21:01:07 <flaper87> #startmeeting Zaqar
21:01:07 <openstack> Meeting started Mon Jul  6 21:01:07 2015 UTC and is due to finish in 60 minutes.  The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:08 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:01:10 <openstack> The meeting name has been set to 'zaqar'
21:01:24 <flaper87> vkmc: ryansb kragniz o/
21:01:30 <vkmc> o&
21:01:31 <ryansb> \l
21:01:32 <flaper87> anyone around ?
21:01:34 <ryansb> \o
21:01:34 <flaper87> awesome
21:01:36 <flaper87> :D
21:01:42 <vkmc> ops, that's a weird arm
21:01:44 <vkmc> o/
21:01:45 <vkmc> there
21:01:49 <flaper87> :D
21:02:01 <flaper87> #topic Agenda
21:02:03 <flaper87> #link https://wiki.openstack.org/wiki/Meetings/Zaqar#Agenda
21:02:09 <flaper87> that's our agenda for today
21:02:15 <flaper87> Not much to discuss, really
21:02:21 <flaper87> #topic Relevant reviews
21:02:29 <flaper87> #link https://review.openstack.org/#/q/status:open+project:openstack/zaqar+branch:master+topic:pre-signed,n,z
21:02:45 <flaper87> Those patches implement the pre-signed URL feature we need for L-2
21:02:53 <flaper87> I'm having some fights with py34 and py27
21:02:59 <flaper87> fun thing is that those tests pass locally
21:03:00 <ryansb> oh, nice
21:03:12 <flaper87> anyway, the feature should be complete with those patches
21:03:24 <flaper87> which will allow for the client side code to be written
21:03:39 <flaper87> It'd be really amazing if we could get as many comments as possible there asap
21:03:44 <flaper87> therve already did some reviews
21:03:59 <flaper87> Any comments/questions?
21:04:15 * ryansb hasn't reviewed those
21:04:24 <ryansb> no q's yet.
21:04:40 <flaper87> ryansb: coolio, drop some on those reviews
21:04:41 <flaper87> :D
21:04:43 <flaper87> moving on
21:04:45 <flaper87> #link https://review.openstack.org/#/q/status:open+project:openstack/zaqar+branch:master+topic:bp/cross-transport-api-spec-messages,n,z
21:04:51 <flaper87> That's the cross-transport fitler
21:05:04 <flaper87> WE gotta get those merged
21:05:11 <flaper87> I +2'd the message impl today
21:05:20 <flaper87> not sure if flwang already got to that
21:05:21 <flaper87> btw, flwang you around?
21:05:34 <flaper87> nope
21:05:36 <vkmc> re presigned, I have been checking those today but I want to reread them, will submit reviews soon
21:05:49 <flaper87> vkmc: sounds awesome, thanks for taking the time
21:05:55 <flaper87> ask as many questions as possible
21:05:56 <vkmc> re cross transport, I have to work a bit more on claims, there are odd things going on there
21:06:03 <vkmc> sure, thanks flaper87 for working on it
21:06:10 <flaper87> we want to make that pre-signed feature stable enough to not shoot ourselves on our feet
21:06:30 <flaper87> vkmc: roger, I think messages is already a good step forward
21:06:31 <vkmc> yeah, that is a good idea
21:06:32 <vkmc> haha
21:06:37 <vkmc> cool!
21:06:38 <flaper87> hahaha
21:06:53 <flaper87> It'd be coold if we could start working on the push notifications feature
21:06:54 <flaper87> :D
21:06:57 <flaper87> anyway
21:06:58 <vkmc> I remember seeing therve asking for notifications with websocket
21:07:02 <vkmc> certainly our next step after claims
21:07:09 <vkmc> lol
21:07:17 <flaper87> yeah
21:07:29 <flaper87> we had a discussion the other day that I have yet to write into a spec
21:07:39 <vkmc> I'll check the backlog
21:07:40 <flaper87> I'll try to find the time this week but I don't promisse much
21:07:52 <flaper87> vkmc: that was in os-zaqar some time last week
21:08:04 <vkmc> flaper87, cool, thx
21:08:12 <flaper87> ok, moving on
21:08:14 <flaper87> #link https://review.openstack.org/#/c/196674/
21:08:18 <flaper87> That's zaqar client
21:08:26 <flaper87> diga has been putting lots of efforts on the CLI
21:08:34 <flaper87> lets give that one some priorities
21:08:41 <vkmc> yeah \o/
21:08:49 <flaper87> erm, I mean, lets bump its priority to something higher
21:08:51 <flaper87> :D
21:09:10 <flaper87> #link https://review.openstack.org/#/q/owner:%22Doraly+Navarro%22+status:open,n,z
21:09:14 <flaper87> last but not least
21:09:18 <flaper87> Those are dynarros reviews
21:09:31 <flaper87> there's still some work there but there are some that are ready to be merged
21:09:46 <flaper87> so, please, take a look there too. Those patches make the library better
21:10:27 <flaper87> ok, that's all I have
21:10:32 <flaper87> since we have ryansb here
21:10:34 <vkmc> cool, thanks dynarro!
21:10:39 <flaper87> #topic News from Heat-land
21:10:45 <flaper87> ryansb: ^ do you have something for us?
21:11:30 <flaper87> apparently we had him
21:11:31 <flaper87> :P
21:11:40 <flaper87> I'll make sure I ask later
21:11:48 <flaper87> #topic Updates from other teams
21:11:48 <ryansb> yeah, I do
21:11:52 <flaper87> #undo
21:11:53 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x946cdd0>
21:11:56 * flaper87 loves the undo
21:12:00 <flaper87> ryansb: goooooo
21:12:00 <ryansb> lol
21:12:27 <vkmc> hahaha
21:12:27 <ryansb> ok, so we merged zaqar as a software config transport
21:12:34 <ryansb> which is great
21:12:52 <ryansb> we're still working on an event format to send user events
21:13:03 <ryansb> and that's it from heat
21:13:18 <ryansb> unless therve is around/has anything
21:13:41 <flaper87> mmh, I don't think he is
21:13:45 <flaper87> those are great news
21:13:54 <flaper87> How are you guys testing things?
21:14:10 <ryansb> therve is mocking zaqarclient, I think
21:14:26 <ryansb> we just kinda trust that zaqar will work
21:14:32 <flaper87> oooook, any chance we can help with a functional gfate? is that something you guys want ?
21:15:21 <ryansb> not sure, I don't know if we're looking at that atm
21:15:28 <flaper87> ok, do let us know
21:15:31 <ryansb> kk
21:15:41 <flaper87> That also reminds me. We gotta fix the functional gate for zaqarclient
21:15:45 <flaper87> oh gosh, so many things
21:15:46 <flaper87> :D
21:15:53 <flaper87> ryansb: thanks for the updates
21:15:57 <ryansb> and (non-heat related) the swift transport is almost passing zaqar's functional tests
21:16:03 <ryansb> so that's cool
21:16:15 <flaper87> ryansb: oh that's coool indeed
21:16:17 <flaper87> :D :D :D
21:16:30 <flwang> o/
21:16:33 <vkmc> oh yes, I was going to bring that up
21:16:34 <flwang> sorry for the late
21:16:40 <vkmc> we need to review that driver
21:16:41 <flwang> blame the morning traffic
21:16:42 <vkmc> is looking really good
21:16:52 <flaper87> that reminds me
21:16:56 <flaper87> (moving on)
21:17:02 <flaper87> #topic News from other projects
21:17:05 <flaper87> #link https://review.openstack.org/#/c/196755/
21:17:11 <flaper87> ^ that's also really cool
21:17:17 <vkmc> hi flwang!
21:17:27 <flwang> vkmc: hi guys
21:17:32 <vkmc> flaper87, sweeeee
21:17:33 <vkmc> t
21:17:39 <flaper87> Swift's notification middleware that uses zaqar
21:17:46 <flaper87> I mean, it's not required
21:17:54 <flaper87> but Zaqar is a totally valid option
21:17:54 <ryansb> neat
21:18:00 <flaper87> w0000h0000
21:18:36 <flaper87> ok, moving on
21:18:43 <flaper87> #topic email publisher
21:18:43 <flwang> i'm looking at https://review.openstack.org/#/c/196755/  very interesting...
21:18:56 <flaper87> flwang: news on $topic ?
21:19:14 <flwang> flaper87:i'm going to submit a patch this week
21:19:22 * flaper87 is making up these topics at runtime. Bare with me, everything is coming to mind as we go
21:19:25 <flwang> still coding
21:19:36 <flaper87> flwang: cooolio, as long as you say you're coding, I'm happy
21:19:44 <flwang> flaper87: hah
21:19:50 <flaper87> it'd be really nice to get it up this week
21:19:59 <ryansb> JIT_agenda++
21:20:12 <flaper87> ryansb++
21:20:29 <flaper87> ok, moving on
21:20:36 <flaper87> #topic Upcoming Summit
21:20:53 <flaper87> yeah yeah, it's probably too early bbut I'd like to have more time to prepare things
21:21:05 <ryansb> summit already?
21:21:11 * ryansb checks calendar
21:21:15 <flaper87> The first thing I'd like to highlight is that there are folks (vkmc, dynarro, cpallares and shaifali) looking forward to present
21:21:29 <flaper87> The CFP is open and it'll close on July 15th
21:21:38 <vkmc> \o/
21:21:43 <vkmc> didn't know about dynarro, cool
21:21:53 <flaper87> so, get your abstracts in there and if you want help to review them and provide feedback on topics etc, please let us know
21:22:00 <flaper87> vkmc: I saw shaifali pinged her earlier today
21:22:08 <ryansb> indeed, I'm happy to review too
21:22:11 <flaper87> I'm not really sure if she's actually going to
21:22:24 <vkmc> flaper87, oh she will
21:22:34 <vkmc> sweet!
21:22:37 <flaper87> vkmc: I know, right? ;)
21:22:45 <vkmc> it would be nice to have a prez of Zaqar + Heat
21:23:02 <flaper87> to be clear, I'm not going to present this time.
21:23:10 <vkmc> flaper87, damn
21:23:19 <vkmc> it was targeted to you directly
21:23:20 <vkmc> hahahaha
21:23:27 <flaper87> vkmc: I think zaqar+heat would be amazing
21:23:45 <flwang> flaper87: what's the current status of zaqar + heat?
21:23:55 <flwang> any link/spec about that?
21:24:02 <flaper87> flwang: ryansb just updated us, mind reading the summary later?
21:24:05 <ryansb> flwang: we just merged zaqar as a transport for softwareconfig
21:24:15 <ryansb> see meeting notes for details
21:24:22 <flwang> flaper87: ryansb: cool, thanks
21:24:25 <flaper87> np
21:24:25 <flwang> i will read the summary
21:24:42 <flaper87> ok, back to the topic
21:24:57 <flaper87> Please, send many abstracts, I think Zaqar+Heat or Zaqar+Swift would be great to have
21:25:10 <flaper87> It'd be awesome to show people real-world use cases
21:25:18 <ryansb> zaqar+swift as in "using zaqar as a swift notifier"
21:25:21 <flaper87> rather than talking a lot about the internals
21:25:24 <flaper87> ryansb: yes
21:25:30 <ryansb> gotcha
21:25:31 <flaper87> (related to the patch I just shared)
21:25:53 <flaper87> mmh, actually, swift+zaqar+Swift
21:25:56 <flaper87> that'd be sweet too
21:25:59 <ryansb> lol
21:26:06 <ryansb> stacks on stacks on stacks
21:26:07 <flaper87> :D
21:26:16 * flaper87 does the containers dance
21:26:45 <flaper87> ok, moving on
21:26:49 <flaper87> #topic Policy support
21:26:52 <flaper87> #link https://review.openstack.org/#/c/179673/
21:27:00 <flaper87> kragniz seems to be very busy lately
21:27:12 <flaper87> I'm not sure if he'll have the time to tackle that one
21:27:28 <flwang> flaper87: i can talk with him
21:27:28 <flaper87> I'll ping him again later but, in case he doesn't have time, would any of you like to take it?
21:27:39 <flaper87> flwang: +1
21:27:47 <flaper87> flwang: that'd be great
21:27:49 <flwang> i can take it back if he don't have bandwidthg
21:27:50 <flaper87> to get a clear view
21:27:57 <flaper87> flwang: do you have the b/w ?
21:27:58 <flaper87> :)
21:28:15 <flwang> flaper87: good question
21:28:22 <flaper87> flwang: any spare minion ?
21:28:32 <flaper87> I have one in mind, diga
21:28:42 <flwang> you can take it either, boss
21:28:50 <flwang> then i can focus on the email notification
21:28:52 <flaper87> but that's the only name that comes to mind
21:29:02 <flaper87> flwang: yeah, it's better to get that email publisher done
21:29:20 <flaper87> flwang: you focus on that one, I'll see what can be done with the policy spec
21:29:32 <flwang> flaper87: ok, coojl
21:29:38 <flaper87> We have other contributors that would be happy to help
21:30:21 <flaper87> ok, now I ran out of topics
21:30:26 <flaper87> #topic Open Discussion
21:30:30 <flaper87> o/
21:30:48 <flaper87> #info There won't be a meeting next week (and probably the one after next week)
21:31:04 <vkmc> whaaa?
21:31:17 <flaper87> vkmc: holidays :)
21:31:21 <flaper87> you're happy to join us
21:31:24 <vkmc> ohh
21:31:26 <flaper87> I know you'll like that
21:31:31 <vkmc> yeah I need holidays
21:31:35 <flaper87> s/happy/welcome/
21:31:38 <flaper87> but also happy
21:31:41 <vkmc> but I have some things to finish
21:31:50 <vkmc> unfortunately >.>
21:32:10 <ryansb> nice. We'll see you when you get back then.
21:32:18 <vkmc> hey guys, its flaper87's birthday
21:32:18 <vkmc> :D
21:32:23 <flwang> flaper87: i would like to know your current opinion about 'queue' of zaqar
21:32:25 <flaper87> vkmc: me too but remember, procrastination makes you better https://www.youtube.com/watch?v=Ux5dRX_9bj0
21:32:29 <ryansb> \o/ happy birthday!
21:32:37 <flaper87> ryansb: oh, thank you :D
21:32:42 <vkmc> flwang, that's a good question
21:32:47 <vkmc> we should revisit that
21:32:57 <flaper87> flwang: you mean the spec to s/queue/topic/ ?
21:33:05 <flwang> flaper87: sorry for the tough question as your birthday  gift
21:33:22 <flwang> flaper87: yes, i'm not a fan to move to topic, TBH
21:33:26 <flaper87> flwang: haha, no worries, I get younger every year ;)
21:33:45 <flaper87> flwang: my current opinion is that we did a big thing by making queue's a non-required resource
21:33:50 <flaper87> it's now a background thing
21:33:57 <flaper87> and messages are all we care about
21:34:06 <flaper87> the rename can be posponed (or never done)
21:34:25 <flaper87> My current thinking is that we should focus on improving what we have and the less we talk about queues the better
21:34:35 <flaper87> That is, Zaqar is a messaging service and that's what it provides
21:34:36 <flwang> flaper87: i asked because i just noticed swift notification middleware is using the 'queue'
21:35:00 <flwang> that said, the integrations are happening,  is depending on 'queue'
21:35:10 <flaper87> Yeah, I need to re-read the patch. I commented on an earlier PS saying that queue's are lazy
21:35:37 <flaper87> I think they shouldn't rely on queue creation
21:36:03 <flaper87> that said, I think we may have to stick with queue as a resource name
21:36:08 * flaper87 shrugs
21:36:21 <flwang> flaper87: that's my point
21:36:57 <flwang> flaper87: i prefer to keep 'queue' as supported, though it maybe not recommend by zaqar's ptl
21:37:03 <flaper87> I guess the TL;DR is that I don't care enough for now. Lets work on making people use API >=1.1
21:37:23 <flaper87> or even >=2
21:37:33 <vkmc> if we could rename queue -> topic
21:37:41 <vkmc> without changing the api too much
21:37:50 <ryansb> I don't see how renaming it matters all that much
21:37:54 <flwang> vkmc: then, what's the benefit?
21:37:59 <flaper87> ryansb: exactly
21:38:09 <vkmc> the thing is that users complain that the queues don't behave like a queue
21:38:20 <flaper87> At some point it seemed quite important to me but the more I thought about it the less I cared
21:38:21 <vkmc> and that is right
21:38:37 <flaper87> We worked on changing the API itself instead of just the name of the resource
21:38:39 <vkmc> I don't care either, but I know that this has generated some noise in the past
21:38:42 <flaper87> and I believe the result was better
21:39:09 <flaper87> vkmc: I think it was a set of things put together mission+queues+name+API
21:39:19 <vkmc> flaper87, yeah of course
21:39:21 <flaper87> now we're just left with that silly name
21:39:29 <flaper87> which we can just tell people to ignore
21:39:31 <flaper87> :P
21:39:36 <vkmc> but the fact that our queues are not FIFO was the biggest complain
21:39:48 <vkmc> hehe yes
21:39:52 <ryansb> yeah, that's a fair complaint I think.
21:40:25 <flaper87> well, our queues were FIFO
21:40:33 <vkmc> not really
21:40:33 <flaper87> we then moved them down the chain
21:40:38 <flaper87> yes they were
21:40:40 <vkmc> we had the get_by_id endpoint
21:40:43 <flaper87> that was a requirement
21:40:51 <flaper87> sure, which we didn't have in 1.1
21:40:55 <vkmc> oh yes
21:41:03 <flaper87> anyway
21:41:17 <flaper87> queue's are behind messages now
21:41:23 <flaper87> FIFO is optional
21:41:29 <flaper87> and it's a storage feature
21:41:34 <flaper87> rather than an API guarantee
21:42:34 <flaper87> One of Zaqar's goals has always been to answer the need of: "Gimme an url to post messages"
21:42:49 <flaper87> I think we have that now without extra requirements
21:42:56 <flwang> flaper87: and i have another topic, redis
21:42:58 <flaper87> we still have queues in the URL but oh well
21:43:04 <flaper87> flwang: sup with that?
21:43:21 <flwang> redis seems like a second citizen compared with mongo
21:43:34 <flwang> thought we don't admit that
21:43:49 <flaper87> oh we do admit that
21:43:50 <flaper87> :P
21:43:58 <flaper87> I want to change that story
21:44:01 <flaper87> FWIW
21:44:06 <flwang> that said, should we start to fill the gap
21:44:23 <flaper87> flwang: I guess you're referring to the lack of support for subscriptions
21:44:25 <flaper87> right?
21:44:46 <flwang> flaper87: not really
21:44:57 <flaper87> flwang: what gap are you referring to?
21:44:58 <flwang> actually, there are many TODO in redis driver
21:45:09 <flaper87> flwang: but is there a feature gap?
21:45:13 <flaper87> other than those TODOs ?
21:45:30 <ryansb> flwang: to be fair, there are TODOs in other places too
21:45:31 <flaper87> ryansb: are those TODOs something you'd like to work on? I know you love redis
21:45:32 <flaper87> :D
21:45:42 <flwang> pool, flavor, catalogue
21:45:44 <flwang> brb
21:45:51 <flaper87> flwang: right, but that's management
21:45:52 <ryansb> heh, I do like redis, but I haven't actually looked at the driver.
21:46:04 <flaper87> we split management and message planes on purpose
21:46:20 <flaper87> I don't think the lack of support for creating pool/flavor/catalogue resources means we don't care about redis
21:46:39 <flaper87> actually, that allows us to focus more on the data plane, which is where we want redis to rock
21:46:50 <vkmc> still... you would need mongo for management
21:46:57 <flaper87> vkmc: or sqla :)
21:47:02 <ryansb> I can add it to my list, but I'd rather do the swift driver first
21:47:05 <vkmc> or sqla
21:47:06 <vkmc> yes
21:47:09 <flaper87> ryansb: +1
21:47:26 <flaper87> we need https://bugs.launchpad.net/zaqar/+bug/1471193 fixed
21:47:26 <openstack> Launchpad bug 1471193 in zaqar "Don't assume the queue controller is in the same storage" [High,Confirmed] - Assigned to Digambar (digambarpatil15)
21:47:28 <flaper87> diga took it
21:47:40 <ryansb> vkmc: fwiw, I'm building swift as data-only, and rely on having sqla available
21:47:41 <flaper87> That bug needs to be fixed to make the redis+sqla combination work
21:47:44 <ryansb> (or mongodb)
21:47:48 <flaper87> ryansb: +1
21:47:52 <flaper87> I think that's totally fair
21:47:55 <vkmc> ryansb, yeah :)
21:47:57 <vkmc> it is
21:48:07 <vkmc> same thing for websocket, in the transport side
21:48:15 <vkmc> if it doesn't make sense to have management its just finde
21:48:17 <vkmc> fine+
21:48:25 <ryansb> yeah, I dunno how I'm going to get swift to do subscriptions yet
21:48:32 <ryansb> claims work fine.
21:48:44 <ryansb> cross that bridge when I get there.
21:49:32 <vkmc> ryansb++
21:49:45 <flaper87> ryansb: if it becomes really hard to get subscriptions there, I'd say you shouldn't even bother
21:49:47 <flaper87> really
21:49:57 <flaper87> swift as a data plane storage is a great thing to have
21:50:04 <flaper87> for management, there are other things
21:50:41 <flaper87> anything else?
21:50:56 <vkmc> one day we should discuss further if the options we now provide for management are good
21:51:17 <flaper87> vkmc: storage options ?
21:51:17 <vkmc> good as in ops friendly
21:51:22 <vkmc> flaper87, yeah
21:51:29 <flaper87> oh, sqla is very ops friendly
21:51:31 * flaper87 ducks
21:51:33 <flaper87> :P
21:51:34 <vkmc> haha
21:51:43 <flaper87> jokes apart, I agree
21:51:52 <vkmc> cool
21:51:53 <vkmc> :)
21:51:58 <flaper87> for now I guess I'm happy because I know ops have experience with both mongo and sqla
21:52:03 <flaper87> more with the later, likely
21:52:30 <vkmc> yeah, if they want to deploy zaqar they don't need any odd dependency
21:53:45 <flaper87> ok, if there isn't anything else, I'll end this meeting
21:53:48 <flaper87> we used the whole slot
21:53:52 <flaper87> I did not expect that
21:53:53 <flaper87> :P
21:54:24 <ryansb> work expands to fill the time available
21:54:35 <ryansb> I think that's a law or something
21:54:38 <ryansb> ;)
21:54:56 <vkmc> hahaha
21:55:17 <vkmc> it was a very productive meeting indeed
21:55:23 <flaper87> ryansb's law
21:55:34 <flaper87> coolio, thanks all for attending and the amazing feedback
21:55:41 <flaper87> catch you all in os-zaqar!
21:55:44 <ryansb> thanks flaper87, seeya
21:55:44 <flaper87> rock on
21:55:47 <flaper87> #endmeeting