15:00:04 #startmeeting Zaqar 15:00:05 Meeting started Mon Sep 8 15:00:04 2014 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:06 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:08 The meeting name has been set to 'zaqar' 15:00:11 #topic Roll Call 15:00:14 o/ 15:00:20 vkmc: malini flwang ? 15:00:25 o/ 15:00:48 I think malini is not around 15:01:01 #topic Review Items from last week 15:01:07 #link https://wiki.openstack.org/wiki/Meetings/Zaqar#Agenda 15:01:12 that's the agenda for today 15:01:18 #link http://eavesdrop.openstack.org/meetings/zaqar/2014/zaqar.2014-09-01-21.00.html 15:01:35 kgriffs to re-run benchmarks using the redis driver 15:02:03 I started working on that last week. goal is to finish today - publish something at least, in advance of the next TC mtg 15:02:16 kgriffs: great news, thanks! 15:02:39 #info kgriffs is working on re-running benchmarks on redis. Hopefully ready before the TC meeting tomorrow 15:02:42 vkmc to add checks enforcement rule to our development docs 15:03:33 vkmc: ? 15:03:57 flaper87, what do you mean by checks enforcement rule? 15:04:03 sorry I'm lost 15:04:15 dunno, that's an action item from last week 15:04:17 :D 15:04:24 * flaper87 tries to figure out what that is 15:04:30 let me check the logs 15:04:35 ahhh 15:04:48 so, we have to document when an API check has to go to the storage layer or the API layer 15:05:07 vkmc: ^ 15:05:27 I can take that 15:05:46 #action To document where API checks should go (API or Storage layer) 15:05:46 flaper87, oh it makes sense 15:05:53 I can do that 15:05:57 #undo 15:05:58 Removing item from minutes: 15:06:01 #action vkmc To document where API checks should go (API or Storage layer) 15:06:10 ok, moving on 15:06:18 #topic Early adoption of dox (flaper87) 15:06:31 not sure if you guys read the email thread about dox 15:06:40 I started hacking on it and it seems like a neat idea 15:06:45 #link https://github.com/stackforge/dox 15:06:52 #link https://review.openstack.org/#/c/119740/ 15:07:06 so, I've proposed a review that suggest becoming early adopters/users of this tool 15:07:13 dox is basically tox on top of docker 15:07:37 ideally, developers shouldn't need to install redis and whatnot to test Zaqar, instead they'd just need docker and dox 15:07:38 good stuff 15:07:53 with that, no one will have an excuse not to run all tests locally 15:07:55 :D 15:08:11 neat! 15:08:12 flaper87: how do you see this in relation to devstack? 15:08:23 I hacked on dox over the weekend and I worte the dox.yml file for zaqar 15:08:41 kgriffs: I think we'll endup having a docker image with devstack in it 15:08:54 the ultimate goal is to have images like `infra/zaqar/trusty` 15:09:01 I like the idea of deploying a "production cluster of dockers" instead of a big gnarly shell script 15:09:03 so that developers don't even need the `prep` section 15:09:16 kgriffs: right 15:09:23 let's us realize the isolation benefits wrt dependencies that a SOA is *supposed* to provide 15:09:31 +1 15:09:33 15:09:39 We'll see how it evolves 15:09:44 kk 15:09:59 So, If you think it makes sense to adopt it, +2 otherwise, I'll STFU and ktanksbai 15:10:17 any other question on dox? 15:10:40 i'll take that as a no 15:10:41 not from me 15:10:45 #topic Goals for the next couple of weeks (flaper87) 15:10:56 Soooo, j-3 was cut last week. w0000000000000000t 15:11:07 we've rc1 ahead but with reaaaaallly few things in it 15:11:18 I'd like us to focus on getting the client back in shape 15:11:27 +1 15:11:39 it's quite behind Zaqar and it'd be nice if we can get to the release with the client fulling supporting v1.1 15:11:44 that should help us also smoke test j-3 and catch bugs 15:11:50 +1 15:12:04 ok, I'll start by adding the v1.1 spec to the client 15:12:14 and we can work on the missing features on top of that 15:12:17 I am finishing up the last Redis patch I wanted to land 15:12:19 we still have to finish the support for v1.0 in the client 15:12:24 important performance patch 15:12:31 #action flaper87 to start adding v1.1 to the client 15:12:33 vkmc: wiat, what? 15:12:36 what's missing? 15:12:44 I thought we had full support for it 15:12:44 shouldn't it make sense to get feedback from the community about v1.1 and then add the support for v1.1? 15:12:56 flaper87: I think vkmc is referring to CLI commands 15:12:57 let me rephrase, what's missing that was not added in Juno? 15:13:02 aaaaaaah ok ok 15:13:04 oh sorry 15:13:04 yes 15:13:09 well, that makes sense 15:13:11 yeah 15:13:18 and a full series of screencasts. 15:13:23 ;) 15:13:25 vkmc: I know you started working on that a couple of weeks ago 15:13:43 what about you keep taking care of the CLI and I worked on the v1.1 support for the *library* ? 15:13:47 flaper87, I did yeah, I was waiting for some feedback to keep working on it 15:13:51 Does that sound good to you? 15:14:01 of course it does 15:14:06 awesome 15:14:15 yeah, we also decided docs was higher priority at the time 15:14:25 indeed! 15:14:26 vkmc has been working miracles w/ docs 15:14:30 sorry I mixed terms 15:14:32 I'm happy with the current state, really! 15:14:43 vkmc++ 15:15:20 ok, any other comment/thought ? 15:15:39 ok so, start adding v1.1 to the *client* and finish the *cli* for v1.0 15:15:39 #topic Final graduation meeting tomorrow (flaper87) 15:15:44 #undo 15:15:45 Removing item from minutes: 15:15:49 vkmc: yeah 15:16:00 the work on the CLI for v1.0 should be good for v1.1 too 15:16:01 +1 15:16:02 I think 15:16:29 The better the library is, the easier it'll be for others to start using zaqar 15:16:30 * kgriffs notest #undo command for future reference 15:16:36 #topic Final graduation meeting tomorrow (flaper87) 15:16:38 kgriffs: ;) 15:16:43 ok ok ok ok 15:16:47 I bet you didn't know this 15:16:51 tomorrow is the final review 15:16:52 :D 15:17:02 A summary of what happened last week 15:17:37 I think the TC meeting went well. There were some concerns raised about NoSQL being added as a requirement, whether we should add an optional project as integrated etc 15:17:48 * flaper87 wishes he had the link to the email thread handy 15:18:18 Those concerns were sent to the mailing list to start a discussion before our meeting tomorrow 15:18:47 https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg33997.html 15:18:48 We haven't gotten as much feedback as I'd have liked 15:18:53 vkmc: thanks :D 15:18:55 :) 15:18:58 #link https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg33997.html 15:19:11 anyway, tomorrow is the big day 15:19:29 is there anything we could do to be 'more' prepared? 15:19:34 Whether we're graduating or not, I think Zaqar is a gread product and ready to be consumed by big players 15:19:46 vkmc: that's actually one of the things I wanted to ask :D 15:19:50 you stole my thunder 15:19:57 sorryyyy 15:19:59 Do you think there's something missing for tomorrow? 15:20:05 vkmc: kidding 15:20:18 today I'm being a pain in the neck 15:20:20 Do you reckon we should prepare more material before our meeting tomorrow? 15:20:33 If any, will that actually help? 15:20:45 flaper87: how is that etherpad coming w/ use cases? 15:20:53 I've prepared - with the help of many gentle folks from other teams - this etherpad with use cases: https://etherpad.openstack.org/p/zaqar-integrated-projects-use-cases 15:20:59 LOL 15:21:03 I think it may be helpful to highlight why those use cases couldn't just be done on top of RabbitMQ 15:21:16 yeah... that etherpad is key 15:21:20 * flaper87 kicks kgriffs out of his brain 15:21:29 because, that is what some folks are hinting at 15:21:32 I need to get some Trove folks to look at it 15:21:36 I'll do that today 15:21:47 user agent is a big one 15:21:56 can you get some guys from that project to comment? 15:22:01 I think there're some really neat/key use-cases in the etherpad already 15:22:05 but we need moar 15:22:09 yup 15:22:19 I'll chase them down to hell if needed 15:22:36 I say that because it would be silly to try to expose a Rabbit endpoint to the guest 15:22:43 * kgriffs in my opinion 15:22:56 They were in the session at the Juno summit and I talked with them afterwards 15:22:59 firewall, and also security issues, and stuff 15:23:09 They mentioned several times the will of adoption Zaqar once it graduates 15:23:15 cool 15:23:15 the use case is precisely the user-agent thing 15:23:42 s/adoption/adopting/ 15:23:50 anyway, me and my silly typos 15:23:53 Rackspace has also talked about using Zaqar to communicate with the backup agents used in Cloud Backup, although that product is not OpenStack, it does highlight an "app/saas use case" 15:24:16 kgriffs: would it be worth it to mention it in the use-cases list? 15:24:22 flaper87: I think so 15:24:30 I think so too 15:24:30 awesome, lets add it 15:24:35 flaper87: try to track down ametts and see if they are still thinking about it 15:25:07 kgriffs: kk, I'll see if he's around otherwise I'll just call his wife and ask where he is 15:25:13 sweet 15:25:24 * flaper87 will chase everybody down!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 15:25:29 no mercy 15:25:45 (isn't that the name of some boy-band ? 15:25:47 ) 15:26:05 anyway. Other than that, is there anything you think we should totally do? 15:26:29 flaper87: I'll send an email to Mr. Metts 15:26:53 kgriffs: awesome, that'd be great 15:27:13 nothing else comes to my mind, really 15:27:22 I believe everything that had to be said has been said 15:27:39 pretty please everone review my redis patchset in 30 mins 15:27:47 (will upload soon) 15:27:50 I'll take a look at the logs of last week meeting and see if I spot something else 15:28:00 kgriffs: +1 15:28:06 has anyone heard from prashanth lately? 15:28:17 kgriffs: yeah, he showed up then dissapeared 15:28:19 vkmc: ? 15:28:19 hmm... and the NoSQL concern? is there something else to say there? 15:28:47 vkmc: girl, you almost made me say something I was going to regret :D 15:28:48 kgriffs, I saw him last week, when we chatted about the gc script 15:28:55 it's a little frustrating - they say SQL is an anti-pattern, so we use NoSQL and they say "ain't nobody gonna deploy that" 15:28:58 vkmc: I don't think so, I mean. We can't do much there 15:29:17 flaper87, :x should I STFU for the rest of the day? lol 15:29:25 WE depend on it and if that is a real issue, which I don't think it is, then I dunno what to do 15:29:36 vkmc: hahaha, nooo. "It's not you, it's me" 15:29:38 (ROFL) 15:29:51 oh c'mon I don't believe you hahaha 15:30:22 well, yeah... there is not much to do there 15:30:32 yeah :/ 15:30:51 I don't buy that of 'nobody gonna deploy that' 15:31:07 well, that's what people said about openstack when it first came out 15:31:09 ;) 15:31:12 someone needs to call them on this catch-22 15:31:16 if they are deploying OpenStack then NoSQL shouldn't be a problem 15:31:52 seriously, I do see the problem about "adding a new technology to the stack" but I don't think that's the real concern 15:31:57 that just happened to be there 15:32:21 the real concern is the "should we add a project that is optional to the integrated team, which also adds a NoSQL dependency to the stack" 15:32:32 or something along that 15:32:51 it's not Zaqar only 15:32:56 vkmc: I know 15:32:58 other projects use NoSQL as well 15:32:59 flaper87: I think their concern was about training people to use it, no? 15:33:24 cpallares: that was also there, yeah. 15:33:35 but still, it's part of the growth and evolution 15:33:39 I'm surrounded by sys admins at work. They have no trouble deploying, say, Elasticsearch if the tool they want to use requires it. 15:33:54 I think admins are smarter than many give them credit. 15:33:57 +1 kgriffs 15:33:58 kgriffs: That's a good argument. 15:33:59 kgriffs: exactly, they'll just figure it out 15:34:21 anyway, we all agree on that 15:34:29 also... with that thinking we wouldn't be able to achieve greater things 15:34:29 lets bring all this comments to the meeting tomorrow 15:34:40 we'll see what happens 15:35:09 I mentioned several times that I'd be happy if Zaqar gets moved to an 'optional-projects' group once that group exists 15:35:27 flaper87: yeah 15:35:28 '640k ought to be enough for everybody' thinking 15:35:47 me too 15:35:55 great 15:36:01 so, I've nothing else to add there 15:36:17 but it is silly to hold up graduation based on pie-in-the-sky-maybe-someday thinking around the future of the integrated release. In the meantime, we could be delivering a lot of value to other projects and app developers. 15:36:45 indeed, that's the point I hope we'll be able to make tomorrow when this'll come up 15:37:00 kk 15:37:22 One way or another, Zaqar will be blessed one day so why should we delay it if technically speaking it's up for the game 15:37:24 ? 15:37:36 cool, that's all I've got. 15:37:36 +1 flaper87 15:37:58 #topic Open Discussion 15:38:24 So, I finally watched all episodes of True Detective seasson 1 15:38:28 LOL 15:38:28 ZOMFG 15:38:43 I was going to mention that I saw Undateable 15:38:44 the seasson two will come out in summer 2014 15:38:48 what's wrong with people ? 15:38:59 vkmc: isn't it fun? 15:39:03 s/saw/watched 15:39:13 vkmc: you should also watch `Mom` 15:39:17 yup 15:39:18 it's really fun too 15:39:25 will do :) 15:39:31 I also gave a shot to The Wire 15:39:50 That one didn't catch me, I should probably try again 15:39:59 I couldn't finish the first episode 15:40:07 lol 15:40:26 but probably it's me 15:40:31 hahaha 15:40:40 it's not you, it's them 15:40:41 :P 15:40:54 anyway, something else folks want to talk about? 15:41:00 otherwise, I'll call it! 15:41:28 ok, thanks guys. Hope to see you all tomorrow at the TC meeting 15:41:32 nobody else wants to discuss about tv shows 15:41:34 * vkmc sad panda 15:41:35 remember, it's at 20:00 UTC 15:41:37 lol 15:41:40 * flaper87 sad panda 15:41:50 I've heard good things about true detective 15:41:57 must watch cpallares 15:42:02 cpallares: it's A.M.A.Z.I.N.G 15:42:09 * cpallares will watch 15:42:23 Watch tru detective you must 15:42:23 be careful with flaper87, he is a spoiler 15:42:26 true* 15:42:29 :O 15:42:30 vkmc: WHA? 15:42:35 that's not trueeeeeeeeeeeee 15:42:41 lol 15:42:54 I had to quit IRC and delete all logs because vkmc was spoiling true detective 15:43:00 my eyes were bleeding 15:43:13 vkmc: :O et tu? 15:43:28 tsk, tsk 15:43:30 cpallares: so now you don't believe me, ah? 15:43:32 >.> 15:43:32 cpallares, I didn't do that 15:43:48 in any case, it's kgriffs' fault! 15:43:55 byeeeeeeeeeeeeeeeeeeeeeeeeeeee 15:43:57 It's always kgriffs fault 15:43:58 #endmeeting