service file is correct but it lacks command socket support | 16:09 |
| openstackgerrit | Merged openstack-infra/zuul master: Add additional steps for configuring Nodepool service on CentOS 7 https://review.openstack.org/564950 | 16:23 |
| SpamapS | mordred: heh, we also calculated how much we'd have spent if we had run all of this in AWS... | 16:24 |
| SpamapS | but that's not really fair | 16:25 |
| SpamapS | because we'd probably not do min-ready's ;) | 16:25 |
| SpamapS | and might even use spot instances | 16:25 |
| corvus | SpamapS: the builds db doesn't include time that nodes spend in the ready state, that should be pretty close to actual usage | 16:26 |
| SpamapS | corvus: indeed, I wonder if I could use the nodepool launcher logs | 16:27 |
| SpamapS | But really, the seconds running jobs is probably within a few percentage points of accurate if we did fully on-demand nodes. | 16:27 |
| *** electrofelix has quit IRC | 16:27 | |
| clarkb | at one time I did math that we were saving over a day of cpu time per day simply by removing the default grub timeout on our images | 16:28 |
| openstackgerrit | Merged openstack-infra/zuul master: Add CentOS 7 environment setup instructions https://review.openstack.org/564948 | 16:29 |
| clarkb | (so that pre job time can add up) | 16:29 |
| SpamapS | Hah good point | 16:29 |
| SpamapS | we have a horrible image | 16:29 |
| SpamapS | that does evil ssh things for the first 60s it is booted | 16:30 |
| SpamapS | so I actually had to write a pre-run that checks if the box has been up for at least 120s and if not, waits that long. | 16:30 |
| *** dkranz has quit IRC | 16:57 | |
| *** acozine1 has quit IRC | 16:58 | |
| *** dkranz has joined #zuul | 16:58 | |
| *** gtema has quit IRC | 17:07 | |
| *** jpena is now known as jpena|off | 17:07 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Add more explanatory text to zuul-from-scratch https://review.openstack.org/569482 | 18:04 |
| corvus | Shrews: we should figure out how to continue the zfs document. the next parts of the tutorial are going to be writing a bunch of changes and pushing them up to gerrit or github | 18:05 |
| corvus | Shrews: (the github branch already goes further than the gerrit branch in that it sets up a pipeline and exercises it) | 18:06 |
| corvus | Shrews: fundamentally, it should be the same process for both, so we probably just need to figure out how to describe actions like "push up a gerrit change / open a github pull request" in neutral terms? | 18:06 |
| corvus | Shrews: or maybe mechanically duplicate the content and switch out those terms. | 18:07 |
| Shrews | corvus: uh, ok. honestly, i haven't put any more thought into that doc. i'm sure it can be improved/expanded | 18:09 |
| Shrews | glad it's getting some attention from contributors :) | 18:10 |
| Shrews | it *is* pretty bare bones, as you point out | 18:11 |
| corvus | yeah, we should honestly probably make it deeper before we make it any broader | 18:12 |
| corvus | we're about to have a document that tells you how to set up zuul 6 different ways, none of which will run your tests. | 18:12 |
| Shrews | oh, um, erm... | 18:13 |
| Shrews | i wanna say that's not unbad | 18:13 |
| Shrews | definitely should expand on the configuration portion | 18:14 |
| corvus | yeah, next we need to walk folks through creating a base job (or using what's in the zuul-base-jobs repo), then constructing a simple job on top of that which actually runs a command. | 18:15 |
| Shrews | wait... that sounds familiar | 18:16 |
| Shrews | https://review.openstack.org/493873 | 18:17 |
| corvus | Shrews: it's going to take a bit more than that to finish, but i'd love your continued assistance. | 18:22 |
| tobiash | we're planning to write a github zuul settings linter github app for our users | 18:23 |
| corvus | tobiash: why not use zuul itself? | 18:23 |
| tobiash | that will on request check repo and branch protection settings for common pitfalls | 18:24 |
| tobiash | corvus: would that make sense? | 18:24 |
| tobiash | our idea is that it would be triggered by a command like '/support' and then add a comment with a checklist | 18:25 |
| tobiash | so to do that in zuul itself I think this would be a separate non-pipeline related functionality of the github driver | 18:26 |
| corvus | tobiash: in openstack-infra, we have linter jobs which run changes to our project-config repo when people add new projects. you might consider doing something like that. | 18:26 |
| corvus | correction: ...which run *on* changes... | 18:26 |
| tobiash | yeah so we have something similar for our tenant config | 18:27 |
| tobiash | but what I often get as a support question from users is 'why do I get merge failed after a successful gate' | 18:27 |
| tobiash | or 'why do I get a gate loop' | 18:27 |
| Shrews | corvus: as long as you can help me define the end product you want delivered, glad to help however i can | 18:28 |
| tobiash | and the branch protection settings are not tracked in any config management | 18:28 |
| tobiash | so our idea was a more or less interactive support bot (which could be zuul itself or a separate app) | 18:29 |
| corvus | Shrews: i will take you up on that post-summit :) | 18:30 |
| tobiash | so using a job for this would feel a bit odd because we would have to add it to every project in every tenant into a special support pipeline | 18:32 |
| corvus | tobiash: is this something that needs to be checked only when you add a new project, or more often than that? | 18:32 |
| tobiash | corvus: more often (as users add/remove branches and branch protection via github ui) | 18:33 |
| tobiash | and also preferably directly on request (user see's something odd, adds '/analize' comment and gets a linting report of what could be wrong) | 18:34 |
| corvus | tobiash: the project-regex feature could add the support pipeline :) | 18:35 |
| tobiash | Yes, but in a multi tenant env that feels clumsy | 18:36 |
| tobiash | So our idea was a simple github app based on probot (and open source it if it works out) | 18:36 |
| corvus | tobiash: you can put the configuration for that in its own repo, and then add that repo to every tenant. so the tenants don't even have to see it. | 18:37 |
| corvus | (or, if you already have a repo for global configuration which applies to all tenants, you can add it there) | 18:37 |
| corvus | tobiash: obviously it's your decision -- but "run an action and leave feedback based on a github pull request event" should be well within zuul's capabalities -- i just want to make sure we're not missing anything :) | 18:39 |
| *** elyezer has quit IRC | 18:40 | |
| tobiash | corvus: thanks for the input, we will think deeper about that | 18:42 |
| *** elyezer has joined #zuul | 18:43 | |
| tobiash | I think we would need the comment text in the zuul vars so we can support several commands without having 50 pipelines | 18:43 |
| *** _ari_ has quit IRC | 18:46 | |
| *** _ari_ has joined #zuul | 18:46 | |
| corvus | tobiash: good point; the idea of selecting jobs based on comment text (or other trigger characteristics) may be relevant too | 18:47 |
| corvus | either, or both, of those may make that approach more feasible | 18:47 |
| tobiash | Indeed that even was a customer request I got a few days ago | 18:47 |
| corvus | (in openstack, we mostly suggest folks just dynamically change the in-repo config to run whatever jobs they want, in the case where they're trying to work through a specific issue) | 18:48 |
| corvus | otherwise, that needs a bit more thought, because you obviously can't do something like that in a gating configuration | 18:49 |
| corvus | putting the text in a zuul var seems pretty straightforward though :) | 18:49 |
| tobiash | the user's use case was that there is a very expensive test (too big for check and gate) that should only run if the user wants it to run based on expert knowledge of the risks of a change | 18:51 |
| tobiash | Such jobs would normally run nightly | 18:51 |
| corvus | tobiash: ah, in openstack we use an 'experimental' pipeline for that | 18:52 |
| tobiash | My advice was a comment triggered experimenal pipeline | 18:52 |
| tobiash | ;) | 18:52 |
| corvus | still, selecting jobs by comment in an experimental pipeline would be nicer, and safe. | 18:53 |
| corvus | so maybe it's a matter of saying "this pipeline allows job selection" and setting that on experimental, but not check/gate. | 18:53 |
| tobiash | Interesting idea | 18:55 |
| *** xhku_ has quit IRC | 19:43 | |
| *** fbouliane has joined #zuul | 19:45 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Replace use of aiohttp with cherrypy https://review.openstack.org/567959 | 20:25 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Convert streaming unit test to ws4py and remove aiohttp https://review.openstack.org/568335 | 20:25 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Add never_capture test decorator https://review.openstack.org/569515 | 20:26 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Add never_capture test decorator https://review.openstack.org/569515 | 20:46 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Replace use of aiohttp with cherrypy https://review.openstack.org/567959 | 20:46 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Convert streaming unit test to ws4py and remove aiohttp https://review.openstack.org/568335 | 20:46 |
| corvus | the never_capture decorator may as well be named never_fail_now_that_i_can_see_your_output | 20:46 |
| clarkb | heisenbugs? | 20:48 |
| corvus | ya | 20:51 |
| corvus | i ran it many hundreds of times locally with no failure :( | 20:52 |
| SpamapS | really toob ad that aiohttp didn't work out :-P | 21:15 |
| SpamapS | seemed nice. | 21:16 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Add never_capture test decorator https://review.openstack.org/569515 | 22:44 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Replace use of aiohttp with cherrypy https://review.openstack.org/567959 | 22:44 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Convert streaming unit test to ws4py and remove aiohttp https://review.openstack.org/568335 | 22:44 |
| *** xinliang has quit IRC | 23:08 | |
| *** xinliang has joined #zuul | 23:20 | |
| *** rlandy has quit IRC | 23:34 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!