16:00:42 <kozhukalov> #startmeeting Fuel
16:00:43 <openstack> Meeting started Thu Mar 26 16:00:42 2015 UTC and is due to finish in 60 minutes.  The chair is kozhukalov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:44 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:47 <openstack> The meeting name has been set to 'fuel'
16:00:50 <kozhukalov> #chair kozhukalov
16:00:51 <openstack> Current chairs: kozhukalov
16:01:00 <kozhukalov> hey guys
16:01:03 <mattymo> hi all
16:01:04 <ikalnitsky> o/
16:01:06 <mkwiek> hello
16:01:10 <kozhukalov> agenda as usual
16:01:20 <kozhukalov> #link https://etherpad.openstack.org/p/fuel-weekly-meeting-agenda
16:01:36 <kozhukalov> is anybody else here?
16:01:39 <agordeev> hi
16:01:44 <akislitsky_> hi
16:01:55 <vkramskikh> hi
16:01:59 <kozhukalov> great at least 6 persons
16:02:07 <kozhukalov> let's start
16:02:09 <akasatkin_> hi
16:02:09 <alex_didenko> hi, 7 now
16:02:11 <dpyzhov_> hi
16:02:15 <angdraug> o/
16:02:21 <kozhukalov> #topic External ubuntu mirrors and pinning (ikalnitsky, brain461)
16:02:29 <prmtl> o/
16:02:45 <ikalnitsky> i'm glad to announce that consume-external-ubuntu feature is in master. it finally merged.
16:02:52 <mattymo> +1
16:02:57 <ikalnitsky> though, there are some improvements we have to do
16:03:14 <kozhukalov> ikalnitsky: including?
16:03:29 <ikalnitsky> for instance, we should allow users to change repos after deployment
16:04:07 <ikalnitsky> or we should add some warning banner that he need additional steps to run his ubuntu cloud
16:04:32 <ikalnitsky> except this, we have issues with apt-pinning
16:04:52 <ikalnitsky> currently, we're using pin template (which is hardcoded in Nailgun)
16:05:08 <kozhukalov> also build image script needs to be re-written in python and even more today bvt failed because of this script
16:05:17 <kozhukalov> btw, floating bug
16:05:31 <ikalnitsky> and it may lead us to some issues, when the remote repo has another.. properties
16:05:32 <kozhukalov> we are working on improving this part
16:05:34 <ikalnitsky> attributes
16:05:49 <ikalnitsky> looks like that's it from my side.
16:06:10 <maximov> to change repos after deployment on master node or on all slaves?
16:06:13 <ikalnitsky> and yeah, we obviously need to sync our jobs with patching story
16:06:25 <ikalnitsky> maximov, on all slaves
16:06:42 <ikalnitsky> for instance, user wants to add new nodes to existing env
16:06:50 <angdraug> ikalnitsky: why not simplify apt pinning preferences by not having separate entry for each component?
16:06:52 <ikalnitsky> and his repos aren't available
16:07:07 <angdraug> we still have the same priority for all components in a single archive
16:07:32 <ikalnitsky> angdraug, what do you mean exactly?
16:07:41 <ikalnitsky> it looks like we misunderstood each other
16:08:00 <angdraug> you have separate pinning entries for main and restricted
16:08:04 <angdraug> both pinned at 1050
16:08:09 <ikalnitsky> yep
16:08:20 <ikalnitsky> it doesn't work in one entry
16:08:21 <angdraug> if you don't specify c=main explicitly, such entry would cover all components of an archive
16:08:49 <kozhukalov> angdraug: because we have template
16:09:19 <kozhukalov> angdraug: and it is not very beautiful to have ifs for plenty of different cases
16:09:33 <angdraug> now you've lost me
16:09:34 <kozhukalov> it is better to have a loop over all sections
16:10:11 <ikalnitsky> guys, here's what we have for mos component
16:10:13 <ikalnitsky> #link http://xsnippet.org/360594/raw/
16:10:36 <ikalnitsky> as far as i understand, angdraug means why do we have two pins declarations instead of one?
16:10:39 <kozhukalov> the issue with pinning now is that we are not able to have different priorities for different repos if they both have the same suite and the same section
16:10:40 <ikalnitsky> am i correct?
16:10:49 <angdraug> and you could have: http://xsnippet.org/360595/
16:11:18 <ikalnitsky> angdraug, and what if i want to pin main from mos6.1 with another priority?
16:11:29 <ikalnitsky> i just add additional repo
16:11:34 <ikalnitsky> with exactly main section
16:11:40 <ikalnitsky> and set another prio
16:11:49 <ikalnitsky> different from restricted section
16:11:55 <ikalnitsky> what i have to do then?
16:11:56 <angdraug> on my debian system, I have different pin priorities for stable, testing, and unstable
16:12:06 <angdraug> they each have the same set of sections
16:12:09 <angdraug> works fine
16:12:25 <kozhukalov> for debian yes
16:12:25 <angdraug> why would you ever want a different pin priority for main and restricted?
16:12:40 <kozhukalov> but sections are not fixed
16:12:47 <ikalnitsky> angdraug, why not? it's possible technically
16:13:00 <ikalnitsky> it doesn't cause any problems
16:13:43 <kozhukalov> the issue is that it looks like we don't have enough data for setting pinning properly
16:13:47 <angdraug> no, no immediate problems, just giving user more ways to shoot themselves in a foot, and making it harder to read the preferences file
16:13:58 <kozhukalov> url + suite + section is not enough
16:14:31 <angdraug> this is taking too much time, lets discuss it later
16:14:56 <kozhukalov> it looks like in the future it is better to download Release file and set pinning according to metadata in this file
16:15:07 <kozhukalov> yes
16:15:13 <ikalnitsky> move on guys
16:15:22 <kozhukalov> if there are no other q let's move on
16:15:46 <kozhukalov> #topic IBP (agordeev, kozhukalov)
16:15:50 <agordeev> hi
16:16:05 <agordeev> I'll start with the listing of current IBP activites:
16:16:14 <agordeev> 1) Rewriting fuel-image in python is in progress now. ETA 1w
16:16:16 <agordeev> 2) Work on progress bar hasn't been started. We are going to ask some help from our fuel-ruby team on next week.
16:16:18 <agordeev> 3) Pinning for image building process isn't finished yet. We are waiting for help from more experienced engineers.
16:16:20 <agordeev> 4) Our ubuntu BVT is red for today due to fuel-image script breakage, I'm working on its fixing, ETA today
16:17:15 <kozhukalov> agordeev: is it still red?
16:17:25 <kozhukalov> i thought it is a floating bug
16:17:41 <ikalnitsky> yep
16:17:45 <ikalnitsky> i see that bvt are green
16:17:51 <ikalnitsky> though ubuntu smoke is still red
16:18:07 <kozhukalov> ok, needs to be fixed
16:18:10 <agordeev> kozhukalov: i don't know. I'm fixing the script
16:18:13 <ikalnitsky> agordeev, please ping devs for review when it's ready
16:18:22 <ikalnitsky> it's important to merge it asap
16:18:22 <agordeev> ikalnitsky: sure, i'll do
16:18:39 <kozhukalov> the error was that build image script could not find loop device
16:19:24 <kozhukalov> agordeev: thanx for working on this
16:20:18 <kozhukalov> and looks like some people still don't see the advantages of IBP, i think we need to write kind of blog post maybe
16:20:28 <kozhukalov> ok, are there any other q?
16:20:59 <agordeev> kozhukalov: the advantages for users to be exact
16:21:35 <kozhukalov> let's skip next topic since aglarendil is not available right now
16:21:52 <kozhukalov> agordeev: yep
16:22:08 <kozhukalov> #topic Bugs status (dpyzhov)
16:22:37 <kozhukalov> dpyzhov_: how far are we from scf? how many bugs are there?
16:22:56 <dpyzhov_> Well, we have five days until SCF
16:23:13 <dpyzhov_> and we have a huge number of bugs
16:23:54 <kozhukalov> some of them are not actually bugs but feature requests. right?
16:24:06 <dpyzhov_> about 700 confirmed bugs and almost 150 bugs in progress
16:24:38 <dpyzhov_> We are trying to separate tech debt and feature requests from this list
16:24:51 <dpyzhov_> I've added 'feature' tag for such stuff
16:25:22 <dpyzhov_> 'feature' means: this bug needs either proper design or it is not a bug at all
16:25:45 <dpyzhov_> Also I'm not sure if we really have 150 bugs in progress
16:26:21 <dpyzhov_> It's a really big number and maybe some of them are orphaned
16:27:05 <angdraug> any in progress bug that wasn't updated for more than a week is orphaned
16:27:27 <dpyzhov_> I think that we really need to walk through 'feature' bugs and decide how to deal with them\
16:27:31 <angdraug> any in progress bug that wasn't updated for more than a day is probably orphaned too
16:27:48 <dpyzhov_> Most of this bugs are travelling from one release to another
16:28:20 <angdraug> are they attached to blueprints?
16:28:27 <kozhukalov> guys, what do you think of the idea to try to deal with loads of bugs using kind of bayesian algorithm?
16:28:43 <angdraug> write an AI to fix bugs?
16:28:45 <kozhukalov> at least for categorizing them
16:28:57 <dpyzhov_> angdraug: yes, we should ping owners of hanged bugs. no, most of this bugs are just suggestions
16:29:26 <kozhukalov> angdraug: don't exaggerate me
16:29:34 <dpyzhov_> our workflow for blueprints is even worse then our workflow for 'feature' bugs
16:29:36 <angdraug> kozhukalov: sorry, couldn't resist :)
16:30:04 <angdraug> I'm just not sure "feature" tag is necessary, we already have wishlist priority for such thing
16:30:29 <dpyzhov_> I think that our list of blueprints was not revised for a long time
16:31:02 <dpyzhov_> angdraug: actually some of 'feature' bugs represent real user pain
16:31:23 <dpyzhov_> so we need to have priorities for this wishlist
16:31:24 <angdraug> if they are more important than a wishlist bug, they must have a blueprint
16:32:04 <dpyzhov_> angdraug: maybe
16:32:37 <dpyzhov_> angdraug: so your proposal is to start working with blueprints more carefully?
16:32:46 <angdraug> yes
16:33:09 <angdraug> maybe make it part of bugs triage duty, or have a separate duty
16:33:20 <kozhukalov> how are you guys able to read them? i mean do you sleep sometimes?
16:34:01 <angdraug> most of our blueprints only have a title, so not a big deal :p
16:34:15 <angdraug> all you need is ESP to read the mind of blueprint creator )
16:34:30 <kozhukalov> -)
16:35:04 <angdraug> seriously, identifying bugs that should be associated with a blueprint should be a part of bugs triage
16:35:15 <angdraug> triaging blueprints themselves is an entirely different story though
16:35:45 <angdraug> but I don't see a way around it
16:35:53 <dpyzhov_> angdraug: we have 381 blueprint
16:36:07 <kozhukalov> no so many, right? -)
16:36:15 <docaedo> holy smoke thats a lot
16:36:28 <angdraug> tbh we have even more product backlog items
16:36:54 <dpyzhov_> so it's about 1000 user stories
16:37:12 <angdraug> most of them overlap, so probably more like 500
16:37:28 <kozhukalov> is there a chance we address them all somewhen before end of the Universe?
16:37:30 <dpyzhov_> should we throw away half of them?
16:37:43 <angdraug> we're addressing several dozen in each release, don't we?
16:37:56 <angdraug> so we only got 2-3 years worth of user stories there
16:38:28 <angdraug> and yes, we should throw away some of them, but it's only part of the problem
16:38:29 <dpyzhov_> angdraug: but we getting new stories faster
16:39:01 <angdraug> that's normal, as long as we have a process in place to prioritize them
16:39:24 <angdraug> which is exactly what product backlog is for
16:40:00 <angdraug> we do need something like that on the community side of the project, though
16:40:16 <angdraug> xarses: around?
16:40:27 <kozhukalov> ok, guys, maybe it is better to move our chatter to the open discussion
16:40:37 <kozhukalov> looks like we are done with bugs
16:40:39 <dpyzhov_> so we should convert all 'feature' bugs to blueprints, add blueprints to the backlog and let backlog owners deal with it?
16:40:55 <angdraug> all feature bugs that are higher priority than wishlist
16:41:12 <angdraug> I think it's fine to leave wishlist bugs as just bugs
16:41:43 <angdraug> and keep them targeted at next instead of current release series, so that we don't have to move them every release
16:42:00 <dpyzhov_> good point about 'next' release
16:42:08 <kozhukalov> +1
16:42:41 <dpyzhov_> I'm tired of moving all the 'volume manager refactoring' bugs each release
16:42:47 <angdraug> )
16:42:57 <kozhukalov> ok, moving on
16:43:09 <kozhukalov> #topic Open discussion
16:43:10 <dpyzhov_> thank you, it was productive discussion
16:43:36 <kozhukalov> how about activities from US and Polish teams?
16:43:49 <kozhukalov> is there anything interesting?
16:44:32 <kozhukalov> angdraug: is there?
16:44:41 <mkwiek> I hit a little bump with nailgun tests, but need to put more work to diagnose it and show eventual solution, so maybe I will have sth for agenda next week
16:45:13 <kozhukalov> mkwiek: great
16:45:34 <kozhukalov> looking forward
16:45:51 <angdraug> nothing interesting on the us side, rmoe is working on bugs, and xarses on openrc related cleanup
16:45:57 <kozhukalov> seriously guys, we need to share info between locations
16:46:49 <kozhukalov> dpyzhov_ angdraug what about merging this one https://review.openstack.org/#/c/155698/?
16:47:18 <kozhukalov> looks like enough +1s
16:48:03 <kozhukalov> this one is also have two +1 https://review.openstack.org/#/c/165537/
16:49:06 <dpyzhov_> kozhukalov: merged. both of them
16:49:09 <kozhukalov> one more time, if you guys have anything to share: status, issue, question, please don't hesitate to add this into agenda
16:49:30 <kozhukalov> it is extremely simple https://etherpad.openstack.org/p/fuel-weekly-meeting-agenda
16:49:44 <kozhukalov> dpyzhov_: great, thanx
16:50:32 <kozhukalov> ok, looks like no one has anything more to discuss
16:50:40 <kozhukalov> ending then?
16:50:56 <kozhukalov> thanks everyone for attending
16:51:01 <IvanKliuk> thanks!
16:51:07 <kozhukalov> #endmeeting