19:01:14 #startmeeting 19:01:15 Meeting started Tue Jun 26 19:01:14 2012 UTC. The chair is mtaylor. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:16 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:02:07 hey everybody 19:02:19 anybody here? 19:02:22 o/ 19:02:27 yup 19:04:14 * mtaylor is fixing a boneheaded breakage at the moment ... Shrews, you wanna tell folks hat you've been up to? 19:04:32 sure 19:05:02 So, I've been looking at adding OpenVZ support (via libvirt) to Nova. 19:05:27 In the process, I've worked with Devananda to test/verify that his openvz patched kernel actually works 19:05:31 and it does 19:06:16 And it works with the libvirt utilities, too. I can create/destroy openvz containers with the command line tools 19:06:53 So I think I know what needs to change with the Nova configuration stuff, but I'm having difficulty actually *testing* changes, atm 19:07:05 hrm. chicken and egg problem? 19:07:21 yeah, kinda. 19:07:44 you need to be running on a system that has the openvz kernel installed ... and then you need to run devstack and have it configured to use openvz instead of qemu, yeah? 19:08:21 damn, my alarm didn't trigger for this meeting 19:08:24 i'm running the kernel, i have devstack installed but it doesn't support openvz 19:08:51 i *think* i can just modify /etc/nova/nova.cfg and restart devstack, but that's what i'm trying now 19:09:09 nova docs suck 19:09:13 devstack docs suck 19:09:14 Shrews: ah - well, adding support to devstack for openvz should also be on the list 19:09:26 because we'll need to be able to test that it works as part f CI 19:09:42 Shrews: see how I'm just making more work for you? 19:09:45 mtaylor: yeah, that's a natural thing to do 19:09:58 but i need nova to understand it first 19:10:20 so, that's where i'm at 19:10:26 ++ 19:10:27 19:10:33 stellar 19:10:57 clarkb: your week has involved much less head-bang-against-brick-wall I think 19:12:03 yes, however that may start again soon :) 19:12:22 * LinuxJedi gives clarkb a crash helmet 19:12:31 I have a working ACLed JavaMelody patch for Gerrit 2.4 19:12:57 basically you go gerrit.server/monitoring when you are logged in as a server administrator and you get awesomesauce 19:13:17 it's teh r0xr 19:13:45 currently working to push the patch upstream. It was suggested that I turn the patch into a plugin, which is in theory a great idea. 19:14:09 clarkb: and in actuality, much more work? 19:14:23 the Http plugin architecture forces your plugin under /plugins/plugin_name/* 19:14:37 that's groser 19:14:40 so I am having a hard time figuring out how I can filter /* for all of the statistical goodness 19:14:42 grosser 19:15:03 clarkb: it might be worthwhile responding to their suggestion with that 19:15:07 /plugins/melody/monitoring is working but it no longer knows about all of the http transactions happening under /plugins/melody 19:15:22 mtaylor: yup going to clean up the patch as suggested then comment on these issues 19:15:38 you know, in this case, because of what melody does, it might not actually be suitable for the plugin architecture 19:16:00 I am beginning to think that is the case. Either that or the plugin architecture needs to be more flexible 19:16:18 that being said - perhaps having knowledge about the http plugin architecture will end up being helpful in the future? 19:16:32 you can add a core plugin which in theory gives you access to all of gerrit but I haven't had luck getting that to work either 19:17:10 mtaylor: definitely I think the plugin stuff will be super useful. Add mod rewrite and your users don't even need to know they are accessing /plugins/foo/stuff 19:17:18 ++ 19:18:29 Upstream does seem interested in this though, so hopefully we can work something out that everyone is happy with 19:18:42 I hope so. I think it's going to be superhelpful to us too 19:19:00 we should be about ready to roll out a new gerrit to production soon, no? 19:19:11 LinuxJedi: are your changes for the theming in trunk? 19:19:20 mtaylor: perhaps we should grab 2.4.2? 19:19:31 clarkb: probably a good idea 19:19:34 mtaylor: in openstack/2.4.1 19:19:41 mtaylor: our branch I mean 19:19:44 LinuxJedi: that's what I meant :) 19:19:54 lemme make an openstack/2.4.2 real quick 19:19:55 mtaylor: yes, and I just pushed up a review to put it live 19:21:07 ok. openstack/2.4.2 now exists 19:21:23 in case anybody wants to rebase and propose all the changes 19:22:09 if no one gets to it first I am sure that will be a good break from this plugin 19:22:14 :) 19:22:37 before I forget. If I do somehow manage to get the plugin working, would we want to host that or try to get Gerrit to own it? 19:22:59 I'd love it if we could get gerrit to own it ... but if we need to host it we can do that too 19:23:07 I take it plugins can happily live in their own repos? 19:23:27 yes 19:23:49 hrm. I wonder if our WIP patch could be re-written as a core plugin :) 19:23:58 * mtaylor likes torturing Shrews 19:24:02 NO 19:24:04 haha :) 19:24:09 so - from my end - I've just about got all of the groundwork done for uploading client libs to pypi and the patches are merging into the client lib branches 19:24:28 hopefully by this week we should be able to have that working automatically on tags 19:24:37 and also hopefully soon we'll have doc changes gating 19:24:58 jeblair is down in Argentina telling people about all the crazy things we do 19:25:07 LinuxJedi: you been doing anything useful with yourself? 19:25:15 mtaylor: a couple of things 19:25:33 so, first we had Jenkins break a little last week 19:25:54 this was due to Jenkins Job filler giving a slightly incorrect but valid XML 19:26:09 the other is Stackforge 19:26:40 we are planning to move Stackforge Gerrit and Jenkins into Openstack Gerrit and Jenkins now that setting up Jenkins stuff is semi-automated 19:26:54 the branches on github will still be separate though 19:27:11 this solves many problems for us and means we have less machines to maintain 19:27:49 it will also make things more reliable too 19:27:58 and that is about if from me 19:28:49 cool 19:28:53 anybody else got anything for us? 19:30:26 I'm going to call that good then. thanks folks! 19:30:30 #endmeeting