17:04:06 #startmeeting XenAPI 17:04:07 Meeting started Wed Feb 20 17:04:06 2013 UTC. The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:04:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:04:10 The meeting name has been set to 'xenapi' 17:04:23 #topic agenda 17:04:33 #link https://wiki.openstack.org/wiki/Meetings/XenAPI 17:04:41 the wiki upgrade changed the link 17:04:49 any more things people wanted to add? 17:05:08 There are a couple of things we can talk about in AOB but nothing to the agenda 17:05:24 Ok, cool, just want to make sure we leave time for them 17:05:32 #topic blueprints 17:05:53 well its the last day, I don't see anything pending, except the OVS quantum related stuff 17:06:12 #link https://review.openstack.org/#/c/15022/ 17:06:13 We do have one more thing that we're trying to get in 17:06:20 what is that? 17:06:27 a bug fix from Mate - he's just joining 17:06:31 Hi 17:06:45 Mate - was just saying that there's one more thing we're trying to get in grizzly-3 17:06:55 Ok, well that will have to wait till tomorrow probably 17:07:15 critial features only today I think they said 17:07:35 It's very close to being ready for review - hopefully will be pushed today. 17:07:36 it'll be fine, I have a bug filed. 17:08:02 cool 17:08:09 It's the generic case of copy images to volume. 17:08:17 #topic docs 17:08:36 now we are past feature freeze I gues it is docs time 17:09:03 you might want to join in with other Xen people 17:09:06 #link http://wiki.xen.org/wiki/Xen_Document_Days 17:09:06 Who wants to do what? 17:09:28 well each do your own feature first seems like a good way forward? 17:09:38 then see where we are later? 17:09:55 Okay, I am planning to deal with the XenAPINFS Stuff. 17:10:00 does anyone fancy doing a full pass of the docs to see what we are missing 17:10:01 cool 17:10:09 Does the other features require docs? 17:10:25 configdrive? 17:10:36 We're planning to start looking at docs v.soon, yes - but unfortunately can't make the Xen document day 17:10:41 #action matelakat to ensure docs are done for XenAPINFS 17:10:46 Both Mate and myself are involved in a long meeting on Monday 17:10:49 OK 17:11:28 matelakat: i think just need to add it works on XenAPI , and document (find) any limitations 17:12:01 I can ping michael still and see if it is complete 17:12:08 I would like to try out that configdrive at some point. 17:12:18 cool 17:12:38 OK any more for any more? 17:12:50 install guides are a big missing bit 17:13:03 and testing them, of course 17:13:22 Installing what sort of setup are you thinking of? 17:13:49 these ones: #link http://docs.openstack.org/install/ 17:14:04 they don't cover XenAPI 17:14:31 We have a proper lab setup here 17:14:33 got it 17:14:57 And hopefully we could amend the docs while trying them out. 17:15:06 sounds good 17:15:12 ping me on IRC if I can help 17:15:16 with questions 17:15:19 :-) 17:15:54 OK, so open discussion time 17:16:04 #topic OpenDiscussion 17:16:22 Me first! 17:16:35 OK 17:16:39 I'd like your input john on https://review.openstack.org/#/c/22441/ 17:16:52 Tried to add you as a reviewer, but gerrit doesn't like your email any more 17:17:04 sure, you can just add my name, and I should help out 17:17:05 ah 17:17:23 hmm 17:17:36 I've tried a few times but it keeps rejecting you saying your email is invalid :D 17:17:37 it doesn't like me adding myself either 17:17:41 even though it auto completed it! 17:18:04 just add "johngarbutt" 17:18:14 not sure why it dosn't like my email 17:18:54 Ah - that added you 17:18:55 ta. 17:18:57 anyway, looks about right. 17:19:09 copyright notice and stuff 17:19:16 I will take a look tomorrow 17:19:28 OK, me next 17:19:42 I have added a summit talk for the XenAPI road map 17:19:53 to help the discussion I have opened an etherpad 17:20:03 #link https://etherpad.openstack.org/HavanaXenAPIRoadmap 17:20:10 #link http://summit.openstack.org/cfp/details/4 17:20:30 Looks good 17:20:36 There are some boiler plate bits, but please contribute 17:20:39 I'll try and update it with my thoughts 17:20:43 cool 17:20:51 I can add mine as well. 17:20:57 feel free to ping me on IRC when you add stuff, not sure I can subscribe 17:20:58 Having a proper Xapi layer in oslo. 17:21:20 it is a nova discussion really, that would be a seprate oslo session really 17:21:43 the can collapse the ones that will not take a whole session 17:21:54 but we can talk about including it in the nova session 17:22:03 What is AOB? 17:22:11 any other bussiness 17:22:19 silly english term I guess 17:22:31 Any Other Business 17:22:35 oh duh 17:22:43 Okay. 17:22:50 didn't see the reply 17:22:55 no worries 17:22:56 I think I need to visit the opticians 17:23:02 So I was busy with this cinder stuff, and I found around 3-4 bugs. 17:23:02 I went last week 17:23:07 apparently I can see 17:23:25 So that's good for Cinder. 17:23:27 Ah, 17:23:46 We were talking about using qemu-img for image conversion inside nova 17:23:53 cool, do we have tempest tests runninon your new code 17:24:10 have you spoken to comstud about that? 17:24:17 matelakat: what'd you find by the way? 17:24:20 I don't have tempest test, I have separate exercise-like tests. 17:25:01 guitarzan: see this link: https://review.openstack.org/#/c/20514/ 17:25:05 matelakat: regular tests should test the new driver OK 17:25:35 john: I haven't spoken to comstud 17:25:58 I 'am a bit confused. 17:25:59 he was preping something here: https://github.com/comstud/python-libvhd 17:26:11 matelakat: gotcha! I was expecting xen cinder stuff :) 17:26:41 guitarzan: The nightmare happened in the manager level... 17:26:51 guitarzan: sorry for that. 17:27:00 john, this https://github.com/comstud/python-libvhd is a bit old 17:27:05 no worries, I'm vested in both 17:27:23 And I don't see how it would solve the issue. 17:27:26 oops 17:27:29 wrong link 17:27:33 It is just a python binding. 17:27:41 I tried it once, but did not work. 17:27:49 #link https://github.com/comstud/vhd 17:28:06 WHat I am talking about: use qemu-img convert to convert the downloaded image. 17:28:23 yes, the plan was to extract vhd code from xen 17:28:30 So in this case, we could launch any type of images. 17:28:33 then link that inside qemu-img 17:28:50 agreed, we could go to raw then boot raw 17:28:57 I think, we are talking about 2 things. 17:29:10 so comstud's plan is to upstream the XS vhd format so qemu-img could convert directly to it? 17:29:18 yep 17:29:31 currently it creates broken vhd files 17:29:50 but we could create raw then boot from those, using the disk attach dd method 17:29:57 "broken" or just "not XenServer enhanced" ;) 17:30:03 dd method. 17:30:12 THat's what I was trying to describe. 17:30:17 I think they break when you take a snapshot, can't remember 17:30:20 replace dd with qemu-img 17:30:38 but what would qemu-img be used for? 17:30:55 johnthetubaguy, yeah - the XS VHD format has extensions that enable sensible snapshots (in particular you can't coalesce sensibly with the stock VHD format) 17:30:55 *->vhd? 17:31:11 BobBall: ah, that would be it then 17:31:21 I knew there were extensions 17:31:24 so I'm not surprised that problems start when you try snapshotting with the stock VHDs 17:31:25 I guess they are missing 17:31:29 If you want to start an instance from a qcow2 image, you could use qemu-img convert to convert it to raw, and pipe the raw bytes to the vdi. 17:31:53 right, OK, I guess that was what I was expecting to hear, cool 17:32:01 I think it helps 17:32:24 but the image in the qcow2 may well not run on Xen, but that is less likely these days, I guess 17:32:47 it would be nice to extened/use cloud-init to install the XenServer tools 17:33:20 config drives could contain the tools 17:33:41 Ok, sounds like we are done 17:33:58 Ok, see you 17:34:00 * BobBall is 17:34:14 #action matelakat to add inline qcow2 conversion to wiki 17:34:19 oops 17:34:22 missed him 17:34:28 and I meant etherpad 17:34:39 OK 17:34:42 thanks all 17:34:46 #endmeeting