19:01:08 #startmeeting swift 19:01:08 Meeting started Wed Jul 30 19:01:08 2014 UTC and is due to finish in 60 minutes. The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:01:11 The meeting name has been set to 'swift' 19:01:37 welcome everyone 19:01:39 who's here? 19:01:43 o/ 19:01:44 o/ 19:01:46 howdy 19:01:48 o/ 19:01:49 o/ 19:01:49 hello 19:01:49 hi 19:01:51 ol/ 19:01:52 o/ 19:02:12 #link https://wiki.openstack.org/wiki/Meetings/Swift 19:02:16 o/ 19:02:17 yo 19:02:18 welcome. there's the agenda ^^ 19:02:26 full hosue today (yay) 19:02:31 hi 19:02:39 #topic housekeeping 19:02:42 got all our aces here, plus a couple jokers 19:03:00 * portante now officially in the joker camp 19:03:02 jsut a few reminders of stuff and things 19:03:07 hello 19:03:17 talk submission deadline for the paris summit has passed 19:03:32 so if you were planning on submitting something to the conference section, I hope you got it in 19:03:47 and you should go ahead and book travel and hotels 19:04:04 the call for topics for the design sessions hasn't yet opened 19:04:21 that should happen in the weeks ahead (I have no idea when, actually) 19:04:42 those are for the dev-oriented design sessions, not the "give a formal talk" conference sessions 19:05:16 any questions on the summit? 19:05:59 ok, moving on 19:06:21 next week (next tuesday) at the TC meeting, I'll be going over the "Swift gap analysis" 19:06:45 the TC has formalized the requirements they have for new projects to be incubated or moved to integrated projects 19:06:47 lots of sea-green blue text 19:07:04 what does it mean? 19:07:11 and so now they are going back to the existing projects and measuring them against the new rules 19:07:28 #link https://etherpad.openstack.org/p/swift_gap_analysis 19:07:32 portante: there. colors removed 19:07:39 heh, thought it was just me :) 19:07:59 so this is mostly a formality, but they want to see where there are gaps, if any, and then have a plan to resolve them 19:08:02 so where are the gaps? 19:08:09 I've given short answers on that etherpad 19:08:28 well that's for the TC to decide :-) 19:08:34 Mr. PTL, we must not allow a mineshaft gap! 19:08:34 ahh 19:08:36 lots of dones... that looks good :) 19:08:38 but I can guess at a few they may or may not hit 19:08:54 do you mean minecraft gap? 19:09:09 first, we'll probably have to come up with a mission statement 19:09:30 easy: make the world a better place 19:09:32 second I don't know if they'll comment on us not doing the mid-cycle milestones 19:09:33 Mission: To be as different as possible from the rest of openstack >:) 19:09:43 creiht_: you're not helping ;-) 19:10:04 and now we're being logged to :) 19:10:07 too 19:10:10 so they want to put the kibash on mid-cycle releases? 19:10:11 oy 19:10:15 portante: no 19:10:18 portante: well, I don't know 19:10:18 good 19:10:26 To boldly store data where nobody has stored data before. 19:10:34 portante: I wouldn't be surprised if it comes up. that's all I'm saying 19:10:47 notmyname: okay 19:11:02 so, overall, I think we're actually in really good shape 19:11:11 looks that way 19:11:19 pretty much everything is "done" or "N/A" 19:11:47 is there anything we can help with re gap analysis? 19:12:13 mattoliverau: look over the doc. see if there is anything you think is incorrect 19:12:13 no issues you think w/lack of oslo stuff? 19:12:27 did we check with apple marketting for the name ? ;) 19:12:48 Reading.. that I can do :) 19:12:49 I saw that evidently Marconi has to change it name 19:12:53 peluse_: I don't know what will come up wrt oslo. I think we do use oslo where appropriate :-) 19:13:08 I figured that would be your response :) 19:13:12 oh and that is a more helpful comment that creiht's? 19:13:16 ;) 19:13:20 lol 19:13:24 if we have to change names, we should call it Zoidberg 19:13:25 haha 19:13:30 lol 19:13:34 notmystorage 19:13:39 +1 19:13:47 torgomatic: because why not? 19:13:51 :) 19:14:40 I think the "oslo question" has always been one of "no copy/paste code externally managed added to swift" and it needs to provide benefit beyond what exists 19:15:11 any other questions around that? the TC meeting is a 1pm pacific next tuesday 19:15:26 * portante is glad you are on top of that one 19:15:35 sounds early 19:15:53 mattoliverau: ya, but it's normal TC meeting time 19:16:21 ok, other housekeeping then 19:17:13 looking at time between now and the integrated juno release, we should probably look at what's landed and cut a release in a few weeks. that will give us time to do another for the integrated juno release 19:17:57 I'll be looking at what's landed in the next few weeks and we'll take it from there 19:18:31 I don't have any questions. more of an FYI 19:19:05 anything else before we move to ec status and patches? 19:19:47 ok, let's move on 19:19:51 #topic EC status 19:20:02 couple of things.... 19:20:04 peluse_ and I have talked about what's going in here 19:20:11 (1) EC end user docs up for review (2) EC spec up for review (3) tsg has EC policy patch up for review 19:20:18 currently working on updating the EC spec and trello to add some traceability/visibility to remaining items 19:20:26 great 19:20:33 ..and small meet-up next week to review/update design (cover open questions) and wrap up any EC policy patch stuff 19:20:43 notmyname: 2 more patches in the series coming up 19:20:46 following that I'd look for another rev of the spec and maybe a few more patches 19:20:50 tsg: nice 19:21:06 peluse_: I thought the docs had landed 19:21:23 notmyname: yup, I just mewan they're still in need of eyes but yes they're on master 19:21:28 shit, I mean feature/ec 19:21:28 ah, yes 19:21:34 right :-) 19:21:41 I mean "shoot" 19:22:00 #link https://review.openstack.org/#/c/106103/ 19:22:01 peluse_: nice recovery in a logged channel :P 19:22:06 #link https://trello.com/b/LlvIFIQs/swift-erasure-codes 19:22:53 yeah, looked for numbered items in Trello that match up with sections in the design doc, coming soon.... 19:23:14 peluse_ and I are working on getting the spec and trello in-line with one another and actually with the things that are outstanding to do 19:23:59 but the general idea, as I'd like to see it, is to get the basic read/write path landed and then do the background daemons (similar to what happened with storage policies) 19:24:15 yup, tsg any quick status update on PyECLib? 19:24:30 and tsg and kevin and others have been doing great work on pyeclib, but there's still a ton to do in swift itself 19:24:45 maybe 2 tons... 19:25:03 the big ton was PyECLib refactor into PyECLib + liberasurecode 19:25:19 we have liberasurecode API stable now and 3 EC backends enabled 19:25:24 nice 19:25:37 the last remaining piece is making PyECLib use the new liberasurecode API 19:25:51 ETA this week on that 19:25:57 tsg: great 19:26:20 I'm hoping that will allow some dev time to move from the ec library to swift itself :-) 19:26:22 notmyname: most of this work is being done on the "split-libec" branch of PyECLib 19:26:33 * tsg nods 19:27:07 notmyname: absolutely, we can focus most of the time next week on Swift EC datapath 19:27:17 also, I'm hoping that as the outstanding tasks are clarified, more people can jump in and work on the implementation 19:27:51 notmyname: Kevin and I will move the "done" cards out of the way on Trello so the outstanding tasks are clearer 19:27:56 so expect to hear more from me bugging everyone about ec implementation work :-) 19:28:00 tsg: thanks 19:28:01 sidenote: looks like voting just opened for Paris presentations 19:28:11 peluse_: ah, thanks 19:28:18 peluse_: can you #link it? 19:28:47 one sec, on my other systesm 19:28:51 I got it 19:28:55 #link http://www.openstack.org/vote-paris 19:29:02 gracias 19:29:15 I mean "merci" :) 19:29:23 any other questions or things to bring up about EC? 19:29:27 from anyone? 19:30:03 ok, let's move on to patches 19:30:10 #topic patches 19:30:34 mattoliverau: do you still have that shortlink for your dashboard additions? 19:30:50 hah the first one that comes up for me is http://www.openstack.org/vote-paris/Presentation/api-battle-s3-vs-swift-vs-cdmi-can-t-we-speak-the-same-language 19:31:05 I'll dig one up.. or make a new one :) 19:31:29 mattoliverau: thanks. actually, can you add it to the priority reviews page? 19:31:42 creiht: CDMI: http://xkcd.com/927/ 19:31:51 indeed 19:31:53 sure 19:32:07 heh 19:32:12 I looked through all the "Need Reverify" patches on the dashboard this morning 19:32:28 there were only one or two that actually needed a reverufy 19:32:33 most have rebase issues 19:32:41 so if one is yours, etiher kill it or rebase it 19:32:43 please 19:34:04 other than that, there hasn't been a lot of movement on the reviews :-( 19:34:13 so what are we going to do? 19:34:42 notmyname: i’m back from vacation, will have a look at the keystone stuff 19:34:50 cschwede: awesome, thanks 19:35:24 I've got 2 to look at (that I haven't yet). the migration middleware and the object sysmeta one 19:35:25 #link http://bit.ly/UHcm1d 19:35:35 Will add it to the wiki now 19:35:42 mattoliverau: thanks 19:36:09 notmyname: mea culpa 19:36:11 the difference with mattoliverau's dashboard link is that it has a "possible abandoned changes" section 19:36:50 related migration: I am working now on another patch, will contain swift-swift and file system - swift. S3 to swift will be different patch. Alistair provided some comments that i am going to implement in the next patch. 19:37:05 gvernik: ah, great. thanks 19:38:34 I think it would be good for each of us to set aside some "review time" over the next week to get some reviews done 19:39:26 I don't want to move to a "review schedule" system yet :-) 19:39:35 so, that being said...patches 19:40:15 we went over the ones on the priority reviews page last week. anything to add there? anyone need some horsetrading for reviews? 19:43:01 * torgomatic is just up to his neck in non-Swift code 19:43:08 ok. 19:43:18 * portante wiggles ear in agreement 19:43:55 ya. it seems that a lot of us got pulled back into internal things after the storage policies push 19:44:43 I'll block off some time to do some reviews. looks like I can do that on friday 19:45:17 try to do the same, and let's see where we are next week 19:45:17 * portante on vacation EOD today thru August 6th, back on the 7th 19:45:28 portante: have a great tiem :-) 19:45:35 portante: so you'll have plenty of time to review :P 19:45:39 lol 19:45:42 cschwede: I've been looking at the keystone stuff, could use a little help if you're around later... 19:45:45 yes, and maybe ... ;) 19:45:56 #topic open discussion 19:46:00 anything else? 19:46:53 notmyname: what about this Swift hackaton that you mentioned two weeks ago? 19:47:04 gvernik: ah, yes 19:47:32 gvernik: we're putting it together right now. it will be on the east coast (boston). red hat is sponsoring it (thanks!) 19:47:48 and in a week or so we'll open up the registration 19:47:57 we are hosting it at our Westford, MA engineering facility 19:48:02 sweet 19:48:16 do we have dates yet? 19:48:28 29th - 2nd sep -oct? 19:48:34 yep 19:49:03 29 sept - 2 oct 19:49:06 sept 29th - oct 2nd 19:49:17 which on is it? ;) 19:49:49 portante: it's like your eyes and hands are running in different threads. with no lock 19:50:08 gvernik: thanks for bringing it up 19:50:20 may be i will come this time :) 19:50:30 * portante tried to make a joke and failed 19:50:52 anything else to bring up? 19:50:57 I have an etherpad open for an email template to send owners of patches whose changes are in danger of being abandoned. Thanks for those who have taken a look, and those who will now that I've mentioned it: https://etherpad.openstack.org/p/abandoner_email 19:50:59 * peluse_ was laughing on the inside 19:51:11 mattoliverau: +1 19:51:48 So if you can all take a look that would be helpful :) 19:52:42 thanks for coming, everyone 19:52:44 #endmeeting