14:02:37 #startmeeting fedora-qadevel 14:02:38 Meeting started Mon Jun 8 14:02:37 2015 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:02:38 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:02:38 #meetingname fedora-qadevel 14:02:38 #topic Roll Call 14:02:38 The meeting name has been set to 'fedora-qadevel' 14:02:43 * kparal is here 14:02:46 * mkrizek here 14:03:11 * kparal pokes jskladan 14:03:19 here! 14:03:23 yay! 14:04:08 we have a "full" house today :) 14:04:17 who wants to go first for status update? 14:04:26 I'll go 14:04:35 #topic mkrizek status report 14:04:48 missing chair 14:05:11 sorry 14:05:17 #chair mkrizek kparal jskladan 14:05:17 Current chairs: jskladan kparal mkrizek tflink 14:05:19 #topic mkrizek status report 14:05:28 #info addressed kparal's concerns about remote execution: 14:05:28 #link https://phab.qadevel.cloud.fedoraproject.org/D380 14:05:28 #info sent a patch for allowing koji directive to download debuginfo rpms 14:05:31 #link https://phab.qadevel.cloud.fedoraproject.org/T494 14:05:33 #link https://phab.qadevel.cloud.fedoraproject.org/D381 14:05:36 #info reported failing test-suite on disposable-develop (my bad, working on a fix) 14:05:39 #link https://phab.qadevel.cloud.fedoraproject.org/T495 14:05:41 #info started to work on distgit directive 14:05:44 #link https://phab.qadevel.cloud.fedoraproject.org/T492 14:05:46 #info fixed incorrect links to execdb in the resultsdb frontend 14:05:49 #link https://phab.qadevel.cloud.fedoraproject.org/T469 14:06:06 great work 14:06:27 agreed :) 14:06:44 if no questions, I'll continue 14:07:09 * roshi steps in a bit late 14:07:10 no questions from me, need to test the koji debug directive, though 14:07:25 I need to test it - just making sure I was clear 14:07:44 ok 14:07:50 #topic kparal status update 14:08:00 #info caught up with disposable clients development 14:08:04 #info some reviews and feedback about latest diposable-clients changes 14:08:07 #info spotted that Bodhi started to display extra \n characters for Taskotron comments, reported, Luke fixed it shortly afterwards 14:08:10 #info discussing the proposed abidiff check 14:08:14 #info now looking at the new formula syntax for disposable clients 14:09:15 I guess we'll talk about the fedmsg/abidiff later according to the agenda 14:09:16 cool, thanks for catching the comment display issue - did they ever figure out what happened? 14:09:49 tflink: I think the root cause is still not known, they just used the same cover-up hack they used elsewhere earlier 14:09:58 kparal: if we're ok with having 2 longer-ish meetings in a row, yeah 14:10:00 to stop it from displaying 14:10:23 * tflink shrugs, looks forward to bodhi2 14:10:34 right, good enough fix for me 14:10:57 any other questions/comments? 14:11:12 none here 14:11:16 nope 14:11:32 jskladan: you and I are left 14:11:33 #link https://github.com/fedora-infra/bodhi/issues/147 14:11:36 #topic jskladan's update 14:11:36 #info finished the Testdays revamp to a state where it provides the same functionality as the first version, but with flask and current resultsdb 14:11:36 #info catching up with Phab tickets and diffs 14:11:36 #info in need of tickets 14:11:47 mine is short, sadly 14:12:30 jskladan: any thoughts on when testdays will be ready to redeploy? 14:12:44 and is there a migration path for the data on the old testdays instance? 14:13:25 the old data can be converted to wiki, I believe 14:13:26 *shrugs* can be easily done tomorrow, if we wanted to IMHO. I do not really think that migrating the data is necessary - just make sure that all the events got exported to the wiki matrices 14:13:48 I think that's good enough 14:14:04 sounds good 14:14:16 let's just make sure it's all converted before deleting the instance 14:14:19 is there a dev/stg instance up? 14:14:26 *exported 14:15:11 where do we want to host the new testdays instance? 14:15:42 * tflink thinks it's a good candidate for putting into infra 14:15:51 unless the wiki export needs shell access 14:16:11 it is quite lightweight, we could easily deploy it next to any other flask stuff we run (same machine as resultsdb-frontend would be my first choice) 14:16:21 the wiki export is done over the web-ui 14:16:45 is it going to share a resultsdb instance with taskotron? 14:16:47 it just creates the wiki data, which are supposed to be manually inserted to the testday's wiki 14:17:21 it can (and I'd go that way), but it is not a necessity 14:17:33 I just do not think there is a need for separate instance 14:17:53 my first thought would be to isolate it from taskotron but I'm open to other ideas 14:18:06 tflink: I don't really care either way, to be honest 14:18:26 * tflink thinks we can figure this out outside of the meeting 14:18:26 I'd rather stay safe and used separate dbs, but I'm not the db expert here 14:18:41 it generates such low amounts of traffic to the resultsdb, that it can easily be either absolutely self-contained, or shared 14:18:46 does it run on el7? 14:19:17 resultsdb or testdays? 14:19:28 both, I suppose - more testdays 14:19:45 haven't tried that, really. But it is based off of the same flask-skeleton 14:20:04 and I'm IMHO not using anything too fancy on the Python side 14:20:19 I'd be more worried about any extensions 14:21:08 we can try, I can't really give an knowledgeable answer - I have not been dealing with python/flask on el7 ever 14:21:10 jskladan: perhaps we can continue this after the meeting - are you going to be around for a bit longer? 14:21:40 tflink: i'll need to check my bus schedule, but I'm not leaving straight after 14:22:38 k, any other questions/comments? 14:23:16 would testdays benefit being stuffed into a docker container? 14:23:47 other than the buzzword powerup? I suspect not 14:23:47 roshi: I'm fairly certain that docker is a no-no for Infra 14:24:11 yeah, AFAIK nothing in infra is run w/ docker 14:24:27 well, I was thinking more for the isolation and it'd be easier for people to hack on since they could launch the container and poke at it :) 14:24:30 ah 14:24:34 nvm then :) 14:25:04 we could add it to the taskotron locally-hacked playbook, I suppose 14:25:47 any other questions/comments? 14:26:09 not here 14:26:29 * tflink assumes silence == no, moves on to status report 14:26:34 #topic status report - tflink 14:26:34 #info reboots and updates during outage 14:26:35 #info still working to get testCloud ready for another release, need to figure out what to do about polkit 14:26:35 #info started on code to get libtaskotron to launch VMs 14:26:35 #info some work on the new beaker client - networking happy fun time 14:26:35 #info discussion and some planning around an ABI checking task 14:26:37 #info wrote blog post about taskotron progress, plan to start doing that more regularly http://tirfa.com/taskotron-update-june-2015.html 14:27:33 thanks for the blog post 14:27:39 note that for some reason, planet hasn't picked up the post - will be looking into that later today 14:28:21 any questions/comments? 14:28:52 none here 14:28:55 nothing here 14:29:13 ok, lets move on 14:29:32 #topic Fedora QA Report to the Council 14:29:47 just wanted to bring this up quickly to make sure we didn't miss anything for today's report 14:30:05 I figure most folks are paying attention in #fedora-qa and/or test@, though 14:31:51 I take that as a "nothing to add or dispute" 14:31:53 * kparal just found out he also missed tflink's email into the test list 14:32:27 do we want to spend the time to go over the proposed abidiff stuff or fedmsg duplication? 14:32:30 I'll try to read it asap and ping roshi if I have anything to add 14:33:01 * tflink just realized that he didn't finish his post to qadevel@ about fedmsgs on friday :-/ 14:33:40 it depends if folks are up to date with the email discussions 14:34:00 if not, there's no point in discussing it right now, I guess 14:34:14 email and phab discussions, to be exact 14:34:17 kparal: was there something you wanted to go over WRT abidiff? 14:34:33 * mkrizek isn't up to date with the abidiff discussion in phab 14:34:39 not really, I think we reached good conclusions in the ticket 14:35:08 there's just one thing to add - that koji->bodhi bridge task might be a bit more complicated than it seems. but we don't need it right now, so no big deal atm 14:35:41 kparal: what's the issue you're seeing? 14:35:48 seeing/anticipating 14:37:01 I was thinking about how you realize which bodhi updates are up-to-date with taskotron results information, and which are not, without constant crawling through bodhi 14:37:18 which is very slow and error-prone (server often sending 500 response codes) 14:37:22 doesn't bodhi emit fedmsgs? 14:38:05 * tflink figured that we could listen for "update creation" or "update push to testing" messages and work off of those 14:38:33 it does, but I'm not sure it will be enough. but as I say, this is not an immediate concern, I just wanted to point out that it might be a potential problematic point once we get to it 14:38:53 I don't have anything prepared or well though out now 14:39:02 *thought 14:39:03 ok, we can deal with it if/when we have to get that done 14:39:14 but I think it's past time for ... 14:39:19 #topic Open Florr 14:39:22 #undo 14:39:22 Removing item from minutes: 14:39:26 #topic Open Floor 14:39:26 I liked it 14:39:46 sounded... pirate-y 14:39:54 :D 14:40:08 arr! i'll be wiping the florr with you scurvy dogs! 14:40:14 or something like that :-P 14:40:38 * kparal always expanding his dictionary 14:40:43 anyhow, other stuff to bring up? 14:41:21 I take that as a no 14:41:25 * kparal typing 14:41:31 * tflink waits 14:42:11 so, jskladan said he's in need of tickets, I wonder, would this be a good time to finally finish that "logs as artifacts" feature, where we link the particular log directly from the execdb/resultsdb overview page? 14:42:28 I think we're very close, I'm not aware of any roadblocks 14:42:56 doing a s/taskbot/taskotron on the templates would also be good 14:43:02 not sure I ever filed a ticket on those 14:43:06 those/that 14:43:19 provided jskladan is interested in working on that 14:43:23 tflink: what exactly do you mean? 14:43:31 sure 14:43:44 kparal: resultsdb_frontend has a "taskbot url" column 14:43:47 hello fedora team. i've joined the qa team (sponsored by jedi master roshi) and i'm also interested in sys admining, packaging, or anything involving shell :) if there's anything i can do, please count on me. 14:44:20 fcon: welcome! is there anything in particular that you'd be interested in? 14:45:06 kparal: do we want to link to taskotron.log or stdout? 14:45:20 tflink: thanks! well, i'm very open - roshi told me about helping in the cloud with some Dockerfiles 14:45:21 * kparal looking for a ticket 14:46:01 there are some tasks tracked here: https://phab.qadevel.cloud.fedoraproject.org/T399 14:46:12 roshi: can you elaborate a bit? I'm not sure what you had in mind 14:46:41 oh, I see what you meant 14:46:42 well, I was pointing out general areas he could get plugged into 14:47:05 kparal: you mean get depcheck to start generating more separated logs for artifacts? 14:47:09 and since I know we could use more admins for stuff on the devel side of things, I thought I'd plug him in here 14:47:13 or am I misunderstanding 14:47:21 my idea was that we would link this page from bodhi comments: http://taskotron-dev.fedoraproject.org/resultsdb/results/1251713 . and that page would contain links to createrepo_c.log file (created by depcheck or upgradepath) as well as usual full taskotron.log/stdio 14:48:24 kparal: ah, I did misunderstand - that sounds good to me 14:48:46 so people would go to this "resultsdb item landing page", will all relevant artifacts linked, and also links to more details like debug logs, if needed 14:49:00 s/will/with 14:49:39 not sure if the ticket for this is created, but I can make sure it will be 14:49:47 did we ever decide whether we wanted to link to artifacts dir from resultsdb_frontend? 14:49:57 kparal: sounds like a plan to me 14:50:14 * kparal will create a ticket and link it in T399 14:50:43 fcon: are you going to be around for a bit after the meeting ends? 14:51:03 tflink: I would rather link to the specific artifacts that the whole dir in general, so that people don't need to hunt for the relevant bits 14:51:11 *than 14:51:13 tflink: yep 14:51:48 the relevant artifacts would be defined for each item in TAP 14:52:10 fcon: I'll ping you shortly after the meeting ends in #fedora-qa to figure out more of what tasks/work/etc that we have which you might be interested in 14:52:30 which reminds me that we haven't discussed the proposal to move away from TAP much yet 14:52:37 but let's not start that today :) 14:52:51 we can't do everything at once :) 14:52:56 deal 14:53:07 #action kparal to file ticket about linking directly to artifacts from resultsdb, attache to T399 14:53:38 anything else for open floor? 14:54:19 * tflink assumes not, ends meeting 6 minutes "early" 14:54:21 :) 14:54:28 heh 14:54:33 thanks for coming, everyone 14:54:38 * tflink will send out minutes shortly 14:54:39 thanks 14:54:41 #endmeeting