15:00:08 <therve> #startmeeting Heat
15:00:12 <openstack> Meeting started Wed Apr 13 15:00:08 2016 UTC and is due to finish in 60 minutes.  The chair is therve. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:14 <therve> #topic Roll call
15:00:16 <openstack> The meeting name has been set to 'heat'
15:00:18 <jdob> o/
15:00:23 <ochuprykov> hi
15:00:30 <jasond> o/
15:01:04 <elynn> o/
15:01:24 <zaneb> wow, meeting-4, this is really the ghetto
15:01:27 <therve> shardy, You found your way? :)
15:01:33 <shardy> o/
15:01:46 <therve> zaneb, 15UTC is busy time! I had to kick someone to get it
15:01:53 <zaneb> lol
15:01:59 <jdob> oh oh, rumble at Summit for the less ghetto rooms
15:02:25 <therve> #topic Adding itmes to agenda
15:02:29 <therve> #link https://wiki.openstack.org/wiki/Meetings/HeatAgenda#Agenda_.282016-04-13_1500_UTC.29
15:02:48 <therve> #topic Summit sessions
15:02:59 <therve> #link https://etherpad.openstack.org/p/newton-heat-sessions
15:03:10 <zaneb> pleased to see there are now some on the list :)
15:03:12 <therve> Alright!
15:03:28 <therve> So the good news is that we have 13 sessions for 12 slots
15:03:50 <therve> We only need to bump one, or to add it to the meetup topics
15:03:53 <zaneb> they're not all the right type of sessions though
15:03:57 <therve> A bunch don't have drivers though
15:04:13 <zaneb> 2 fishbowl for 4 slots
15:04:29 <zaneb> 11 working for 8 slots
15:04:43 <therve> There 3 TBD ?
15:04:52 <therve> So now we know where they'll get :)
15:05:40 <zaneb> didn't notice that
15:05:42 <shardy> FWIW re convergence switchover, I did some testing and it appears to be mutually exclusive to performance improvements
15:05:54 <zaneb> changed them to working, because they are all working-style proposals
15:05:58 <shardy> we can discuss that further another time tho ;)
15:06:03 <zaneb> shardy: lol
15:06:38 <therve> zaneb, I'll move some to fishbowl anyway...
15:07:20 <zaneb> shardy: alternative formulation: performance improvements may help convergence even more than they help the legacy code :)
15:07:43 <therve> Regarding the one to boot, I think 11 may be a bit early. We can discuss it friday afternoon too
15:07:44 <prazumovsky> Ok, hello, sorry for lating
15:08:05 <shardy> zaneb: Yeah, I was really just commenting on the "why haven't we yet" aspect of the session description
15:08:25 <zaneb> that is good feeback :)
15:08:30 <zaneb> feedback*
15:10:08 <therve> I don't care that much about 8 too.
15:10:17 <therve> Comments?
15:11:15 <zaneb> I think the idea in 13 needs to die, but maybe having a fishbowl session where we publicly stab it with users present would be a Good Thing
15:11:22 <shardy> I'm not sure it warrants an entire session unless we turn it into a more general deprecation session
15:12:13 <shardy> e.g not just features, but the deprecation issues around resource plugin and template interfaces
15:12:14 <therve> zaneb, Well it's not a user who put that in :)
15:12:21 <jdob> zaneb: that was... graphic.
15:16:14 <zaneb> I'd be tempted to spike 8 and re-run the very large stacks session from last summit in a fishbowl
15:16:57 <shardy> zaneb: we could, does that overlap with performance, or can we justify a separate session?
15:17:24 <therve> zaneb, We need to remove another one if you add one for large stack though
15:17:35 <jdob> shardy: are performance concerns cropping up for anyone besides tripleo?
15:17:51 <zaneb> I think a fishbowl session to find out from other projects (Sahara, TripleO) what the problems still are, and then a working session to decide what to do about that and other performance issues
15:17:58 <jdob> because I read the large stacks one as pretty much saying "what tripleo is doing"
15:18:09 <shardy> jdob: Yeah, sahara was the main one last summit IIRC
15:18:17 <jdob> interesting
15:18:27 <zaneb> Sahara uses *very* big stacks
15:18:53 <zaneb> not super-complicated like TripleO AIUI, but very very big
15:19:02 <jdob> big meaning lots of resources or lots of nesting?
15:19:07 <shardy> zaneb: Yeah, some of those issues will improve if we generally improve performance, but there's some special concerns around throwing giant stacks at relatively resource constrained heat installations
15:19:09 <zaneb> lots of resources
15:19:09 <jdob> (nevermind, we can talk about that outside of this)
15:19:34 <zaneb> mainly, lots of nova servers to run Hadoop on
15:19:52 <shardy> the interesting thing is, AFAICT from testing breaking things up via RPC (as we did for nested stacks, and are now doing much more for convergence) is the worst possible thing you can do in that case
15:20:11 <shardy> I'm experimenting with backing out the StackResource RPC patch (too early to share results)
15:20:16 <zaneb> yep, which is kind of expected
15:20:58 <therve> Yeah you trade off some horizontal scaling against pure time
15:21:15 <zaneb> scalability and efficiency on a single node are often at odds
15:21:22 <zaneb> (see also: Hadoop)
15:21:55 <shardy> Sure, I guess it's a good discussion then, e.g to what extent we can support both use-cases
15:22:22 <shardy> then there is the fact that beyond scaling a few nodes, you run into DB scale limitations (again made worse by convergence)
15:22:38 <shardy> so it's not like we can just suggest everyone runs 20 node environments
15:23:29 <zaneb> really? typical nova deployment has a lot more nodes than that hitting the DB
15:23:55 <zaneb> seems like if we are running into limitations then there may be more efficient ways to use the db than what we are doing
15:24:02 <therve> Anyway, drifting a bit
15:24:08 <zaneb> yes, sorry
15:24:10 <therve> zaneb, So you want one session on large stacks?
15:24:23 <therve> Then we can remove 8 and 11
15:24:36 <therve> I'll see how to arrange sessions
15:24:38 <shardy> therve: Yeah, sorry, bit of a tangent there
15:25:50 <zaneb> so 5, 7, 13 & v. large stacks get fishbowls
15:25:55 <therve> The schedule is a bit strange with 3 working session, then the 4 bowls and the 5 remaining working
15:26:02 <zaneb> everything else except 8 & 11 get working sessions?
15:26:14 <therve> Sounds good
15:26:53 <zaneb> +1
15:27:17 <zaneb> not entirely sure about 13, but we can go with it
15:28:34 <therve> Yeah it's a discussion worth having I believe
15:30:02 <therve> Urg etherpad
15:30:08 <zaneb> ok, what just happened to the etherpad
15:32:01 <therve> So I reordered the session a bit
15:32:07 <therve> Should be okay now :)
15:33:39 <therve> Alright moving on
15:34:08 <therve> #topic Open discussion
15:37:21 <prazumovsky> Want to note that me and ochuprykov are interested in duscussing validation improvements and will be happy to get involving in it implementation
15:37:21 <prazumovsky> *involved
15:37:37 <ricolin> nice
15:38:08 <zaneb> jasond: was that you asking about abandon/adopt bugs?
15:38:08 <jasond> zaneb: yes.  i wasn't aware of those
15:38:08 <jasond> zaneb: i will look into that
15:38:08 <zaneb> jasond: see my comment at the bottom of the etherpad as well
15:38:38 <jasond> zaneb: wow, will look into that too.  thanks