15:02:19 <tflink> #startmeeting fedora-qadevel
15:02:19 <zodbot> Meeting started Mon Feb 20 15:02:19 2017 UTC.  The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:19 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:19 <zodbot> The meeting name has been set to 'fedora-qadevel'
15:02:21 <tflink> #meetingname fedora-qadevel
15:02:21 <zodbot> The meeting name has been set to 'fedora-qadevel'
15:02:23 <tflink> #topic Roll Call
15:02:39 <tflink> #chair kparal roshi mkrizek garretraziel lbrabec
15:02:39 <zodbot> Current chairs: garretraziel kparal lbrabec mkrizek roshi tflink
15:02:45 <roshi> .hello roshi
15:02:46 <zodbot> roshi: roshi 'Mike Ruckman' <mruckman@redhat.com>
15:02:47 * garretraziel is here
15:02:51 * mkrizek 
15:02:51 * kparal is here
15:02:58 * lbrabec is here
15:04:11 <tflink> welcome to the yak shaving marathon, I trust that you all were issued a yak and a razor as you entered
15:04:31 <tflink> #chair jskladan
15:04:31 <zodbot> Current chairs: garretraziel jskladan kparal lbrabec mkrizek roshi tflink
15:04:32 * jskladan is here
15:04:46 <jskladan> this is better than #fedora-meetign-1 :)
15:04:54 * kparal received two razors and no yaks
15:05:14 <tflink> kparal: sounds like you now have 2 yaks to find
15:05:23 <tflink> :)
15:05:37 <tflink> lets get this party started and hopefully it'll end early-ish
15:05:42 <tflink> #topic Announcements and Information
15:05:50 <roshi> there's a butcher in town that can handle yak... you know, if you finish harvesting the hair :p
15:05:52 <tflink> #info allow for Jenkins to be used in the Taskotron stack WIP - mkrizek
15:05:52 <tflink> #link https://phab.qa.fedoraproject.org/T908
15:05:52 <tflink> #info depcheck crashes have been resolved - kparal
15:05:52 <tflink> #link https://phab.qa.fedoraproject.org/T906
15:05:52 <tflink> #info testcloud race condition leaving a lot of zombie VMs should be resolved - kparal
15:05:53 <tflink> #link https://phab.qa.fedoraproject.org/T905
15:05:54 <tflink> #info new resultsdb* pushed stable, should be updated on prod soon - kparal
15:05:56 <tflink> #link https://phab.qa.fedoraproject.org/T877
15:05:58 <tflink> #link https://phab.qa.fedoraproject.org/T890
15:06:17 <tflink> roshi: eh, it's extra fiber and adds a certain ... crunch
15:06:34 <tflink> any comments or questions?
15:06:55 <kparal> I also created a new testcloud release today, pending updates-testing
15:07:15 <roshi> #info tunirtests have been ported to ansible
15:07:17 <tflink> 0.1.11?
15:07:19 <roshi> #link https://pagure.io/fedora-qa/cloud-atomic-ansible-tests
15:07:21 <kparal> tflink: yes
15:07:23 <jskladan> I'd just add a call to read and comment on the trigger doc
15:07:39 <tflink> #info testcloud 0.1.11 has been released and pushed to updates-testing
15:07:47 <tflink> fewer zombie VMs!
15:08:04 <tflink> now it's just imagefactory that will leave around extra vms :)
15:08:15 <kparal> roshi: so, when we port tunirtests into ansible, is that something that upstream is going to maintainer, or will we have to? or is it just a wrapper that requires little maintenance?
15:08:18 <tflink> roshi ++
15:08:42 <tflink> from what I understand, though, roshi +50 and a case of beer
15:08:44 <roshi> the intent, aiui is for the cloud SIG to maintain the cloud tests and same for the atomic
15:08:46 <kparal> tflink: there might be still a few edge cases unresolved. but it seems to work fine atm
15:08:55 <tflink> glad to hear it
15:08:57 <roshi> I'll handle the sheaperding of that though
15:09:20 <roshi> but my goal is to get their input and feedback, as well as PRs for new tests and whatnot
15:09:39 <tflink> it'd be good to get that running on new atomic composes
15:09:40 <kparal> so, they'd maintain it as taskotron tasks, right?
15:09:59 <roshi> that'd be the plan, but at the very least they can maintain the playbooks
15:10:31 <kparal> just in a nutshell, why was the conversion even needed?
15:10:31 <roshi> sine my hope is that we just have a generic ansible_runner task that you feed a compose and url for the ansible repo and it runs
15:10:44 <roshi> like the atomic-host-tests
15:10:45 * kparal is not very familiar with tunirtests
15:11:01 <roshi> it turns out that it was a good thing I went through them
15:11:12 <roshi> there were several tests that weren't actually testing things
15:11:41 <kparal> hmm, ansible_runner task, sounds good
15:13:04 <roshi> I think ansible is easier to maintain for non-python people
15:14:00 <roshi> as well as if the cloud/atomic folks succeed in getting python out of the image (which has been a goal for a while, to reduce image size), we can use raw calls in ansible but would require a complete rewrite to port from the python unittests that tunir uses now
15:14:32 <kparal> was there any communication with kushal (who is I believe the primary author of tunirtests)?
15:14:52 <roshi> so I converted it for those reasons: 1) easier for outside patches, 2) future proofing, 3) allows us to run them in a generic runner, 4) the upstream atomic tests are already ansible
15:15:07 <roshi> there was, I told him that I was going to port them to ansible
15:15:27 <kparal> great reasons, thanks for explanation
15:15:57 <roshi> np, it wasn't exactly clear from the outside looking in
15:17:10 <tflink> cool, any other questions/comments/items?
15:18:19 * tflink takes that as a no
15:19:27 <tflink> #topic Trigger and ExecDB changes
15:20:06 <jskladan> ad execdb: I moved the "acked" doc from gdocs to a wiki page: https://phab.qa.fedoraproject.org/w/taskotron/execdb_rewrite/
15:20:25 <jskladan> There is a similar doc on Trigger: https://docs.google.com/document/d/1BEWdgm0jO4p5DsTO4vntwUumgLZGtU5lBaBX5av7MGQ/edit?usp=sharing
15:20:30 <jskladan> please read and comment
15:20:39 <tflink> you beat me to the links :)
15:20:40 <jskladan> (thanks kparal for posting some thoughts)
15:20:41 <kparal> jskladan: "phriction page" ;)
15:20:57 <jskladan> #link https://phab.qa.fedoraproject.org/w/taskotron/execdb_rewrite/
15:21:03 <tflink> sounds like someone is wanting to do some php work ...
15:21:05 <jskladan> #link https://docs.google.com/document/d/1BEWdgm0jO4p5DsTO4vntwUumgLZGtU5lBaBX5av7MGQ/edit?usp=sharing
15:21:25 * roshi dusts off his firefox google profile
15:22:11 <tflink> jskladan: other than getting comments, is there anything else you wanted to discuss about the proposals?
15:22:44 <jskladan> not really, just wanted to scout whether there is somebody longing to do the execdb rewrite
15:22:53 <jskladan> if not, I'll do it myself (should not take more than a couple of days)
15:23:10 <jskladan> but I won't be putting together comprehensive list of ticket just to work on them myself
15:24:29 <jskladan> so, seems like no volunteers, so I'll be doing the rework this week
15:24:54 <jskladan> #task jskladan to do the execdb rewrite this week (don't forget to create at least high-level tickets to track progress)
15:25:12 <jskladan> (is this how you do the action items? /me does not really remember)
15:25:21 <tflink> it's #action
15:25:22 <kparal> it's #action
15:25:26 <kparal> damn
15:25:27 <tflink> jinx!
15:25:27 <jskladan> aaah
15:25:33 <jskladan> #action jskladan to do the execdb rewrite this week (don't forget to create at least high-level tickets to track progress)
15:25:36 <jskladan> thx
15:25:59 <tflink> anything else on this topic?
15:26:06 <jskladan> so, please, read the docs, comment, and I'll do my best to make it final & transfered to a "phriction page" once it is done
15:26:10 * jskladan out
15:26:23 <tflink> sounds good
15:26:34 <tflink> thanks for working on these proposals
15:27:23 <tflink> #topic Meta CI for libtaskotron
15:27:44 <tflink> jskladan has been busy the last couple weeks :)
15:27:47 <tflink> #link https://pagure.io/taskotron/task-taskotron-ci
15:28:05 <kparal> jskladan++
15:28:09 <jskladan> the framework is ready, we need some "real"-ish tests now
15:28:33 <jskladan> I'd be glad if kparal took the lead on that, but not forcing...
15:28:35 <tflink> my goal here was to highlight the meta-task and see if there are any ideas about where to run it (in taskotron, in phab etc.)
15:29:02 <jskladan> what do you mean by "where to run it"?
15:29:13 <jskladan> it is a taskotron task, so I'd run it there :)
15:29:31 <jskladan> trigger changes need to be deployed (new consumer) for that, but those are all the changes I know of
15:29:53 <tflink> jskladan: but that's crazy talk!
15:30:15 <jskladan> in the future, I'd love to be able to trigger on PRs in pagure or DRs in phab instead of the pushes
15:30:45 <tflink> agreed but this is still an improvement
15:30:46 <jskladan> but I'm not sure whether putting much work into the Phab thing is reasonable, since we are still not sure about keeping it
15:30:53 <tflink> agreed
15:30:54 <kparal> I wanted to concentrate more on the user experience part. but I can try to come up with some "real-ish" task
15:31:25 <jskladan> kparal: what do you mean by "user experience part"?
15:31:41 <kparal> I meant the overall user experience with taskotron results. unrelated to CI.
15:31:52 <kparal> readable task artifacts, etc
15:32:30 <jskladan> ok
15:32:37 <jskladan> whatever you think makes sense
15:32:46 <jskladan> the task now does the basic sanity check IMO
15:32:53 <jskladan> so it is not of a high urgency
15:33:01 <kparal> I can look through our example tasks and pick something that would be good to run in the CI
15:33:10 <jskladan> just for the record: kparal mentioned it might be great to use the example tasks there
15:33:16 <jskladan> ah, ninjad again...
15:33:34 <kparal> the current ninja ladder is: tflink > kparal > jskladan
15:33:37 <tflink> that sounds good to me as well
15:34:03 <tflink> kparal: hush, the first rule of ninja is not telling anyone about it
15:34:12 <roshi> lol
15:34:20 <tflink> the second rule is not talking about fight club, though
15:34:22 <tflink> anyhow
15:34:26 <jskladan> ..thus kparal falls of the ladder, as the real, invisible ninjas chop it from underneath him...
15:34:31 <tflink> :)
15:35:16 <tflink> #info we do need more tests in the CI task, the example tasks may be a good place to start
15:35:22 <tflink> anything else on this?
15:36:35 <jskladan> nope
15:36:55 <tflink> cool, moving on
15:37:12 <tflink> #topic Documentation
15:37:46 <tflink> I don't have much to say here, just a note that we should go through our docs and make sure that everything is up to date, not showing more "warning, this stuff is alpha software" etc.
15:38:04 * tflink is planning to make time for it later this week
15:38:07 <jskladan> yay! docs!
15:38:11 <kparal> we're no longer alpha?
15:38:16 <jskladan> no, we are pre-alpha
15:38:20 <kparal> right!
15:38:31 <kparal> we should not claim alpha then
15:38:44 <jskladan> or maybe pre-ci-post-cd-weekly-release-rolling project
15:39:00 <tflink> not enough buzzwords
15:39:16 <jskladan> also microservice-enabled, customizable and modular
15:40:01 <roshi> in the cloud!
15:40:04 <jskladan> !!
15:40:12 <tflink> continuous integration and delivery for a modular, microservices based framework that is rapidly innovating and moving to empower customers in a new age of container based infrastructure
15:40:22 <jskladan> love it! let's make t-shirts
15:40:22 <roshi> it promotes reverse synergy
15:40:35 * tflink just died a bit writing that
15:40:39 <roshi> is .word a TLD?
15:40:39 <jskladan> :D
15:40:43 <roshi> we should get buzz.word
15:41:22 <jskladan> getting back to the old-boring-stuff - as kparal mentioned in the Trigger doc, it might be usefull to get that "initial state" section, and use it as docs for trigger
15:41:36 <jskladan> (or as base for the docs)
15:41:56 <jskladan> I know some people (threebean?) were asking for them lately
15:41:59 * kparal filed https://phab.qa.fedoraproject.org/T912 for the CI "real"-ish task (which is its official name now, including the quotes)
15:42:02 <jskladan> so...
15:42:32 <tflink> think it'd be worth creating a sphinx projects for resultsdb and trigger?
15:42:57 <kparal> jskladan: I was thinking we could simply create "docs" directory and store that as a markdown document. pagure can render it, and it would be very quick to set up. we can convert it to proper docs later
15:43:05 <jskladan> dunno. I just don't feel like rewriting the API docs from Apiary to Sphinx, but if there are volunteers... :)
15:43:23 * tflink has no idea how much work that'd be
15:43:32 <kparal> resultsdb has enough docs, now?
15:43:41 <kparal> I thought we're mainly talking about trigger here
15:43:48 <kparal> *now -> no?
15:44:00 <jskladan> well, if people can use the Apiary doc, then yes, it does
15:44:24 <tflink> we can deal with that at another time if it makes sense
15:44:25 <jskladan> but from the reactions I'm getting all over the place... *sigh* if only apiary did not make it so easy to write those things..
15:44:33 <kparal> don't worry, it's usable. we were teasing you a bit :)
15:44:54 <tflink> +1 to putting a markdown/rst doc in pagure for now
15:44:56 <kparal> (if you know the headers are clickable)
15:44:58 <tflink> for trigger
15:45:41 <tflink> anything else on this?
15:46:59 <tflink> I assume that's a no
15:47:16 * roshi has nothing
15:47:17 <tflink> kparal: you OK taking the action item to put the current state stuff into a file in pagure for trigger?
15:47:37 <kparal> sure
15:47:50 <kparal> copy paste is my thing
15:48:06 <tflink> #action kparal to take information from trigger proposal and make a quick trigger doc out of that in pagure
15:48:15 <tflink> #topic Open Floor
15:48:33 <kparal> I have one topic
15:48:35 <tflink> I feel like I'm forgetting a topic for this week but as I can't remember what that might have been ...
15:48:40 <tflink> kparal: go for it
15:48:59 <kparal> I was thinking about some of you moving teams, and what it means for maintenance of certain of our apps
15:49:00 <jskladan> burrito kittens: http://static.boredpanda.com/blog/wp-content/uploads/2016/08/cute-kittens-60-57b32505e2b5e__605.jpg discuss!
15:49:18 <tflink> jskladan: looks tasty :-P
15:49:25 <kparal> since taskotron is hopefully continuing unchanged, but I guess you'll not have time to maintain everything
15:49:48 <kparal> so far I'm thinking about blockerbugs and testday app
15:49:58 <tflink> at this point ... who knows but I suspect that you're correct
15:50:12 <kparal> is there anything else that's QA-specific and not automation specific that we should add to that list?
15:50:15 <jskladan> testday app needs a redo anyway, does not work on top of resutlsdbv2
15:50:15 <tflink> adamw has been working on blockerbugs lately :)
15:50:20 <roshi> I think those are it
15:50:51 <jskladan> and it would make sense to stop trying to emulate the wiki pages, and actually sit down and have a conversation about the needs and requirements for that
15:50:52 <kparal> ok, so right now I just wanted to get a list of such tools, and we might start thinking who's going to be the new owner
15:50:54 <jskladan> but...
15:50:58 <kparal> or even ask in the community
15:51:00 <tflink> unless you count testcloud but that's roshi's baby anyways
15:51:36 <kparal> jskladan: are we able to continue using testday app unchanged, even with resultsdb v1 backend?
15:51:43 <jskladan> the question still stands - tflink is most probably leaving us for sure, so there is this huge void of missing sysadmin....
15:51:53 <garretraziel> maybe we can consider blockerbugs app as a target for platform-ci team (or what the hell is its name now)?
15:52:03 <jskladan> kparal: as long as there is a resultsdbv1 backend deployed, yes
15:52:04 <tflink> there have been repeated assurances that there will be no drastic changes
15:52:13 <jskladan> it won't work with the v1 fallback in v2, though (no write access)
15:52:28 <kparal> jskladan: does it depend on fedora packaging? or how do we have it deployed?
15:52:28 <tflink> but that doesn't mean that planning for change is a bad thing
15:52:39 <roshi> I'll also be working with tim to pick up slack as it comes around
15:52:46 <jskladan> kparal: who am I to know? Some kind of sysadmin fella? :D
15:52:53 <jskladan> but honestly, I don't know
15:53:05 <kparal> ok, so that's one thing to figure out
15:53:08 <tflink> jskladan: you're the one with the certification :-P
15:53:23 <jskladan> is it rhel6? :)
15:53:45 <tflink> jskladan: that's your new job - porting everything to work on RHEL6 and use that as the primary systems
15:53:48 * tflink runs
15:54:14 <jskladan> OK
15:54:15 <garretraziel> yay for python<2.7!
15:54:25 * jskladan prepares the flammenwerfer
15:54:33 <roshi> sudhir had asked me to be ready to cover for Tim as things shift
15:54:46 <roshi> kparal: ^^
15:55:10 <roshi> so I'm just kinda waiting until it gets to that point, and we actually know what all will need it
15:55:27 <tflink> yeah, so much is up in the air right now
15:55:27 <kparal> ok, I just wanted to raise awareness. no need to resolve that now
15:56:26 <tflink> anything else for open floor?
15:56:37 <tflink> the last 4 minutes of open floor, rather
15:56:44 <roshi> nope :)
15:56:45 * tflink did think it was going to be a shorter meeting today
15:57:05 <tflink> ok
15:57:09 <tflink> thanks for coming, everyone
15:57:13 * tflink will send out minutes shortly
15:57:18 <tflink> #endmeeting