14:01:54 <tflink> #startmeeting fedora-qadevel 14:01:54 <zodbot> Meeting started Mon Apr 10 14:01:54 2017 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:54 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:01:54 <zodbot> The meeting name has been set to 'fedora-qadevel' 14:02:00 <tflink> #topic roll call 14:02:07 * mkrizek is here 14:02:14 * kparal is here again 14:02:29 * roshi is here 14:02:41 <tflink> kparal: again? 14:02:49 <kparal> after two weeks of sickness 14:02:54 <kparal> seems nobody noticed 14:02:58 <roshi> feeling better? 14:02:59 <tflink> ah, good to have you back 14:03:02 * roshi did 14:03:02 <kparal> so I can just quietly leave again... 14:03:20 <kparal> feeling better, yes :) 14:03:28 <tflink> the plague certainly seems to be going through the office over there 14:04:22 <x3mboy> .hello x3mboy 14:04:23 <zodbot> x3mboy: x3mboy 'Eduard Lucena' <eduardlucena@gmail.com> 14:05:42 <tflink> ok, lets get started 14:05:45 <tflink> #topic Announcements and Information 14:05:56 <tflink> #info modularity patches merged and released - mkrizek 14:06:07 <tflink> any additional items? 14:06:23 <tflink> #chair kparal mkrizek roshi 14:06:23 <zodbot> Current chairs: kparal mkrizek roshi tflink 14:07:00 * tflink takes that as a no 14:07:03 * roshi has nothing 14:07:15 <tflink> ah, release validation :) 14:07:33 <tflink> any comments or questions? 14:07:53 <roshi> pretty much sucks us all right in :P 14:08:37 <tflink> ok, moving on 14:08:43 <tflink> #topic Deployment of new features 14:09:03 <tflink> Not much to say here. I'd like to get some of the new features deployed to prod today if there are no objections 14:09:09 * roshi has none 14:09:09 <tflink> new trigger, new libtaskotron 14:09:36 * kparal is not yet up-to-date 14:09:44 <tflink> I'd also like to enable docker layered image tests in stg to make sure everything is working and prod later this week 14:09:55 <mkrizek> no objections, if there were no issues in dev/stg 14:10:04 <tflink> not that I'm aware of 14:10:04 <kparal> what do we use to test docker layered images? 14:10:10 <tflink> bash 14:10:29 <kparal> so it's a custom task running some upstream test suite, right? 14:10:30 <roshi> adam miller had asked me about the status of that 14:10:45 <roshi> that part of the dist-git stuff? 14:11:10 <tflink> it's allowing image authors to wrtie something in response to an image build 14:11:20 <kparal> because I replied today to test list and the question was similar, testing Boltron. tflink you might want to look into the list and correct my statements if necessary 14:11:38 <tflink> AFAIK, there is no good "universal" way to test layered images 14:11:48 <tflink> ok, will do. haven't gotten through my email yet this morning 14:12:58 <tflink> roshi: pretty much, yeah. the thing that's left is to make sure that the trigger works correctly on docker image build 14:13:30 * tflink was going to do it the other week but stg OSBS was down and I got distracted after it came back up 14:14:02 <roshi> ok, cool 14:14:28 <tflink> also, getting nested virt enabled on the prod virthosts so we can start running the atomic stuff 14:14:48 * tflink wonders if there is a more global outage planned this week or next week for updates 14:15:13 <kparal> have we tested nested virt on dev/stg already? any issues? 14:15:31 <tflink> dev, yes 14:15:41 <kparal> how does it work? 14:15:43 <tflink> roshi: did we put it in stg as well 14:15:57 <tflink> it works fine AFAIK. roshi was doing more of the work on that part 14:16:54 <tflink> roshi: can you elaborate on how well the atomic tests were working in dev/stg? 14:17:17 <kparal> I guess I should look into failure emails folder from taskotron hosts :) 14:17:43 <kparal> I haven't dared yet 14:17:45 <tflink> hopefully that'll start getting more sane 14:17:52 <tflink> once we get your patch into prod 14:18:04 <kparal> one drop in an ocean 14:18:04 <roshi> so far they seem to work? 14:18:25 <roshi> I'm seeing results, which are expected 14:18:32 <kparal> I think the most frequent bug is "failed to boot in 60 seconds" due to cloud-init probably 14:18:40 <roshi> that was the main reason I wanted to get things in docker, so I could test locally 14:18:54 <roshi> so now I'm playing with the taskotron-docker repo jan pointed me to 14:19:34 <kparal> roshi: the repo contains dockerfiles for stable releases, and each project repo contains a dockerfile for development releases 14:19:38 <kparal> IIUIC 14:19:45 <roshi> yeah, poking at it now 14:19:48 <kparal> garretraziel can tell you the details 14:19:53 <roshi> good to know 14:20:21 <tflink> it's not quite the same thing 14:20:28 * tflink will respond to emails later 14:20:37 <roshi> yeah, they're a bit different 14:20:49 <tflink> roshi: sorry for not telling you about them but I want installation from git, not rpm 14:20:50 <roshi> nginx instead of apache, for one 14:20:59 <roshi> right 14:21:04 <roshi> this is something to work off of though 14:21:10 <roshi> known working, which is good 14:21:46 <tflink> which is a nice segue into 14:21:52 <tflink> #topic Resultsdb Performance 14:22:03 <tflink> This'll be short since we have no josef this morning 14:22:17 <tflink> but we're still seeing occasional issues with latency/load in prod 14:22:29 <kparal> are we still seeing 5xx errors? 14:22:40 <kparal> or just slow responses? 14:23:01 <tflink> not as much as we were for a while but the alarms are going off and alerting pagers 14:23:38 <tflink> waking up sysadmins on a weekend is not something that I'd very much like to stop doing 14:24:03 <roshi> meaning you want to stop doing that 14:24:08 <kparal> or they might visit us one night and staby staby 14:24:09 <tflink> yep 14:24:14 <roshi> lol 14:24:20 <kparal> agreed 14:25:06 <tflink> the docker stuff that roshi is talking about is an attempt to get some performance testing into resultsdb so we can start figuring out exactly what's going on 14:25:24 * tflink has some load generation scripts that will be part of that as well 14:25:33 <kparal> sounds great 14:26:02 <tflink> we should also get moving on putting prod resultsdb on its own db host so we can isolate some issues 14:26:13 <tflink> the db host we're using is arguably overloaded as it is 14:27:17 <tflink> anything else on this? 14:27:54 * roshi has nothing 14:28:11 <tflink> ok, moving on 14:28:19 <tflink> #topic Dist-Git Task Storage 14:28:37 <tflink> not much to say on this one, it should be pushed to prod this week (today) if nothing goes wrong 14:29:00 <tflink> I suspect we're still lackign some docs but otherwise, I'm not aware of any problems 14:29:29 <tflink> so this was pretty much meant to be a status update and a poke to maybe get more docs/examples written :) 14:29:36 <tflink> (poke for me as well) 14:30:08 <tflink> anything else on this? 14:30:32 * tflink assumes not 14:31:03 <tflink> #topic CI 14:31:10 <tflink> my favorite word! 14:31:38 <roshi> whee 14:32:06 * kparal plays a dancing pony video 14:32:09 <tflink> there is a push to get more "CI" into fedora 14:32:31 <tflink> Which will become more obvious in the coming weeks 14:32:43 <tflink> A new list has been created for discussions surrounding CI 14:32:44 <tflink> https://lists.fedoraproject.org/archives/list/ci@lists.fedoraproject.org/ 14:32:54 <tflink> no traffic yet but I expect that to change this week 14:33:21 * kparal prepares for some big revelations, e.g. that we don't need to do any more work and just relax 14:33:30 <tflink> ha 14:34:04 <tflink> well, "CI is a development activity" so you could be on to something :) 14:34:16 <tflink> in reality, I don't see anything changing too much for a while 14:34:37 <tflink> just wanted to give folks a heads up in case they were interested 14:35:15 <tflink> any comments/questions ? 14:35:17 * kparal has enough information overload already 14:36:26 <garretraziel> hmm, that list url takes too much to load for me 14:36:41 <tflink> yeah, mailman seems to be having issues right now 14:37:58 <tflink> any other comments/questions? 14:38:51 * tflink assumes not 14:38:55 <tflink> #topic Tasking 14:39:02 <tflink> Anyone in need of stuff to do? 14:39:37 <kparal> I need a person to read all my emails 14:40:17 * tflink would love to harness machine summarizing algorithms to help with emails :) 14:40:18 <kparal> responding would be a great bonus feature 14:41:49 <tflink> ok, if anyone finds themselves in need of tasks, let me know 14:41:59 <tflink> there are a few things that would be nice to have soon 14:42:08 <tflink> but that brings us to ... 14:42:13 <tflink> #topic Open Floor 14:42:26 <tflink> any additional topics that folks want to bring up? 14:42:36 <kparal> nothing here 14:43:48 <tflink> ok, thanks for coming, everyone 14:43:55 * tflink will send out minutes shortly 14:44:02 <tflink> #endmeeting