00:02:48 <zaneb> #startmeeting heat
00:02:49 <openstack> Meeting started Thu Apr 17 00:02:48 2014 UTC and is due to finish in 60 minutes.  The chair is zaneb. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:02:50 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:02:52 <openstack> The meeting name has been set to 'heat'
00:03:11 <zaneb> #chair stevebaker
00:03:12 <openstack> Current chairs: stevebaker zaneb
00:03:23 <stevebaker> \o
00:03:29 <bgorski> \o
00:03:33 <spzala> Hi
00:04:30 <zaneb> stevebaker: chair is yours if you want it. or not, I don't care ;)
00:04:36 <stevebaker> oh, ok
00:04:49 <stevebaker> #topic Adding items to the agenda
00:05:00 <stevebaker> I've got a couple of firefighting items
00:05:06 <wirehead_> o/
00:05:43 <stevebaker> anything else to add?
00:05:48 <stevebaker> #link https://wiki.openstack.org/wiki/Meetings/HeatAgenda#Agenda_.282014-04-17_0000_UTC.29
00:06:06 <spzala> seems like people are in easter holidays mood :-)
00:06:12 <stevebaker> yeah
00:06:21 <stevebaker> #topic actions from last week
00:06:23 <wirehead_> Yeah, trying to do too much in this meeting might leave people with egg on their face
00:06:27 <stevebaker> shardy to write first Heat security page wiki
00:06:40 <stevebaker> that has been done, but there is still a TODO on the release notes
00:07:09 <stevebaker> #topic https://wiki.openstack.org/wiki/ReleaseNotes/Icehouse#OpenStack_Orchestration_.28Heat.29
00:07:18 <stevebaker> I mean
00:07:20 <stevebaker> #link https://wiki.openstack.org/wiki/ReleaseNotes/Icehouse#OpenStack_Orchestration_.28Heat.29
00:07:51 <stevebaker> feel free to take a look, add any known issues, correct any features
00:08:50 <stevebaker> our icehouse release has been cut, so announcements will happen very soon
00:09:04 <zaneb> stevebaker: didn't we deprecate some resources?
00:09:22 <zaneb> RouterGateway comes to mind
00:09:59 <stevebaker> zaneb: yes, I think that is the only one
00:10:20 <stevebaker> I assume deprecated properties are not release-notes worthy
00:10:56 <stevebaker> zaneb: could you add that to the Upgrade Notes?
00:11:30 <stevebaker> #topic https://bugs.launchpad.net/heat/+bug/1306743
00:11:33 <zaneb> #action zaneb to add deprecation of RouterGateway to Release Notes
00:11:56 <lifeless> oh hai heat ;)
00:12:16 * zaneb runs
00:12:50 <lifeless> stevebaker: I see some progress in analysis
00:13:32 <stevebaker> so metadata polling currently has a high overhead because we currently need to parse the stack. There are a bunch of ways we could optimise and it would be great if we could all carve off bits of the problem
00:14:13 <zaneb> lifeless: are y'all using multiple Heat engines?
00:14:16 <lifeless> is it amenable to a cache?
00:14:19 <lifeless> zaneb: not yet
00:14:21 <asalkeld> +1 to storing the parsed stack to swift
00:15:48 <stevebaker> we've never even considered caching, memcached is another option. But I see that as more medium term
00:16:07 <stevebaker> zaneb: could you take over for a sec, I'll be biab
00:16:13 <zaneb> sure
00:16:55 <lifeless> storing the parsed stack in swift would be a cache, no?
00:17:35 <asalkeld> yip
00:18:04 <zaneb> it's not clear to me that these calls once every 30s are what's causing 300-400 DB requests per second...
00:18:47 <zaneb> having to load the whole stack is unfortunate, but it doesn't seem like it should be this... disasterous
00:18:59 <stevebaker> back (sick kid)
00:19:30 <Qiming> not sure Heat should cache the parsed stack itself rather than relying on other projects
00:20:02 <stevebaker> I believe memcached is used by other projects for caching, such as keystone
00:20:28 <stevebaker> Anyway, we're continuing to look into the issue
00:20:37 <stevebaker> #topic https://bugs.launchpad.net/heat/+bug/1297560
00:20:40 <zaneb> yep, it's been a lot of fund debugging the config option for it >:|
00:20:47 <zaneb> s/fund/fun/
00:20:54 <stevebaker> #link https://bugs.launchpad.net/heat/+bug/1297560
00:22:07 <stevebaker> this one has lead to the heat-slow job becoming non-voting while we figure out why wait conditions time out %50 of the time. So *please* check the reason for any heat-slow failures before +2ing any heat changes
00:22:35 <stevebaker> #topic Meeting times for the Juno cycle
00:22:38 <stevebaker> zaneb: over to you
00:22:42 <Qiming> sdague asked whether we can have cfntools installed in a cirros
00:23:06 <zaneb> stevebaker: cheers, and thanks for all your hard work on Juno
00:23:08 <stevebaker> Qiming: no we can't, but we could have a cirros test which signals with curl. I'll be looking at that
00:23:32 <asalkeld> or docker tests?
00:23:36 <asalkeld> much faster
00:24:03 <zaneb> so, is this meeting time working for us, and do we want to keep it for Juno?
00:24:03 <asalkeld> (stacks with nova docker driver-  when it's back in)
00:24:22 <stevebaker> This time slot has had mixed success, maybe a slot which works for east-coast, europe and maybe asia would be better
00:24:31 <asalkeld> +1 from me
00:24:53 <stevebaker> or at least russia
00:25:00 <zaneb> I going to go out on a limb and guess that Qiming is based in asia
00:25:10 <zaneb> but I don't see anyone else from there present
00:25:39 <zaneb> so it seems like we've lost Europe folks for not much benefit
00:25:39 <stevebaker> yeah, I was hoping for more of Liang, nanjj
00:26:14 <zaneb> and west coast US folks seem to struggle with this time too
00:26:38 <zaneb> wirehead_: I assume this is in the middle of commute time for most?
00:27:09 <Qiming> stevebaker: it's 8am in the morning here, for Liang and Jun Jie as well
00:27:15 <stevebaker> #link http://xkcd.com/1335/
00:27:23 <wirehead_> zaneb: it's dangerously close to beer-o-clock.
00:27:32 <stevebaker> Qiming: I thought it was 9. Do you have DST?
00:27:34 <wirehead_> commute hour for the 9to5er
00:27:45 <zaneb> figures
00:27:59 <wirehead_> I mean, I catch the 6pm train, so
00:28:03 <Qiming> stevebaker: no, we don't.
00:28:06 <zaneb> stevebaker: no DST in China afaik
00:28:56 <zaneb> ok, so if we're agreed the current time is not working that well
00:29:03 <stevebaker> zaneb: oh well, do you want to experience the pure joy of finding a new time?
00:29:20 <zaneb> #action zaneb to propose options for alternate meeting times on Mailing List
00:29:21 <wirehead_> Heavy is the head that wears the crown
00:29:35 <wirehead_> That's one of the mostly-useless pieces of advice a prior manager gave me about management.
00:29:45 <zaneb> #topic Oslo Liason
00:29:59 <zaneb> #link https://wiki.openstack.org/wiki/Oslo/ProjectLiaisons
00:30:05 <zaneb> this is a new thing ^
00:30:11 <zaneb> therve has volunteered
00:30:22 <zaneb> so y'all are off the hook
00:30:41 <zaneb> #topic Design Summit sessions
00:30:47 <zaneb> oh ho!
00:30:59 <stevebaker> I wondered if the full oslo sync might help https://bugs.launchpad.net/heat/+bug/1306743, but I guess not
00:31:06 <wirehead_> radix added updates to the one you sent back, zaneb
00:31:25 <zaneb> wirehead_: ok, thanks, I haven't seen that yet
00:31:40 <zaneb> #link http://summit.openstack.org/
00:31:58 <zaneb> the really interesting link is http://summit.openstack.org/cfp/topic/8
00:32:08 <zaneb> but I don't think anyone else can see that
00:32:13 <wirehead_> yah
00:32:25 <stevebaker> :(
00:32:33 <zaneb> so, every project has lost slots this time
00:32:44 <zaneb> the good news is way have 8
00:32:53 <Qiming> aha, simple page says 'Forbidden'
00:32:55 <zaneb> same as Portland, one less than Hong Kong
00:33:30 <zaneb> so, remember if you click on 'Topic' it will sort by topic
00:33:46 <stevebaker> zaneb: when is the deadline for proposing a session?
00:33:50 <zaneb> we have 13 sessions proposed
00:33:56 <zaneb> deadline is this Sunday
00:33:58 <wirehead_> 13 sessions enter... 8 leave
00:34:00 <stevebaker> I'll be adding one
00:34:35 <asalkeld> http://summit.openstack.org/cfp/details/85 and http://summit.openstack.org/cfp/details/11 could maybe be merged
00:34:37 <zaneb> so, at least one genuinely important session WILL be cut
00:34:38 <bgorski> What do you think about session for template catalog?
00:34:53 <stevebaker> bgorski: that is a glance thing, surely ;)
00:34:55 <zaneb> so make sure you have lots of information in yours ;)
00:35:30 <bgorski> stevebaker, why glance ? :)
00:36:04 <zaneb> asalkeld: it's not clear to me that sending notifications and receiving notifications are closely related
00:36:19 <spzala> stevebaker: thanks for your glance related comment on http://summit.openstack.org/cfp/details/68 .. will follow up on this.
00:36:31 <zaneb> one day somebody will quote that out of context
00:37:04 <stevebaker> bgorski: the template catalog will be implemented in glance, it will become the project that catalogs all the things (images, templates, solum/murano apps)
00:37:12 <zaneb> is Rabi Mishra here?
00:37:48 <zaneb> evidently not
00:37:49 <Qiming> stevebaker: IIRC, tuskar intends to store templates on Swift
00:38:09 <zaneb> I'm just going to reject http://summit.openstack.org/cfp/details/308 and merge it with http://summit.openstack.org/cfp/details/86 if no-one objects
00:38:13 <stevebaker> Qiming: maybe they would store them in glance if they could
00:38:36 <stevebaker> zaneb: they look quite similar
00:39:08 <stevebaker> fyi, my session will be about splitting out resources into different source repositories
00:40:45 <asalkeld> yikes stevebaker
00:40:52 <stevebaker> zaneb: should these be merged? http://summit.openstack.org/cfp/details/11 http://summit.openstack.org/cfp/details/85
00:40:52 <zaneb> stevebaker: rejected :D
00:41:00 <stevebaker> lol
00:41:38 <asalkeld> I am more inclined to push the resources back into {neutron/nova/cinder}
00:41:47 <zaneb> stevebaker: asalkeld just asked that
00:41:51 <zaneb> and I said
00:41:59 <zaneb> <zaneb> asalkeld: it's not clear to me that sending notifications and receiving notifications are closely related
00:42:41 <asalkeld> zaneb it makes sense to talk about them together
00:42:51 <asalkeld> esp. if we are pushed for time
00:42:59 <zaneb> asalkeld: can you leave comments to that effect?
00:43:01 <stevebaker> maybe if the topics are small enough not to need a whole session
00:43:12 <asalkeld> not suggesting it is the same functional topic
00:43:24 <asalkeld> k
00:43:30 <zaneb> the other thing I want to note for y'all is that the design summit is not the *beginning* of the design process
00:43:45 <zaneb> so you should be proposing stuff on the mailing list...
00:43:56 <zaneb> socialising the various implementation options
00:44:13 <zaneb> in preparation for nailing down the details at summit
00:44:30 <stevebaker> zaneb: as an aside, all the cool kids seem to be moving to git repos and gerrit to review blueprint specs. Something to consider for heat
00:44:32 <zaneb> and you should also be starting implementation _now_ that Icehouse is released
00:44:53 <asalkeld> mabe one for the (#ifdef's in the templates?)
00:44:57 <asalkeld> ;)
00:46:06 <stevebaker> zaneb: biab again. you can take over the remainder of the meeting
00:46:24 <zaneb> anyway, if you haven't started the leadup work we will be less inclined to approve your session
00:46:47 <zaneb> there's motivation for everyone ;)
00:47:24 <zaneb> any thoughts about the ops session that shardy proposed?
00:47:30 <wirehead_> The beatings will continue until morale improves?
00:47:35 <zaneb> do we have enough ops people interested?
00:47:49 <zaneb> wirehead_: you are our designated ops person for the day
00:48:45 <zaneb> my plan is to roll andrewplunk's heatclient session into that one too
00:49:03 <wirehead_> Isn't there a seperate operators summit?
00:49:22 <zaneb> there's an ops track at the main summit
00:49:54 <zaneb> but this was an initiative to get projects to make space for ops-developer communication at the design summit
00:50:22 <zaneb> ok, 10 minutes left so let's move on
00:50:31 <wirehead_> Doing it as one of the design talk slots feels weird, but we do need to have this discussion.
00:50:55 <zaneb> #action everyone to submit design summit session proposals by the end of the week
00:51:10 <zaneb> #topic Heat/Mistral collaboration
00:51:12 <wirehead_> It does overlap with the heatclient
00:51:43 <zaneb> are the Murano folks here?
00:52:15 <zaneb> gokrokve: o/
00:52:27 <gokrokve> Hi
00:52:35 <asalkeld> you mean mistral?
00:52:54 <asalkeld> tnurlygayanov, ..
00:52:56 <zaneb> oh, I totally misread that
00:52:59 <zaneb> thanks
00:53:14 <zaneb> gokrokve: sorry. btw, what's Renat's IRC handle?
00:53:38 <asalkeld> only me and tnurlygayanov in #mistral
00:53:54 <gokrokve> rakhmerov
00:54:05 <gokrokve> But it is deep night for him now
00:54:17 <asalkeld> maybe next meeting
00:54:27 <gokrokve> I actually thought that Heat meeting is at 1PM today
00:54:28 <asalkeld> only 5 mins left
00:54:35 <zaneb> good point, I wonder why he wanted it on the agenda of this meeting
00:54:43 <zaneb> #topic Open Discussion
00:54:49 <zaneb> any takers?
00:55:38 <mspreitz> ill bite
00:56:17 <mspreitz> I will bite
00:57:22 <zaneb> mspreitz: 3 minutes
00:57:35 <mspreitz> just wanted to post a heads up. I am writing a nova bp for simultaneous sched of server group
00:58:05 <mspreitz> would later want to discuss how to use it thru heat
00:58:27 <zaneb> mspreitz: it would look like a resource to heat?
00:58:50 <mspreitz> thats all gor now
00:59:11 <asalkeld> hopefully not just an option to nova.boot
00:59:26 <asalkeld> so we can track the sub servers
00:59:37 <asalkeld> or is that totally hidden?
00:59:46 <mspreitz> dub servers?
00:59:53 <mspreitz> sub
01:00:02 <asalkeld> well do we get a server group
01:00:08 <asalkeld> (resource)
01:00:26 <mspreitz> yes
01:00:32 <asalkeld> ok
01:00:36 <zaneb> mspreitz: this is a different blueprint to https://blueprints.launchpad.net/nova/+spec/instance-group-api-extension ?
01:01:15 <zaneb> time's up
01:01:18 <zaneb> #endmeeting