15:00:02 #startmeeting Marconi 15:00:06 Meeting started Tue Apr 15 15:00:02 2014 UTC and is due to finish in 60 minutes. The chair is malini. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:07 o/ 15:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:10 The meeting name has been set to 'marconi' 15:00:11 o/ 15:00:17 o/ 15:00:22 I thought I'd forget 15:00:29 This dependent types paper is *too* good 15:00:35 anywho 15:00:38 o/ 15:00:38 welcome, all! 15:00:39 >—o-> 15:00:42 hello jchai! thanks for joining 15:00:46 * sriram jumps right in 15:00:46 Hi hi :) 15:00:51 others get no gratitude :D 15:00:54 Hello 15:00:54 hello vkmc 15:01:13 anybody else here for Marconi? 15:01:14 jchai: hey! glad to have you join us. :D 15:01:18 vkmc: w00t! good to see you! 15:01:45 jchai is working on adding tests in Tempest for Marconi 15:02:06 Thank You jchai for your behind the scenes effort :) 15:02:07 awesome 15:02:35 Good to see you all :) 15:02:45 https://wiki.openstack.org/wiki/Meetings/Marconi#Agenda get to the agenda 15:02:46 yo yo yo 15:02:54 o/ 15:02:56 #link https://wiki.openstack.org/wiki/Meetings/Marconi#Agenda 15:03:05 hello amitgandhi 15:03:17 First action items from last time 15:03:32 hi 15:03:35 Obulpathi: hello 15:03:46 * malini balajiiyer to get report on queue metadata usage by Rackspace customers 15:04:04 balajiiyer: did you see a lot of queue metadata usage ? 15:04:31 malini: no, not really. it was less than 1% in all datacenters 15:04:32 hello flwang! 15:04:44 malini: Hey :D 15:04:45 flwang: 'sup buddy ? 15:04:48 balajiiyer: that's interesting 15:04:55 flaper87: good, you? 15:05:10 balajiiyer: do you have any insights on how customers use the metadata? 15:05:35 #info queue metadata usage is < 1% @ Rackspace 15:06:03 malini: I looked at some of them, didnt make much sense to me, probably test data. But I can send you a dump of metadata we have for analysis 15:06:31 so, its <1% users or <1% queues? 15:06:38 balajiiyer: did you mean most of the metadata are test/fake data? 15:06:50 Obulpathi: good question 15:07:00 :) 15:07:02 yup 15:07:11 balajiiyer: can you clarify wrt to Obulpathi's question? 15:07:27 flwang: the first set of data I poured my eyes on were test data. I can get you the dump for analysis 15:07:36 Obulpathi: 1% of queues 15:07:45 I think what we need to figure out is: Would they freak out if the next version of the API doesn't have support for metadata? 15:07:48 balajiiyer: cool, thanks 15:08:03 ok 15:08:39 I'm not sure what 1% is in numbers but it's probably already worth an upgrade plan 15:08:58 can we email the customers using them metadata? 15:08:59 flaper87: got it. it might mean that we need to reach out to some of the customers to understand how they are using it and how it will impact them 15:09:04 #info clarification: 1% of all queues created @ Rackspace contain metadata 15:09:10 yeap..the key question is will we break existing users 15:09:14 balajiiyer: that would be awesome 15:09:18 balaji: yep .. thet would be great 15:09:32 megan_w: can help with reaching out to those customers 15:09:48 megan_w: can you take care of that plz? 15:09:52 also, FWIW, if in doubt, we can just provide the upgrade plan we talked last week 15:10:06 create an action item, but megan_w told me once earlier that customers aren't really responsive, so we have to see how it goes 15:10:27 we need more followup on this 15:10:28 it's not hard to do, it's very conservative and plays nice with user data 15:10:41 the feature will go away for sure, lets just make sure we don't blow user's data 15:11:01 #action: balajiiyer, megan_w to investigate customer impact when queue metadata support is removed 15:11:20 any more questions on this? 15:11:34 none from me 15:11:39 nope 15:11:42 none from my side 15:11:53 ok..moving on 15:12:08 We dont have any other action items listed 15:12:33 But we need to track where we are on FAQ updates, ask 15:12:49 Anybody has the FAQ etherpad link handy? 15:12:59 also, any update on the status of the POP? 15:13:14 #link https://etherpad.openstack.org/p/draft-marconi-faq 15:13:49 I have not started on the test wiki, but will do tht this week 15:14:07 #action: Malini to build a test wiki for Marconi 15:14:22 thanks, malini 15:14:25 re: the faq 15:14:30 Please, guys, lets finish that FAQ 15:14:33 I haven't investigated oslo.messaging yet 15:14:45 I think the sooner we start the discussion on the ML the better 15:14:46 alcabrera: can you do it this week? 15:14:51 yup 15:14:58 actionize me! :) 15:15:03 that will give us enough time to iterate over the FAQ several times 15:15:10 flaper87: agreed 15:15:17 #action: alcabrera to compare marconi with oslo.messaging 15:15:22 cool 15:16:13 flaper87, flwang: do you have pending tasks for the FAQ? 15:16:37 malini: there are several open questions there 15:16:50 I tackled some of them today, I'll go through the remaining before the enxt week 15:16:51 Can we assign folks to the open questions now? 15:16:56 malini: nothing from me, but I'm available to be assigned something 15:17:04 please assign one to me 15:17:11 I'd prefer folks to go there and add their names right next to the question 15:17:19 flaper87: +1 15:17:28 +1 15:17:31 Obulpathi: feel free to pick whatever question you want to work on 15:17:31 Lets do it before the end of meeting , so we can actionize them 15:17:53 We'll come back to this towards the end of meeting, to add more actions 15:18:21 Lets tackle the FAQ this week & create a similar plan for ask next week 15:18:34 Anything else on this? 15:19:02 moving on.. 15:19:13 #topic: Split graduation list to smaller actionable chunks for each week 15:19:35 I'm a favorable towards this proposal 15:19:38 err 15:19:41 * I am 15:19:45 the idea behind this is to make progress on our pending tasks at a steady pace 15:19:58 we dont want to make a scramble towards the end 15:20:10 +1 15:20:14 ideally I would like the tempest/ gating tasks done before the summit 15:20:16 +1 to malinis proposal 15:20:29 that would be lovely 15:20:33 and I think we can do it 15:20:35 malini: I agree. However, I think we are still not 100% sure what those tasks are. I mean, we know some of them for sure 15:20:35 #link https://etherpad.openstack.org/p/marconi-graduation-chunks 15:20:52 I think the summit is the best moment for us to brainstorm about this 15:20:52 flaper87: good point..we still have a lot of cloudy areas 15:21:01 it's perfect because we'll *all* be there 15:21:03 But we have some tasks we need to get done for sure 15:21:18 (Don't worry, I'll convince flwang to be there) 15:21:21 pun intended malini? 15:21:33 you are the only one who got it mpanetta ;) 15:21:39 * flaper87 still doesn't get how the pun stuff work 15:21:42 * amitgandhi i just sighed 15:21:44 cloudy, cloudy 15:21:44 flaper87: haha 15:21:50 I can see the silver-lining, at least 15:21:56 :D 15:22:00 :) 15:22:00 I'm still reviewing the list to pick one :D 15:22:06 ahhh, "cloudy areas" 15:22:08 got it 15:22:14 * flaper87 is so fucking slow 15:22:24 :P 15:22:25 flaper87: Can we start tackling the areas we know for sure? 15:22:38 tht will leave us free to tackle any new stuff tht comes up 15:22:40 malini: absolutely, I was just about to say that 15:22:48 getting mongo up and running in the gate with Trusty out would be great 15:22:51 awesome 15:22:52 since 15:22:54 so, pretty much everything related to new features is still "cloudy" 15:22:54 that's like 15:22:55 :P 15:22:56 3 days away 15:22:58 I think 15:23:09 hah 15:23:14 but things like tempest 15:23:15 alcabrera: I dont know when infra will have Trusty, but I'll foloow up 15:23:22 malini: thanks! 15:23:26 gate, devstack, etc have to be done 15:23:28 #action: Malini to follow up on Trusty with openstack-infra 15:23:36 flaper87: exactly 15:23:59 Currently our biggest blocker is we still dont have marconi running on devstack 15:24:14 This blocks horizon, tempest & making the jobs voting 15:24:29 We have two bugs tht need to be fixed, to get around this 15:24:44 (In addition to Trusty, which is not a blocker) 15:24:52 well, marconi running on the gate 15:24:57 I think 15:24:58 mmg 15:25:08 ok, I kinda lost track there 15:25:11 flaper87: thanks for the correction 15:25:12 * flaper87 trusts malini 15:25:20 mmg the merciless 15:25:25 this bug is fixed right: https://bugs.launchpad.net/marconi/+bug/1287490 15:25:34 #info Marconi runs on devstack, but not on gate in devstack 15:25:47 amitgandhi: It came back 15:25:52 mpanetta: heh. :P 15:26:08 amitgandhi: with this one https://bugs.launchpad.net/marconi/+bug/1301268 15:26:12 ugh hate it when bugs dont squish right 15:26:42 amitgandhi: the original fix changed the server behavior, making it hard for devs 15:26:53 yeh =/ 15:27:06 * flaper87 should really pay attention to what's going on in LP 15:27:06 anyways Can I get someone to volunteer on https://bugs.launchpad.net/marconi/+bug/1287490 ? 15:27:21 We really need this fixed this week 15:27:34 since it blocks a lot of upstream integration 15:27:34 now that malini has the powerz, I see more bugs assigned to me. NOt sure if it's related to that, though 15:27:35 :P 15:27:37 It's low-hanging fruit, imo. This is a great one for new-comers. 15:27:39 :) 15:27:46 I can do it 15:27:52 cool! 15:27:59 Let me assing to myself 15:28:00 Obulpathi: awesome..Let me assign tht to you 15:28:06 Obulpathi: even better 15:28:09 :) 15:28:15 #action: Obulpathi to work on https://bugs.launchpad.net/marconi/+bug/1287490 15:28:42 Lets get this fixed this week..Nag everybody on openstack-marconi if you need help ;) 15:29:00 the next one is https://bugs.launchpad.net/devstack/+bug/1294068 15:29:19 I am not so sure we want to fix this..but want to hear everybody's thoughts 15:29:26 #link https://bugs.launchpad.net/devstack/+bug/1294068 15:29:35 i vote no if we arent going to support mysql 15:29:52 would rather focus energy around mongo and X 15:29:57 amitgandhi: +1 15:30:00 where X is redis or whatever 15:30:07 malini: Is it *just* a mysql issue? 15:30:16 I know the bug report says mysql but I want to make sure 15:30:17 flaper87: yes..sqlite runs fine 15:30:25 on gate i.e 15:30:39 I'd say, lets not worry (prio low) about it now 15:30:44 malini: maybe we missed some mysql specific config 15:30:46 I dont know why mysql does not work at the gate, but does outside 15:30:48 flaper87: +1 15:30:51 +1 15:30:52 but saying we won't fix it sounds bad 15:31:05 I propose we drop sqla as a core platform once we have redis|amqp available 15:31:06 flwang: do you want to take a look at this one? 15:31:08 are we deprecation mysql? 15:31:08 especially because, although not recommended, we have support for it 15:31:28 amitgandhi: not now, for sure 15:31:35 probably in the not so far future 15:31:48 flaper87: how far? before our next grad review? 15:31:51 malini: I can if it's targeted in Juno 15:31:56 dropping sqla, in particular, leads to easier installs (no longer require mysql_config to be installed) 15:31:58 if its a low hanging fruit then just do it 15:31:59 malini: after, I'd say 15:32:09 if its time and energy, then dont bother (my opinion) 15:32:16 alcabrera: does sqlalchemy require mysql_config? 15:32:17 alcabrera: +1 15:32:18 that's weird 15:32:30 flaper87: In that case, we need this fixed - since we'll be asked what backends are used in the gate 15:32:34 I mean, we *HAVE* to remove python-mysql form the requirements *before* the graduation 15:32:34 that's been my experience so far, flaper87. :/ 15:32:41 or 15:32:43 hmm 15:32:45 which is the dep requiring mysql_config 15:32:49 maybe it's python-mysql 15:32:51 yes, that 15:32:55 ok 15:33:09 #action flaper87 remove python-mysql dependency 15:33:18 +7.25 15:33:27 alcabrera: waaa??? Why so low ? 15:33:27 +2.5 15:33:33 Obulpathi: <3 15:33:41 :) 15:33:46 who gives me the remaining .25 ? 15:33:46 triple core, 1 regular, and a quarter to go 15:33:48 #action: flwang to work on https://bugs.launchpad.net/devstack/+bug/1294068 for Juno 15:33:54 +0.25 :P 15:33:59 mpanetta: <3 <3 15:34:00 I din't do my math right :( 15:34:01 ok 15:34:24 Other items in https://etherpad.openstack.org/p/marconi-graduation-chunks 15:34:30 Obulpathi: still <3 15:34:34 Item # 2 is blocked by #1 15:34:45 So lets get 1.1 fixed this week 15:34:50 so devstack really needs a fix 15:34:55 alcabrera: yes 15:34:58 got it 15:35:07 Tempest Updates 15:35:18 will push it to my top priority 15:35:30 jchai: want to update on API tempest tests? 15:35:50 Just getting started 15:36:20 I should be able to submit something later in the week 15:36:22 jchai: do you think you can get the queues API tests patch submitted this week? 15:36:32 Tht will be poistive tests for 5 more APIs 15:36:55 I think so, yes 15:37:15 #action: jchai to work on Queues API +ve tests 15:37:27 #action: Malini to work on adding CLI tests 15:37:36 ok..moving on 15:37:50 #topic: Meeting with TC to discuss graduation 15:38:12 We need to get into the TC schedule again 15:38:22 +1 15:38:29 couple of thoughts there 15:38:39 flaper87: sure..this is all yours 15:38:48 1) We need to have the FAQ ready and more marketing-like docs written 15:39:06 2) We should probably do it after the summit and participate in the cross-project sessions 15:39:18 3) I think we should also have the tasks sorted out 15:39:32 with that plan, we can request a TC meeting to discuss what's expected from us 15:39:38 +1; +1; +1; 15:39:39 also, we need more than 1 meeting 15:39:46 agreed 15:39:50 flaper87: what are specific actions that we need to do between now & the summit? 15:39:51 for instance, we need to discuss the whole falcon + pecan thing, one more time 15:39:58 we need to be regulars to have our voices heard 15:40:06 malini: FAQ and docs 15:40:14 we addressed the FAQ 15:40:23 What do we need for the docs? 15:40:24 also, gather more info about our take on pecan, benefits and downsides 15:40:29 balajiiyer: did an amazing work 15:40:33 flaper87: what's the mean of "participate in the cross-project sessions" 15:40:53 but again, the summit is *THE* right moment to reach to the community 15:40:56 we need to be more active on ask (post questions as well as answers) 15:41:02 flwang: http://summit.openstack.org/ has cross project sessions 15:41:13 we need to talk to the guys working on pecan, we need to talk to the TC members again and to everyone in the community 15:41:25 flaper87: +1 15:41:28 we need to ask for feedback on our current, high-level plan 15:41:30 etc 15:41:45 I think we are in the right direction & have the right ideas 15:41:57 so, between now and the summit, I think we should focus on building the missing FAQ and doc foundations for Marconi 15:42:01 malini: +1 15:42:02 But we are still missing a solid list of stuff that we will do to accomplish these 15:42:33 flaper87: Can you detail the doc foundations? what can we do to get it done? 15:42:44 sure 15:42:50 so, besides the FAQ 15:42:59 It will be ideal if it ends with #actions ;) 15:43:18 I think we need to document our architecture a bit better 15:43:24 +1 15:43:27 it's not clear for people what marconi looks like 15:43:46 A series of cheese laden tubes? 15:43:48 we need to document what we mean with scaling marconi horizontally and what *sharding* means in MArconi's context 15:43:56 maybe obul can do that while he learns it? 15:44:08 sure .. I can document that 15:44:10 we need to document better what transports are or simply *stop* talking about them until we add another transport 15:44:11 etc 15:44:17 all this besides the FAQ 15:44:28 #action: Obulpathi to document Marconi architecture 15:44:31 Once that's done, we should then invite folks to read them 15:44:38 and then chase them down at the summit 15:44:48 mpanetta: Can you or oz do the 'document what we mean with scaling marconi horizontally and what *sharding* means in MArconi's context' ? 15:44:49 sounds like a problem tackled in multiple directions -- e.g., defining marconi 15:44:49 to get feedback 15:44:52 docs, blogs, articles 15:45:17 alcabrera: yeah 15:45:20 malini: Hmm, didn't Oz already do a doc, does it not include that? 15:45:42 Our biggest problem in our graduation attempt was that basically no one knew what marconi looks like 15:45:54 not even the TC 15:46:04 I almost wish it was not called sharding, I always get confused between marconi and mongodb 'sharding' 15:46:04 mpanetta: I dont think we talked with the focus on scaling 15:46:04 not until after we withdrew our graduation request 15:46:05 or what marconi was ;-) 15:46:20 mpanetta: it is sharding after all :P 15:46:24 hmm..even if we create all the docs, will anybody outside of us read that? 15:46:27 mpanetta: but I agree 15:46:38 If creating awareness is our goal... 15:46:40 malini: we don't know that but that doesn't mean we shouldn't have them 15:46:49 I think we can tie this in with our next topic 15:46:58 malini: our goal is to clear all those points among ourselves too 15:47:11 flaper87: good point 15:47:23 +1 15:47:24 for isntance, we know we want to graduate but we don't have a detailed list of how we benefit from graduating 15:47:28 we need to have that too 15:47:37 does everybody agree on stopping the talk about transport, till we have something other than http? 15:47:38 this will help the project and the community 15:47:53 malini: agreed 15:47:59 malini: +1 15:48:02 (note I meant stop talking about it, not stop working on it) 15:48:12 tht is easier than creating a new document ;) 15:48:15 there's some work Cindy is doing that is worth moving forward 15:48:19 that said, I agree 15:48:24 if the focus is on another storage then we dont need to keep talking about transports 15:48:33 yeah 15:48:45 we're 2 weeks away from the summit 15:48:48 (maybe 3) 15:48:49 cool.Lets have a new transport that really works, & then lets start talking abt it 15:48:56 +1 15:49:09 so, I think it is worth dedicating the upcoming weeks to writing docs instead of code 15:49:09 Guess its time to move on to the next topic 15:49:19 flaper87: +1 15:49:27 * amitgandhi dusts of microsoft word 15:49:30 argh 2 weeks... If we are going to do the marconi cluster thing I need to find something else to run it on... 15:49:45 Someone bought all the wandboards heh 15:49:47 can we move on to the next topic? 15:49:55 * flaper87 STFU 15:49:56 malini: yes 15:49:57 yes 15:49:58 #topic ATL summit 15:50:10 mpanetta is eager to talk abt wandboards 15:50:15 So lets hear that first :) 15:50:16 * amitgandhi 10 minute warning 15:50:23 amitgandhi: +1 15:50:28 thanks amitgandhi 15:50:41 can we add these docs into source code under /doc? 15:50:46 Well, someone purchased them all and they have a 6 week lead time, so we have to find another platform. 15:50:55 flwang: +1 15:51:19 If we really want to do this (I think it would be a way to get people looking at it anyway) 15:51:20 mpanetta: its your territory, you get to choose - unless somebody else has other ideas 15:52:15 mpanetta: I hope you are still typing :) 15:52:26 I'm looking 15:52:30 but nothing is as cheap :( 15:52:41 I am always open to suggestions :) 15:52:58 #action: mpanetta to investigate wandboard cluster alternatives 15:53:17 Cool :) 15:53:19 #topic • 'Marconi-Meet the team' dinner 15:53:33 this was kgriffs|afk idea 15:53:35 note - atl hack day in next thur/fri ( flaper87, flwang are free to join us ) 15:53:49 w00t 15:53:58 amitgandhi: good point..tht wud be awesome 15:54:12 amitgandhi: how to join? 15:54:37 amitgandhi: is it f2f? or online? or.... 15:54:40 flwang, amitgandhi: Lets discuss this in #openstack-marconi 15:54:49 malini: sure 15:54:50 We are almost out of time 15:54:50 physically you can fly to the Rackspace Atlanta office ;-) else we can hangout in the #marconi channel 15:55:13 +1 for continuing discussion on hackday in #marconi 15:55:18 kgriffs|afk had the idea of a Meet the team dinner during the summit 15:55:24 * amitgandhi stfu 15:55:28 amitgandhi: send me the air tickets, please 15:55:41 It'll be cool if we can get some of the interested parties from other projects there 15:55:42 meet the team dinner would be lovely 15:55:45 * flaper87 wants to meet everyone on saturday 15:55:49 yes! 15:55:52 hmm.. 15:55:54 malini: yep, it's targeted on 10 IIRC 15:55:54 that is May 10th 15:56:02 I'll mark my calendat 15:56:05 I believe this is going to be us + everybody else in openstack 15:56:06 *calendar 15:56:11 so I *might* remember, :P 15:56:12 flaper87: I'm not sure if I can make it :( 15:56:26 flwang: :( 15:56:35 flaper87: given the moving, you know 15:56:40 I wil join on saturday 15:56:54 I believe we are talking abt diff dinners :D 15:57:16 flwang: dang, that's right. you'll be missed. :( 15:57:24 We probably need a hanging out internally + hanging out with prospective Marconi users 15:57:40 yes 15:57:43 malini: increasing inclusivity is always a plus 15:57:51 alcabrera: +1 15:57:53 alcabrera: yep, bad 15:57:57 I'll see if GSoC students can make it out as well 15:58:03 I don't mind if other people join as long as you all are there <3 15:58:05 *GSoC/OPW 15:58:06 <3 <3 <# 15:58:08 awesome 15:58:11 <3 15:58:28 If there are folks you can reach out personally to invite, please do 15:58:40 remember we are looking for prospective customers too :) 15:59:01 1m -- final thoughts? 15:59:06 #action: everybody to create a list of prospective dinner invitees 15:59:19 I guess tht ends our meeting 15:59:21 * flaper87 invites the marconi team 15:59:24 w000t 15:59:27 haha 15:59:31 we have to do this at marconi grill 15:59:39 Lets continue on #openstack-marconi 15:59:44 is there a marconi grill ? 15:59:45 Thanks everyone!! 15:59:46 O.O 15:59:49 malini: thank you 15:59:49 #endmeeting