15:01:55 <flaper87> #startmeeting Marconi Team Meeting 15:01:56 <openstack> Meeting started Tue May 20 15:01:55 2014 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:58 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:02:00 <openstack> The meeting name has been set to 'marconi_team_meeting' 15:02:02 <flaper87> adrian_otto: no worries :) 15:02:10 <flaper87> #topic Roll Call 15:02:12 <kgriffs> o/ 15:02:19 <flaper87> #chair kgriffs 15:02:19 <Obulpathi> o/ 15:02:19 <balajiiyer> o/ 15:02:20 <openstack> Current chairs: flaper87 kgriffs 15:02:20 <sriram> o/ 15:02:25 <flaper87> kgriffs: go ahead :) 15:02:28 <flaper87> o/ 15:02:30 <mpanetta> <-- 15:02:33 <megan_w> heyo 15:02:36 <alcabrera> o/ 15:02:46 <AAzza> o/ 15:03:23 <flaper87> first and foremost. It was really great to finally meet you all! 15:03:28 <kgriffs> yay! 15:03:35 <mpanetta> Yes :) 15:03:37 <tjanczuk> yes! 15:03:38 <sriram> yes! :) 15:03:46 <flaper87> We need more contributors so we can keep doing the same thing every summit :D 15:03:52 <malini> o/ 15:03:54 <flaper87> but don't tell malini 15:03:56 <flaper87> oooooooooooopsssssssssssssss 15:03:58 <alcabrera> haha 15:04:01 <Obulpathi> hahaha 15:04:01 <megan_w> haha 15:04:07 <malini> Thank you znc! 15:04:12 <flaper87> damnit 15:04:20 <malini> flaper87: :-P 15:04:22 <flaper87> but you just joined 15:04:25 <flaper87> znc is not that smart 15:04:27 <flaper87> :D 15:04:30 * flaper87 thinks 15:04:45 <kgriffs> #topic conference retrospective 15:05:39 <cpallares> \o/ 15:05:43 <cpallares> o/ 15:05:45 <kgriffs> First off, I have to apologize - I haven't had a chance to write up the roadmap we discussed on Friday. I'll get that done today and send an email to the ML. 15:05:58 <tjanczuk> I have a picture if you need it 15:06:18 * ametts thinks kgriffs should run the easel chart through an OCR reader 15:06:19 <flaper87> tjanczuk: we took all those papers w/ us :P 15:06:34 <flaper87> tjanczuk: well, I took them down and made kgriffs carry them around :D 15:06:36 <flaper87> hahaha 15:06:39 <kgriffs> tjanczuk: thanks; I've got one too. I'll try to post all the pics of the board as well. 15:07:06 <kgriffs> generally speaking, how does everyone feel about the summit? 15:07:11 <megan_w> in general, there seemed to be confusion in the rest of the community about marconi under/over cloud 15:07:14 <malini> still on? 15:07:32 <megan_w> people didn't seem to know if they were suppose to use it between services 15:07:45 <tjanczuk> I think there was also a bit of confusion around whether Marconi attempts to be the message server for rest of marconi or for end users, or both 15:07:54 <tjanczuk> sorry, for rest of openstack 15:07:58 <megan_w> right 15:08:01 <sriram> yes, The people I talked to at the summit had the same question. 15:08:14 <kgriffs> that's been a point of confusion for quite a while. We need to clear that up once and for all. 15:08:14 <Obulpathi> +1 15:08:27 <tjanczuk> and the answer is...? 15:08:29 <megan_w> do we have clarity amongst ourselves yet? 15:08:29 <flaper87> I don't think we will 15:08:42 <flaper87> we need to put that in the FAQ (which I think it is) 15:08:45 <kgriffs> I think an entry on the FAQ and a mention on the home page should help. 15:08:48 <flaper87> and point people there 15:08:51 <flaper87> kgriffs: >.> 15:09:16 <flaper87> me starts his brain firewall 15:09:22 <kgriffs> I think it is mentioned in one or two answers in the FAQ 15:09:22 <kgriffs> but 15:09:38 <kgriffs> maybe we should call it out specifically as a standalone question? 15:09:45 <Obulpathi> +1 15:10:06 <flaper87> kgriffs: sounds good, I thought it was 15:10:34 <kgriffs> #note basically, we should say that is is primarily over cloud but we have requests from people who want to use it as an event aggregator and normilizer 15:11:13 <flaper87> We're targetting the over cloud but we're not planning to prevent people to use it in the under cloud 15:11:13 <kgriffs> I can take a first stab at it. Then I'll need others to help me edit it for clarity. 15:11:29 <ametts> +1 15:11:31 <megan_w> if its primarily over cloud, then we need to make sure our roadmap isn't filled with aggregator/normalizer items 15:11:31 <flaper87> kgriffs: actionize it or it ain't happen :D 15:11:52 <malini> On tht note, one of TC memebers concern was 'we are not integrated to other products' 15:11:55 <flaper87> megan_w: what aggregators ? you mean the notification? 15:12:03 <tjanczuk> So what was the deal with gathering notification requirements from other open stack teams? Isn't that a bit contradictory? 15:12:05 <kgriffs> #action kgriffs to address over vs. under cloud in FAQ, mention on home page 15:12:08 <malini> So probably we shud clarify tht it is not our intent? 15:12:12 <flaper87> To be clear, by under cloud I mean "replacing rabbitmq" 15:12:21 <megan_w> sorry, i mean we shouldn't focus on features needed for the aggregate and normalize use case 15:12:50 <flaper87> I think it is fair for other projects to consume Marconi (Telemetry, trove, swift, etc) 15:13:10 <flaper87> but I don't think Marconi is a drop-in replacement for rabbitmq in openstack 15:13:26 <megan_w> got it 15:13:30 <ametts> flaper87, kgriffs: Should we clarify our AMQP plans wrt to not "replacing rabbmq"? 15:13:32 <kgriffs> I don't think so either, but it may make sense in a complimentary role 15:13:57 <kgriffs> ametts: I'll try to make that clear in the FAQ 15:14:09 <flaper87> ok, lets move on 15:14:24 <flaper87> besides that, I think the summit was very positive for marconi 15:14:34 <flaper87> community wise and project wise 15:14:57 <malini> w.r.t QA, we need to refactor the functional tests to be Tempest friendly 15:15:03 <flaper87> #info I started this thread: http://lists.openstack.org/pipermail/openstack-dev/2014-May/035341.html 15:15:06 <malini> The neutron folks have a POC 15:15:19 <Obulpathi> POC? 15:15:31 <flaper87> malini: what does that mean? We should probably discuss it further off-line 15:15:40 <malini> flaper87: sure 15:15:43 <abettadapur> proof of concept 15:15:50 <ametts> flaper87: Can you really call it a thread if yours is the only message? :) 15:15:53 <flaper87> not sure why tempest needs us to refactor our functional tests 15:15:59 <Obulpathi> abettadapur: thanks 15:16:03 <flaper87> ametts: LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL 15:16:10 <kgriffs> megan_w: I think that much of what people will want to do with notifications in the overcloud will apply to the internal notifications use cases; let's see how much we can accomplish with the same code 15:16:11 <flaper87> ametts: a hope-for-a-thread 15:16:17 <malini> flaper87: tempest doesnt need us to..But will help us avoid some duplication of effort 15:16:24 <megan_w> sounds good 15:16:40 <malini> flaper87: it'll be using tempest http libs etc. 15:16:52 <malini> flaper87: lets chat more in #openstack-marconi 15:17:00 <flaper87> malini: mmmh, that alone scares the hell out of me 15:17:03 <flaper87> malini: sure 15:17:21 <malini> flaper87: me too :D 15:17:59 <flaper87> next topic ? 15:18:20 <kgriffs> #action kgriffs to document roadmap and send to ML for feedback 15:18:55 <kgriffs> #topic Change the program name 15:19:25 <flaper87> kgriffs: project* 15:19:28 <kgriffs> https://etherpad.openstack.org/p/marconi-scratch 15:19:46 <kgriffs> so, we need a new "code name" 15:20:21 <kgriffs> after we figure that out, we also need a new "regular" name 15:20:52 <flaper87> Program: Queuing, Project: Marconi. We're discussing Marconi now, right? 15:21:04 <megan_w> any sort of theme we want to portray in our name? 15:21:16 <kgriffs> flaper87: right, Marconi 15:21:17 <malini> do we need something tht will encompass notifications too? 15:24:46 <kgriffs> swift, sahara, nova, etc. don't have much to do with what their projects actually do. let's not try too hard. 15:25:33 <notmyname> (marconi is a really great name for a system that broadcasts messages to listeners) 15:25:46 <notmyname> kgriffs: also, are you saying swift is slow?!?!? ;-) 15:26:12 <kgriffs> notmyname: that's *precisely* what I was implying. :) 15:26:26 <kgriffs> unfortunately, someone owns the Marconi trademark... 15:26:45 <malini> will somebody do the vetting tht our new name wont have legal issues too? 15:26:52 <megan_w> yes 15:26:54 <megan_w> i'll do that 15:27:13 <megan_w> marketing recommended we come up with three options and submit them for legal review 15:27:38 <malini> great! 15:28:13 <kgriffs> take your pick 15:28:13 <kgriffs> https://en.wikipedia.org/wiki/List_of_rivers_in_India 15:28:48 <malini> kgriffs: which of those can you remember the easiest? 15:28:57 <notmyname> carrier (as in pigeon), {copper | fiber} (as in what carries messages), express (as in the pony express) 15:28:59 <malini> qn open to all non-Indians 15:29:29 <malini> notmyname: Cuprum 15:29:57 <kgriffs> copper 15:29:58 <kgriffs> hmmm 15:29:59 <kgriffs> not bad 15:30:29 <megan_w> yeah, i like copper too 15:31:07 <malini> Can we call it Cuprum ? the Latin name http://en.wikipedia.org/wiki/Copper 15:31:28 <mpanetta> Hmm 15:31:57 <malini> we are too creative with names! 15:32:26 <Obulpathi> Cuprum +1 15:32:44 <tjanczuk> In Polish, "Cuprum" has a very close connotation to the end of the intestinal tract opposite of mouth. But I can survive it. 15:32:59 <kgriffs> ROFL 15:33:12 <vkmc> lol 15:33:12 <kgriffs> tjanczuk: what is 15:33:14 <flaper87> people, lets get some votes on the etherpad 15:33:16 <kgriffs> "river" in polish? 15:33:23 <tjanczuk> rzeka 15:33:26 <megan_w> ok, why don't we pick one "river" name, one message carrier name, and maybe one more 15:33:27 <malini> tjanczuk: we definitely dont want Cuprum, in tht case 15:33:31 <kgriffs> flaper87: let's get a final, master list. 15:33:55 <megan_w> then if legal says they are all free, we can do a vote 15:33:55 <kgriffs> Let me make a new list at the bottom. Everyone put your 2 favorites on it 15:33:58 <flaper87> yup, that's what I meant. I put some +1 on the ones I like 15:34:17 <tjanczuk> One last one: tamtam. It is used for sending signals. and as far as I can tell it is trademark free. 15:34:20 <abettadapur> river = rzeka(polish) = rio(spanish) ... 15:35:56 <malini> I like tamtam 15:36:11 <cpallares> Zaqar = messaging god in Mesopotamian mythology 15:36:28 <megan_w> cpallares: thats cool 15:36:33 <mpanetta> Oh sweet 15:36:45 <alcabrera> Zaqar, nice 15:36:45 <mpanetta> There is a messaging god? hah that is cool. 15:36:51 <sriram> Zaqar sounds really cool! 15:37:32 <kgriffs> hmm 15:37:33 <kgriffs> rio 15:37:35 <kgriffs> not bad either 15:37:50 <megan_w> ok, are we cool with submitting copper, raven, and zaqar? 15:37:51 <kgriffs> but, probably trademarked all over the place 15:38:01 <malini> how do u pronous Zaqar? 15:38:02 <megan_w> or do we want to keep it open a while longer? 15:38:03 <AAzza> how zaqar is pronaunced?) 15:38:05 <malini> pronounce* 15:38:10 <malini> hard via IRC :D 15:38:16 <megan_w> zah-car? 15:38:17 <mpanetta> I am guessing like za car? 15:38:18 <abettadapur> za - kar? 15:38:18 <malini> AAzza: same concern 15:38:19 <mpanetta> yeah 15:38:39 <tjanczuk> Did anyone think of putting up to to a vote via twtpoll? 15:39:30 <malini> Tamtam will work well with notifications too https://www.google.com/search?q=Tamtam&es_sm=91&tbm=isch&tbo=u&source=univ&sa=X&ei=-3Z7U-PlEY2VqAbwuYHgDw&ved=0CEMQsAQ&biw=1152&bih=551 15:39:57 <abettadapur> people seem to like naav 15:40:09 <malini> I dont like how it sounds :( 15:40:11 <flaper87> naav FTW 15:40:15 <flaper87> :P 15:40:23 <megan_w> ok, copper, naav, zaqar? 15:40:40 <flaper87> megan_w: looks like, I like those 3 15:40:40 <alcabrera> "nube" would amuse me greatly 15:40:43 <malini> Tamtam & Zaqar are tied megan_w 15:40:46 <flaper87> and sounds like folks do too 15:40:47 <alcabrera> megan_w: yup 15:40:48 <megan_w> we could always submit 5 and then vote after legal reviews them 15:40:58 <malini> except someone rigged the vote for Zaqar 15:41:00 <mpanetta> alcabrera: nube? 15:41:01 <cpallares> alcabrera: how do you pronounce nube? 15:41:10 * cpallares reads it as noob 15:41:11 <alcabrera> nube -> cloud (spanish) 15:41:13 <alcabrera> noo-beh 15:41:13 <cpallares> ah 15:41:15 <mpanetta> Ah! 15:41:18 <abettadapur> hehe 15:41:26 <abettadapur> lack of language context there 15:41:34 <alcabrera> :) 15:41:39 <kgriffs> meldung 15:41:45 <kgriffs> = message in German 15:41:50 <malini> kgriffs: sounds a lil yuck 15:41:53 <cpallares> LOL 15:41:53 <kgriffs> lol 15:41:58 <kgriffs> mell-doong 15:42:06 <malini> not so bad now 15:42:06 <megan_w> time check... 19 minutes left in meeting 15:42:12 <flaper87> oooook, I guess we have our options 15:42:16 <flaper87> megan_w: btw, thanks a lot :) 15:42:17 <kgriffs> but yeah, nobody will know how to say it right 15:42:35 <kgriffs> megan_w: if they will check 5, let's do that 15:42:39 <mpanetta> That is half the fun heh 15:42:43 <megan_w> cool 15:42:49 <megan_w> i'll take this from here then 15:42:57 <kgriffs> #action megan_w to check trademarks for our shortlist of names 15:43:31 <kgriffs> #topic open discussion 15:43:52 <tjanczuk> I'd like to touch on AMQP, in particular 0.9 vs 1.0 15:44:03 <flaper87> I want to propose changing the "Open Discussion" topic with "Open Party" 15:44:09 <flaper87> tjanczuk: shoot 15:44:30 <tjanczuk> I think someone mentioned during the summit there was some community feedback gathered around this? Can you point me to it? 15:44:59 <flaper87> tjanczuk: it was probably me, I'll get the emails from the mailing list 15:45:10 <flaper87> there are also IRC logs from oslo meetings 15:45:25 <flaper87> and well, the session we had last week at the summit 15:45:46 <tjanczuk> During the session I felt I did not have the data you did, hence my question now. 15:45:53 <kgriffs> flaper87: maybe you can do a summary and include that on a wiki page for the AMQP driver blueprint? 15:46:19 <flaper87> kgriffs: sounds like a good thing to have 15:46:37 <kgriffs> flaper87: is there a blueprint yet? 15:46:54 <flaper87> tjanczuk: sure, probably because I've been working closely in the amqp 1.0 implementation. I'll make sure you get the same info I have 15:47:01 <flaper87> (which is all up there) 15:47:05 <flaper87> kgriffs: I believe so, 2 secs 15:47:08 <tjanczuk> What AMQP 1.0 product are you planning to use? 15:47:31 <kgriffs> #action flaper87 to summarize 0.9 vs. 1.0 discussion in the context of openstack, add to AMQP driver bp, send to ML 15:47:37 <flaper87> I'm planning to use a amqp1.0 library. It's called qpid-proton 15:47:52 <flaper87> it can target any broker supporting amqp1.0 15:48:06 <tjanczuk> Does this implement the actual broker logic, or is this a client lib? 15:48:15 <flaper87> kgriffs: https://blueprints.launchpad.net/marconi/+spec/storage-amqp 15:48:30 <flaper87> client library 15:48:59 <flaper87> The broker can be: rabbit+plugin, qpid, apollo, activemq, or even qpid-dispatch with a mix of brokers 15:49:19 <flaper87> well, I wouldn't recommend mixing them but, just sayin' 15:49:20 <tjanczuk> I see. Would this implementation come with a recommendation for an actual broker to use? I think TC's expectation after the Mongo feedback was that there is an actual software stack that is a viable alernative. 15:49:58 <flaper87> tjanczuk: it'll likely be rabbit or qpid 15:50:14 <tjanczuk> rabbit's amqp 1.0 implementation is a toy. 15:50:41 <tjanczuk> So that would make it qpid? If you were deploying it for your mother, what would you use? 15:50:50 <flaper87> tjanczuk: I know, but it's still there. Rabbit and qpid are both supported by the community 15:51:34 <kgriffs> flaper87: can you bring this up with Devandanda and also post our plan to the ML? I would like operators to chime in before we commit. 15:52:16 <flaper87> kgriffs: I can but this has already been discussed community wise. The same library is being used for oslo.messaging 15:52:24 <tjanczuk> Basically I am concerned with this picture: http://www.google.com/trends/explore#q=rabbitmq%2C%20qpid&cmpt=q 15:52:30 <kgriffs> flaper87: is oslo messaging going to remove support for 0.9 then? 15:52:45 <flaper87> kgriffs: not remove support for 0.9 but it'll support 1.0 15:52:47 <tjanczuk> To me this is one data point about what community at large is using. 15:53:28 <tjanczuk> It also cannot be ignored that OpenStack project generally use Rabbit. 15:53:45 <flaper87> tjanczuk: TBH, that says pretty much nothing to me. We've the same in openstack. 80% of the community is using rabbit and the rest is using qpid. However, they both have serious issues 15:55:00 <tjanczuk> I am having a hard time with 80% of community using rabbit and at the same time requesting something else. 15:55:27 <flaper87> I'd like to check w/ folks in the rabbitmq team and get feedback from them about the amqp 1.0 support. Nontheless, I don't think supporting an 0.9 driver in the code base is the right thing to do 15:55:39 <flaper87> tjanczuk: I didn't say the 80% of the community want's amqp 1.0 15:55:44 <flaper87> wants* 15:56:21 <tjanczuk> I understand you didn't. So what did the community actually say? 15:56:22 <kgriffs> if we make a really awesome Redis driver, then people who don't want AGPL can use that if they don't want to deal with AMQP 1.0 15:56:50 <kgriffs> but that will still leave some subset 15:57:32 <kgriffs> ...of operators who want to deploy Marconi on AMQP and are invested in Rabbit 15:57:35 <tjanczuk> From the deployment perspective this is another product to deploy, maintain, and have a know-how about, regardless if this is redis or amqp 1.0 product. 15:57:38 <vkmc> talking about this, I would appreciate your comments on which storage backend you would like to see supported by Marconi... I've wrote a small list with the candidates here https://etherpad.openstack.org/p/storage-backend-marconi 15:58:05 <flaper87> openstack deployments already have a broker deployed 15:58:29 <tjanczuk> ... and that is not an amqp 1.0 product in most cases. it is rabbit. 15:58:41 <kgriffs> tjanczuk: yes, some operators have that concern. Others don't worry so much about adding Redis, since it is already pretty commonly used in app development... also should be more common once we add support for redis to keystone middleware via oslo.cache 15:58:55 <kgriffs> hard to please everyone. :p 15:59:19 <flaper87> every openstack module has a minimum requirement. Either it is a broker, database, cache system etc 15:59:20 <kgriffs> vkmc: let's chat in #openstack-marconi 15:59:38 <flaper87> they're not for free 15:59:40 <tjanczuk> Don't get me wrong, I am not pushing against amqp 1.0 if that is what folks want, I just want to understand the rationale. is this community? Is this technical? Or is this something written between the lines? 15:59:58 <flaper87> tjanczuk: I believe is both 16:00:15 <vkmc> kgriffs, sure! 16:00:17 <flaper87> It's* 16:00:23 <flaper87> time's up guys 16:00:26 <flaper87> great meeting 16:00:26 <alcabrera> yup 16:00:39 <flaper87> Thanks everyone! 16:00:41 <kgriffs> I'll try to get some feedback on AMQP 1.0 from Rackspace ops 16:00:56 <Obulpathi> Thanks and nice meeting you all at the summit 16:00:58 <kgriffs> #action kgriffs, megan_w to get some feedback on AMQP 1.0 from Rackspace ops 16:01:02 <kgriffs> #endmeeting