18:01:12 #startmeeting 18:01:13 Meeting started Thu May 17 18:01:12 2012 UTC. The chair is jgriffith. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:01:28 volume folks around today? 18:01:36 I'm here 18:01:51 rnirmal: howdy! 18:02:22 hey jgriffith ... hope it's not another silent meeting :) 18:02:25 rnimral: can you take a look at https://review.openstack.org/#/c/7537/ when you get a chance? 18:02:44 rnirmal: me too 18:03:07 Folks are typically a few minutes late, or jump in when they see their name mentioned :) 18:03:10 jgriffith: I looked at it... and fell I'll have to run it to get a gauge, that's a lot of chagnes... great job putting it together 18:03:43 rnirmal: Yeah, it's a big set, but the good thing is it's realy just a copy of novaclient 18:04:02 jgriffith: that's good, yeah I'll look thru it. 18:04:06 rnimral: anotherjesse made a bunch of changes to strip extras out and get it running on his cloud 18:04:11 do you know if this will get merged in soon https://review.openstack.org/#/c/7406/ 18:04:52 rnirmal: Just waiting on an approve... changes to transfer to cinder I don't believe 18:05:06 changes *don't* transfer to cinder 18:05:22 jgriffith: ok just checking... atleast it affects us in nova right now 18:05:31 and would like to get in soon 18:05:40 rnirmal: Yeah, it's out of my hands now :) 18:05:50 I'll ask vishy if it's pending other +1's 18:06:40 jgriffith: do you just want to talk about cinder client and where we are with it. others can atleast look at the meeting mins. 18:07:02 rnirmal: yep, I was going to run through cinder status updates real quick 18:07:25 #topic cinder status 18:07:35 oooops, hi all 18:07:39 Yo 18:07:51 So the big news right now is python-cinderclient 18:08:02 It's working, unit tests passing etc 18:08:31 anotherjesse tweaked an os setup yesterday and plugged it into horizon 18:08:41 was able to test it against nova-volume 18:08:46 create/delete etc 18:08:50 Looks good 18:09:17 dtroyer is working on devstack integration... he had one submittal for it already 18:09:45 I tried it out this morning and hit a couple of issues, but there are some changes to what we're going to do and I believe he's working on them 18:10:01 The idea is to be able to use cinderclient to do all of the nova-volume stuff 18:10:14 This way we have something to test against 18:10:14 cool... so is the plan to remove nova volume-* from novaclient. or keep it till the next release 18:10:38 and switch to cinderclient for all volume purposes 18:10:49 Depending on how things go I could see it out for Folsom and switched to cinder 18:11:01 But we'll have to prove things out pretty well before that happens 18:11:20 Or we may do like quantum and have both, don't know where those discussions ended but we'll have them again 18:11:32 ok 18:12:03 So we have a working client, we're getting things set up now for devstack to let us actually use cinder or nova-volume 18:12:50 There are some new blueprints to look at: https://blueprints.launchpad.net/cinder 18:13:17 If you pull up: #link https://blueprints.launchpad.net/cinder/+spec/python-cinder-client 18:13:42 you cna see jesse put together a dependency graph that you can click on for the various pieces 18:14:20 cool that looks good. we have a set of blueprints to move forward on 18:14:37 are we going to start targeting them for milestones? 18:14:56 Yes, I started that already I believe 18:15:01 cinderclient is f1 18:15:23 I believe extract-nova-volumes is f2 18:15:34 or it "is" 18:16:09 I think the others are sort of partials in f1 and targetted for f2 18:16:48 I think that's the quick low down... 18:17:07 anything people are wondering about specifically or might know that I missed? 18:17:47 nothing that I can think of for the immediate short term 18:17:51 The next thing I was going to start looking at was implementing the ability to specify cinder in nova 18:17:56 ie alternate volume driver 18:18:09 for a while we want to keep both obviously 18:18:29 jgriffith: hmmm alternate volume manager or volume driver? 18:18:38 sorry...neither 18:18:40 API :) 18:18:55 plugin like glane and quantum do... 18:19:15 or "configurable" I guess is the right way to say it 18:19:33 yeah... should be easy just switching endpoints 18:20:01 rnirmal: famous last words 18:20:16 :) 18:21:06 don't think I have much else if there aren't any items other folks want to talk about? 18:21:14 jgriffith: do you want to briefly discuss keeping things in sync between nova-volumes and cinder 18:21:23 rnirmal: sure 18:21:28 back porting or forward porting 18:21:40 So we're comitted to staying in sync obviously 18:21:56 I don't know of any big changes/bp's coming to worry about 18:22:10 that's one thing I think we kinda decided upon 18:22:20 no new features in nova-volumes.. they'll just go into cinder 18:22:20 at the stage cinder is at right now it shouldn't be a major issue 18:22:28 but it's mostly about bugs and minor enhancements 18:22:33 rnirmal: ahhh... well.... 18:22:47 I tihnk if a feature is proposed it has to be addressed on a case by case 18:23:03 cinder isn't in and core yet so nova-volume has to continue IMHO 18:23:30 +1 on that. but how do we keep track of the changes 18:23:40 In most cases it might mean it's implemented in both places like cpq16's work 18:23:45 I guess we'll just have to follow and volume changes 18:23:48 and/any 18:24:15 yeah I asked cp16 to get it into both 18:24:19 So I check gerrit constantly... also I would hope that nova core keeps transition in mind 18:24:26 so looks like we'll just have to be on the lookout 18:24:59 so there's one bug/enhancement I submitted to nova.. haven't done to cinder yet.. I'll do that soon 18:25:00 also although it may be lame I'm doing diffs between cinder and nova-v every once in a while 18:25:12 rnimral: which one? 18:25:45 jgriffith: https://bugs.launchpad.net/nova/+bug/992729 18:25:47 Launchpad bug 992729 in nova "Add 'host' name to initialize/terminate volume connection" [Undecided,Fix committed] 18:26:26 rnirmal: they can't just work off of the ip address? 18:26:53 it's easier to for managing with hostnames.. they can work off ip address 18:27:07 ok, thus enhancement versus bug :) 18:27:20 yeah.. not all drivers have to use it 18:27:44 Ya know I could solve this problem more easily for you..... ;) 18:28:08 jgriffith: yeah I hear you 18:28:16 LOL 18:28:35 alrighty... 18:28:41 so coming back to the point... how do we solve this... just make then as affecting both projects 18:28:48 make/mark them 18:29:03 Yeah, fill in the also affects project 18:29:11 That's what I did at least 18:29:51 Seems like the right way to go to me 18:30:08 ok cool that works. just added that one.. so I'll do that for any volume changed in nova I come across 18:30:16 excellent! 18:30:45 #action jgriffith mention nova/cinder relationship for changes at next project meeting 18:31:16 sorry I just joined, for cinder I'll look at the logs and try to be on time next week 18:31:16 I'll bring this up again at next project meeting, maybe if everybody on nova core is aware that will help 18:31:28 clayg: No worries... 18:31:52 clayg: If you had anything specific though feel free... 18:31:59 clayg: awesome...more participation is needed at this point 18:32:21 jgriffith: nothing specific, just trying to be more present 18:32:42 :) 18:33:17 Craig Vyvial's notification changes are the only thing I've really looked at this week. Been trying to get an answer off our billing people about whether we've had to add anything that isn't currently provided by Craig's patch. No definite answer yet 18:33:52 Yeah.. nice catch you and rnirmal on those reviews 18:34:27 DucnanT: keep us all in the loop if there ends up being more to come 18:34:38 again keeping nova-v and cinder in sync 18:34:38 DuncanT: that's good would like to get more input on that 18:35:10 we went with what we felt had the most info. but we would be missing some pieces in the notifications 18:35:15 DuncanT: Also if you get a chance https://review.openstack.org/#/c/7537/ 18:35:20 :) 18:35:45 I'd like to get that sucker pushed and close my personal repo ASAP 18:35:55 jgriffith: I'll review that in the morning if not before 18:36:04 thanks! 18:36:15 Oh yeah... what time is it there? 18:36:16 jgriffith: are you going to reset the openstack/python-cinderclient after that merge? 18:36:31 rnirmal: yes, that was the plan 18:36:42 or I'm going to ask mtaylor to do it 18:36:47 7:30 pm 18:37:18 DuncanT: That reminds me, I think I promised you opening discussions on a differnt meeting time? 18:37:28 I was going to ask if anybody minded making things an hour earlier, since most of us are on summer time now and the meeting stayed at the same time utc 18:38:43 DucanT: QA meets in that slot but we could try before that maybe? 18:38:54 I'm ok with 1 or 2 hrs earlier. need to make sure there's no other conflicting meetings 18:39:07 Earlier is certainly better for me 18:39:15 clayg: work for you? 18:40:13 yup! 18:40:32 Cool... I'll send an email out for the benefit of those that aren't here 18:40:42 5pm is metering, 6pm is QA, 7pm is our current slot, 9pm is orchestration 18:41:01 Ahh...metering did grab that slot? 18:41:32 According to the ICAL feed, yes 18:41:33 Oh, Thu!! I searched on Thur 18:41:54 hmph... that closes out Thursdays 18:43:30 Wednesday seems to be open during that time 18:43:48 I can do wednesdays 18:43:58 yes, is fine 18:44:21 ok... I'll send a note out 18:44:41 anything else? 18:44:52 anybody wanna write some cinder code? 18:44:53 :) 18:45:45 ok, well let me know via email or irc if something comes up or you have ideas/bandwidth 18:46:03 If there's nothing else I'm going to go eat lunch :) 18:46:25 Bit late for lunch for me, but the bar is open.... 18:46:35 DuncanT: that sounds better... 18:47:20 #action jgriffith email proposal to switch meeting to Wed 2 hours earlier 18:47:29 Alright, see ya guys 18:47:34 #endmeeting