14:00:31 #startmeeting fedora-qadevel 14:00:31 Meeting started Mon Oct 17 14:00:31 2016 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:31 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:00:31 The meeting name has been set to 'fedora-qadevel' 14:00:31 #meetingname fedora-qadevel 14:00:31 The meeting name has been set to 'fedora-qadevel' 14:00:31 #topic Roll Call 14:00:41 #chair kparal jskladan lbrabec garretraziel 14:00:41 Current chairs: garretraziel jskladan kparal lbrabec tflink 14:00:57 * garretraziel is here 14:01:03 * mkrizek is here 14:01:03 It's monday and we all know what that means! 14:01:05 * omiday is here 14:01:18 * jskladan nods 14:01:18 #chair omiday 14:01:18 Current chairs: garretraziel jskladan kparal lbrabec omiday tflink 14:01:40 * tenk is here 14:01:47 #chair tenk 14:01:47 Current chairs: garretraziel jskladan kparal lbrabec omiday tenk tflink 14:02:38 ok, let's get started 14:02:48 * kparal is here 14:02:52 * lbrabec is here 14:02:59 #topic Announcements and Information 14:03:11 #info taskotron-trigger-0.4.1-1 deployed to dev - mkrizek 14:03:11 #link https://phab.qadevel.cloud.fedoraproject.org/T854 14:03:27 any other items, questions or comments? 14:04:35 Not really, I'd like to have https://phab.qadevel.cloud.fedoraproject.org/D1027 merged before we build resultsdb 14:04:37 but that's about it 14:05:12 other than that, are you aware of any reason not to do releases and builds for resultsdb/resultsdb_frontend? 14:05:29 and get them into dev as soon as we're done testing trigger 14:05:44 nope 14:05:51 cool 14:06:35 * tflink assumes that silence means no more questions or comments 14:06:37 #topic Docker Testing Status 14:07:07 just to make sure I understood correctly, the docker-relevant bits were put into the 0.4.1 release of trigger, right? 14:07:20 yep 14:08:01 does that mean that we just need to have a repo to look at for tasks and adjust the configuration accordingly before we can call that "done for F25"? 14:08:07 testing aside 14:08:43 well, we should write the actual discover code :) 14:09:00 oh, i thought that was done for some reason 14:09:02 :-/ 14:09:25 but that was/is stalled, because we don't really have (or at least I'm not aware of) the spec for how the repos will be structured 14:09:27 and so on 14:09:38 for now, it's going to be a single repo for the cockpit image 14:09:53 since that's the only thing that is being produced for F25 14:10:16 once the dist-git proposal for rpms is figured out, we'll mirror that decision (in tree or separate namespace) for docker 14:10:52 is that enough information to finish the discovery code for now? 14:11:21 if so, how long would it take to get a patch for that proposed? We're supposed to have that into production before F25 final freeze 14:12:38 jskladan, lbrabec ^^ 14:13:15 depends on where (and how structured) will it be in the cockpit repo, TBH 14:13:34 which, I think, ties right back to the dist-git thing 14:14:09 could we do it the same way we handle our tasks, at the moment? 14:14:14 yes 14:14:51 so, I'd rather do that, for starters 14:14:59 that works for me, I assume that changes/enhancements will be needed when more images are produced for F26 and/or when someone figures out more of what testing docker images means 14:15:02 and when we come up with the "proper" way, we can switch to it 14:15:05 agreed 14:15:18 *thumbs up* 14:15:34 do you think that could be done in the next day or two? 14:16:56 * jskladan shrugs I'm not the expert on that matter, honestly 14:17:24 but I think that it's just changes to ansible/deployment right? 14:17:31 so it copies the right repo, and so on 14:17:47 I'll rephrase a bit, I want to have something in dev in the next couple days, even if that's as brain dead as assuming all the tests live in directories inside a single repo 14:18:12 or even a single directory if that makes the difference between getting something ready or not :) 14:18:18 the only problem I see there is the fact that we expect "runtask.yml" (or however is it called) in the repo's root 14:18:29 oh, or did you mean the discover code? 14:19:00 isn't that what we were talking about? 14:19:12 confusion reigns, apparently :) 14:19:19 I thought we agreed on handling it just the same as we do rpmlint/depcheck and so on 14:20:07 * tflink doesn't care much so long as it works for the short term 14:20:40 we can make improvements once there is a conclusion on how stuff is layed out and there's more than 1 docker image to worry about 14:21:19 ad discover - I can't really say, at the moment. I'll have a look at the code, and will try and put something together tomorrow 14:21:41 but I'd start with handling this one case the same as we do our stuff (depcheck/rpmlint/...) 14:21:45 jskladan: would it make a difference if we just pointed it at a repo like depcheck/rpmlint? 14:21:48 works for me 14:22:03 as far as I know, we expect runtask.yml in the repo's root 14:22:16 like I said, I just want something working in the short term for now. I'd really like to meet this deadline if at all possible 14:22:19 that works for me 14:22:32 that is the only hurdle that I can think of 14:22:48 AFAIK, one of us will be writing the task for now. I was just planning to steal the example from the docs that lbrabec wrote 14:23:03 once that works, we'll start talking to the cockpit folks about adding mroe 14:23:05 more 14:23:22 yeah, those are written in "discover will be done some time in the future, and we will want to use it" 14:23:47 anyway - I'll work on the discover thing tomorrow, and if you (or mkrizek) could set up the "standard" way in the mean time 14:24:02 just in case it's a tougher nut to crack... 14:24:30 but by tomorrow's afternoon, I'll know how much effort needs to be put into the discover code 14:24:49 don't have a good idea right now 14:25:48 just to make sure we're on the same page - that means treating the triggering the one docker image that's being produced for F25 in the same way that we're currently doing depcheck/rpmlint/etc. until more details have been figured out for how the checks will be layed out in the repo we end up using 14:26:04 but as I say that, I think I finally understand what you were getting at earlier 14:26:38 we can get that ready right now by just pointing docker build messages at a single repo which works because there's only 1 image being built ATM 14:26:57 yes 14:27:26 apparently the coffee hasn't quite kicked in yet :-/ 14:27:26 and that gives us both "look we delivered" and "some time to get it done properly" 14:27:34 sounds good 14:27:44 awesome 14:27:50 it isn't so terribly urgent if we're going that route 14:28:14 I still want to get it done well before F26 freeze but it can wait until we know what the repo layout/location will look like 14:28:23 anything else on this topic? 14:29:11 * tflink assumes not 14:29:30 #topic Trigger Re-Write Status 14:29:38 I think we've covered most of this already 14:30:07 I'll hop on it, though, it's a missing piece, and it seems to have a bigger priority than the cli tool.. 14:30:07 new trigger 0.4.1 is deployed to dev, better support for docker testing will be looked into once we know what that's going to look like 14:30:43 anything else on this? 14:30:48 nope 14:31:02 ok, moving on 14:31:04 #topic Dist-Git Task Storage Proposal 14:31:25 just another reminder - there is a thread on devel@ talking about our proposal for storing tasks in dist-git 14:31:57 * tflink kicked the thread last night to get more conversation on namespaces/directories 14:32:02 not sure there's much else to say here 14:32:29 which brings us to 14:32:33 #topic Open Floor 14:32:48 Any topics that didn't come up earlier? 14:33:37 I have nothing 14:34:01 nothing here 14:34:54 if there's nothing else, I'll set the fuse for 0-5 minutes 14:34:58 thanks for coming, everyone 14:35:10 thanks 14:35:28 thx 14:35:56 * tflink will send out minutes shortly 14:36:00 #endmeeting