14:01:52 <tflink_> #startmeeting fedoraqa-devel 14:01:52 <zodbot> Meeting started Mon Aug 24 14:01:52 2015 UTC. The chair is tflink_. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:52 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:01:52 <tflink_> #meetingname fedoraqa-devel 14:01:52 <tflink_> #topic roll call 14:01:52 <zodbot> The meeting name has been set to 'fedoraqa-devel' 14:02:37 <tflink_> #chair tflink kparal 14:02:37 <zodbot> Current chairs: kparal tflink tflink_ 14:02:53 * lbrabec is here 14:02:55 * kparal is here 14:03:10 * garretraziel is here 14:03:12 <bexelbie> .hello bex 14:03:14 <zodbot> bexelbie: bex 'Brian (bex) Exelbierd' <bex@pobox.com> 14:03:30 <bexelbie> sorry, wrong meeting 14:03:47 <kparal> Josef replied a while ago that he wouldn't be able to be here 14:03:54 <tflink> bexelbie: no worries 14:04:05 <tflink> and Martin is on PTO until tomorrow IIRC 14:04:29 <tflink> so I think we have everyone who's expected 14:04:41 <kparal> yes 14:04:55 <tflink> #topic status updates 14:05:21 <tflink> #info got taskotron mostly working with bodhi2 - tflink/kparal 14:05:21 <tflink> #info disposable clients are working well enough to use for dev work 14:05:21 <tflink> #info Added test coverage for weak dependencies to depcheck, all tests pass - handsome_pirate 14:05:21 <tflink> #info switch from TAP to YAML for inter-directive communication and results representation in libtaskotron. Early WIP of the result-yaml format: https://bitbucket.org/fedoraqa/resultyaml - tflink, jskladan 14:05:21 <tflink> #info OpenQA - added updates.img via local media test - jsedlak 14:05:34 <tflink> any questions/comments about the above? 14:05:40 <tflink> or additional things to add? 14:06:02 <kparal> nope 14:06:12 <kparal> well 14:06:14 <garretraziel> no 14:06:28 <kparal> what is the ETA for the TAP replacement? 14:06:42 <kparal> i.e. how fast should I rush to review that? 14:06:54 * handsome_pirate waves 14:06:59 <kparal> very soon/in a few months? 14:07:11 <tflink> haven't really talked about it yet but I'd like to have it by the time we have dist-git tasks in place 14:07:28 <tflink> kparal: I wish I could say very soon but I think that's a bit on the optimistic side 14:07:35 <kparal> ok 14:07:39 <tflink> switching over would make packaging somewhat easier 14:07:54 <tflink> and less confusing when folks start writing tasks 14:09:03 <tflink> kparal: if you could look at it to see if we came up with something horrible or if we're missing an important use case, that would be great but a detailed review can wait, I think 14:09:22 <tflink> it's mostly a collection of use cases right now 14:09:50 <tflink> there's some small inconsistencies between them but that'll be worked out with time 14:10:01 <kparal> ok, thanks for info 14:10:09 <tflink> if there's no other questions/comments, moving on 14:10:19 <tflink> #topic Bodhi2 and Fedmsg emission 14:10:39 <tflink> in case anyone missed it, bodhi was upgraded to bodhi2 last week 14:10:53 * kparal didn't miss that 14:11:05 <tflink> kparal and I spent most of last week trying to prepare for the upgrade and cleaning up after it 14:11:18 <tflink> fun times ... 14:11:53 <tflink> anyhow, most of our stuff is working now with the notable exception of bodhi comments 14:11:58 <kparal> bodhi2 comments still don't work, but I'll poke infra today to move it forward 14:12:48 <tflink> we're trying not to put too much effort into supporting bodhi2 comments, though since we want our commenting code to DIAF 14:13:45 <tflink> as i understand it, mkrizek will be working on the fedmsg emission stuff again when he gets back tomorrow 14:14:35 <tflink> once we have fedmsg emission working from resultsdb and appropriate filters in place for fmn, we can get rid of comments 14:15:20 <tflink> one other important note - fakefedorainfra no longer represents a mock bodhi interface and we're not planning to take the time to update it. the project will likely be killed off soon 14:15:43 <tflink> any comments/questions on this or anything I missed? 14:16:36 * tflink takes that as a no, moves on to next topic 14:16:45 <tflink> #topic Remaining Disposable Client Work 14:16:57 * kparal likes DIAF acronym 14:17:48 <tflink> We had a successful-enough demo of the disposable client bits at Flock - folks seem to be interested and looking forward to the dist-git tasks 14:18:02 <tflink> but there's still work left before the disposable clients are ready for production 14:18:31 <tflink> mostly around image production, image handling, system maintenance and polish 14:19:43 <tflink> I'd like to get this done soon. other than the fedmsg/bodhi_comment stuff, this is the higest priority for taskotron work 14:20:13 <tflink> does everyone involved in dev understand what the disposable client feature is and why we're doing it? 14:20:56 * kparal feels he should say yes 14:21:05 <garretraziel> yup, I think so 14:21:19 * lbrabec nods 14:21:26 * handsome_pirate does 14:21:31 <tflink> good - I was hoping so but it doesn't hurt to ask 14:22:09 <tflink> i think the rest of it will be handled when we start going over tasking 14:22:29 <tflink> unless there are comments/questions about the disposable client demo or the remaining work 14:23:04 * tflink assumes not, moves on to tasking 14:23:14 <tflink> #topic Tasking 14:23:27 <tflink> who all is looking for taskotron work to do? 14:23:45 <garretraziel> I think that jskladan did 14:23:50 * handsome_pirate waves 14:23:52 <lbrabec> me too 14:23:54 <tflink> yeah, he mentioned that 14:24:09 <tflink> lbrabec: how do you feel about https://phab.qadevel.cloud.fedoraproject.org/T575 14:24:42 <tflink> kparal: we started talking about this on Friday but do you think you'll have time for https://phab.qadevel.cloud.fedoraproject.org/T568 ? 14:24:47 <kparal> what about creating json metadata for those images? 14:25:03 <kparal> I can do that easily in taskotron-vmbuilder 14:25:27 <tflink> what would the metadata be used for? 14:25:35 <kparal> instead of filename parsing 14:26:34 <kparal> just an idea 14:26:59 * handsome_pirate likes this idea 14:27:00 <tflink> when we (jskladan at least, don't recall off hand who else was there) were talking about this at Flock, the idea was that filename parsing isn't perfect but it's easy and pretty straight forward 14:27:09 <handsome_pirate> Could store it in db 14:27:24 <handsome_pirate> wow, have signal in tunnel 14:27:26 <kparal> well, fine by me 14:27:39 <tflink> is there an advantage that I'm not thinking of 14:27:40 <tflink> ? 14:27:53 <kparal> I'm talking about T575, just to be sure 14:27:59 <tflink> oh 14:28:10 <tflink> yeah, same thing 14:28:48 <tflink> I just really like the simplicity of filename parsing 14:28:53 <kparal> ok 14:29:15 <tflink> this is going to be complex enough as it is :-/ 14:30:05 <kparal> let's talk about T568 after the meeting, I'm still not clear about the use cases 14:30:14 <tflink> but if there is an advantage in using json metadata instead, I'm not totally against the idea 14:30:29 <tflink> it'd be better defined and less reliant on convention 14:31:29 <tflink> but we might have to add testcloud support for the json files. or at least make sure it doesn't puke when the json files are in it's backingstore/ dir 14:32:12 <tflink> thoughts? 14:32:21 <kparal> too many 14:32:23 <kparal> :P 14:32:59 <kparal> let's go with filename parsing for T575 for the time being 14:33:34 <tflink> my thought process was that if we run into problems, we can look into something more sophisticated 14:34:12 <tflink> lbrabec: are you OK with T575? 14:34:24 <lbrabec> tflink, yep 14:34:43 <tflink> cool, did I miss anyone? 14:34:54 <tflink> other than josef - will sync up with him after the meeting 14:35:29 <tflink> wow, this is going much faster than I thought it would :) 14:35:35 <tflink> but those are famous last words 14:36:13 <handsome_pirate> tflink: Ahoy 14:36:26 <handsome_pirate> tflink: Want me to do something? 14:36:53 <tflink> handsome_pirate: https://phab.qadevel.cloud.fedoraproject.org/T576 is a quick task that comes to mind 14:37:40 * handsome_pirate looks 14:38:34 <handsome_pirate> grabs 14:39:18 <tflink> any other comments/questions? 14:39:35 <tflink> ok, moving on to 14:39:43 <tflink> #topic Upgrading to Fedora 22 14:40:04 <tflink> we never really circled back to the conversation around whether to migrate to el7 or not 14:40:32 <tflink> and since f21 is going EOL soon, we need to upgrade our deployments to f22 before long 14:40:56 * tflink was thinking of doing it between beta and final 14:41:23 <kparal> there's still one month after F23 is out 14:41:35 <kparal> too short? 14:41:41 <tflink> shouldn't be 14:41:53 * nirik notes you should likely check dnf vs yum compat for anything you are doing on them... 14:42:09 <tflink> but our dev/stg systems are relying on HW that's going EOL at the end of october 14:42:23 <tflink> seems silly to reinstall the clients on dev/stg and not upgrade them at the same time 14:42:40 <tflink> nirik: have you heard anything about the dnf module in ansible not working? 14:42:52 <tflink> it's been working for me so far but I've not used it much 14:43:13 <nirik> it should work fine for the most part... was more meaning things in mock, etc... 14:43:20 <tflink> ah 14:43:50 <tflink> which reminds me of more fun 14:44:07 <tflink> nirik: thoughts on what to do about roles we depend on that are written for yum? 14:44:36 <nirik> we need to modify them to handle either... shouldn't be too bad I wouldn't think. 14:44:49 * tflink isn't sure doing 'when: is_fedora' and duplicating all the yum directives is the best idea every 14:44:51 <nirik> actually most of them would likely just keep working if you also installed yum-depreciated 14:45:10 <nirik> some folks handled this by making the module itself a variable... 14:45:16 <nirik> package_manager: dnf 14:45:38 <nirik> {{ package_manager }}: state=installed name=whatever 14:45:50 <tflink> didn't realize that would work 14:46:07 <tflink> my first thought was to use old-style syntax if it's still supported 14:46:10 <nirik> yeah, it supposedly does... we can figure something out tho. 14:46:34 <tflink> all the more reason to start working on the upgrade sooner than later :-/ 14:47:18 <tflink> any other thoughts/comments on the upcoming upgrade? 14:47:33 <nirik> yeah. I think I already fixed up base... not sure how much else is affected you use 14:47:55 <tflink> I'm not sure either, haven't tried looking much 14:48:37 * tflink will need to remember to work with the docs folks - they wanted to have their buildbot on el7 14:49:05 <tflink> #action tflink to sync up with docs team to make sure that we keep the buildbot roles compatible with el7 14:49:35 <tflink> any other thoughts/comments? 14:49:51 <Corey84> will pass on docs is mtg in -meeting btw 14:50:45 * kparal shakes his head 14:50:48 <tflink> Corey84: yeah, I don't think it's of immediate concern, though. IIRC, there are bigger fish to fry before we start worrying about compatibility 14:51:45 <tflink> ok, moving on to 14:51:51 <tflink> #topic Open Floor 14:53:15 <tflink> anything else that folks want to bring up? 14:54:11 * tflink looks through his notes from flock again to see if he missed anything 14:54:23 <tflink> oh, i did forget one thing 14:54:29 <tflink> #topic new taskotron logo 14:54:38 <tflink> will try to do this quickly 14:55:10 <tflink> we have a new logo for taskotron - https://mashaleonova.wordpress.com/2015/08/18/a-logo-for-taskotron/ 14:55:21 <tflink> one question is when to start using it 14:55:37 <lbrabec> relevant link https://phab.qadevel.cloud.fedoraproject.org/T584 14:55:53 * tflink would like to hold off for a bit until we at least have disposable clients in place, maybe dist-git tasks 14:56:00 <kparal> lbrabec: you can use #link 14:56:03 <tflink> my thought process goes something like this: 14:56:15 <tflink> 1) test/task automation doesn't really have a great reputation in Fedora 14:56:38 <tflink> 2) we haven't done much with Taskotron in production which AutoQA wasn't capable of (from a packagers perspective) 14:57:09 <tflink> 3) waiting until we have more bits that AutoQA could never do would add a visual change to emphasize that "hey, this isn't AutoQA. it's different and does much more" 14:57:27 <lbrabec> that sounds reasonable 14:57:47 * kparal agrees with all of that, but he wouldn't wait more than a month 14:58:01 <tflink> it would also give us a bit more time to do some frontend redesign/rework so that it doesn't look like we just slapped it on something 14:58:14 <kparal> it would be impolite to the logo author, in my eyes 14:58:27 <tflink> kparal: even if we tell her what we're thinking? 14:58:49 <lbrabec> also, with jsedlak we did quick draft https://db.tt/iXYwltw6 14:58:58 * kparal shrugs. sure 14:59:03 * tflink wouldn't think that saying "we want to use the logo you designed to emphasize a big change that's coming but isn't done yet" 14:59:06 <tflink> would be that bad 14:59:31 <tflink> ugh, bootstrap :-P 14:59:41 * tflink lost that battle a long time ago, though 14:59:43 <kparal> I'm not that optimistic that all of that will be in place in a short timeframe. I think we're again talking about 3-6 months 15:00:25 <tflink> kparal: I'm just worried that we'd be wasting a chance to have the visual emphasis that new things are in place 15:01:40 <kparal> I don't have strong feelings. just expressed my thoughts. if we wait and tell it to Masha and she's not disappointed about that, no big deal 15:01:59 <garretraziel> I can send Masha (author of our logo) draft of our webpage and tell her that we will use it in a few months 15:02:24 <tflink> yeah, I think that not telling her anything would be disrespectful 15:02:49 * garretraziel agrees 15:03:07 <tflink> garretraziel: I'd like to explain some of the reasons behind why we're not planning to use it immediately as well 15:03:39 <kparal> just a note, QA meeting has just started 15:03:51 <garretraziel> tflink: sounds reasonable 15:04:09 <tflink> but we're over time now 15:04:22 <tflink> and can finish this up outside of the meeting - sound reasonable? 15:04:33 <tflink> #topic Open Floor 15:04:36 <tflink> any other topics? 15:04:47 <tflink> otherwise, we're over time and the QA meeting has started 15:05:38 <tflink> thanks for coming, everyone 15:05:45 * tflink will send out minutes shortly 15:06:09 <tflink> feedback on how well the new method for status updates works would be appreciated 15:06:12 <tflink> #endmeeting