15:00:22 #startmeeting XenAPI 15:00:23 Meeting started Wed Mar 18 15:00:22 2015 UTC and is due to finish in 60 minutes. The chair is BobBall. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:25 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:27 The meeting name has been set to 'xenapi' 15:00:34 * johnthetubaguy waves 15:00:37 Well hello there 15:00:48 * BobBall waves back 15:00:56 So - what have we got on the agenda today 15:01:04 #topic libvirt+xen 15:01:34 * johnthetubaguy wonders if its time to rename the meeting… 15:01:35 Not really XenAPI but since we're both involved + interested in these things I wonder if we should change the meeting to Xen based 15:01:40 to include XenAPI and libvirt+xen 15:01:53 I am cool with it been a general Xen meeting 15:02:12 It will, of course, change the location of the eavesdrop logs 15:02:24 So just be aware 15:02:27 yeah, ups and downs 15:02:43 I'll send an email to openstack-dev if we are going to change to Xen-based 15:02:49 Anyway 15:02:54 #link http://zuul.openstack.xenproject.org/ 15:03:00 Exciting things in the world of libvirt+xen 15:03:18 I've set up a CI, above, which is running in the Rackspace cloud 15:03:27 cools 15:03:36 http://jenkins.openstack.xenproject.org:8080 is the jenkins 15:03:39 (we have very original names 15:03:40 so nested Xen+libvirt in XenServer? 15:03:41 +) 15:03:59 Well rackspace is running XenServer yes 15:04:04 then we have an Ubuntu HVM guest 15:04:09 and we're running libvirt+xen inside that 15:04:17 right, cools, just checking 15:04:29 i.e. the CI is testing libvirt+xen but the infrastructure is virtualised on XenServer 15:04:43 http://jenkins.openstack.xenproject.org:8080/job/dsvm-tempest-xen/buildTimeTrend 15:04:59 how are the neutron races going? 15:05:05 As you can see from the build trend, which is not the prettiest of graphs, we still have some work to go on the pass rate 15:05:17 Currently we're passing around 50%ish 15:05:24 wait until we get to that item johnthetubaguy :) 15:05:39 There is a request out on xen-devel for those who want to know more 15:05:49 so libvirt+xen is doing neutron right now? 15:06:02 no - default devstack setup which is nova-network 15:06:07 OK 15:06:15 I believe 15:06:19 certainly still using the default 15:06:25 which I am pretty sure is still nova-network 15:06:29 that might not be the best choice, long term, but OK 15:06:36 you have to start somewhere 15:06:48 but because it's libvirt we don't anticipate any issues with it running in neutron 15:06:55 since architecturarlly it's the same 15:07:05 Might be the odd missing function from libvirt's xen support 15:07:10 but I don't have the details there 15:07:11 you might find it works better, its hard to know I guess 15:07:25 These issues aren't networking related 15:07:28 or at least do not seem to b 15:07:30 +e 15:07:41 http://osdir.com/ml/general/2015-03/msg12714.html 15:07:47 That's the call for interest 15:07:57 If you want to join a walk through of the CI reply there or offline 15:08:18 Any questions on libvirt+xen? 15:08:36 nothing more at this point 15:08:45 leaving this mostly with ant 15:08:49 #topic XenAPI CI 15:09:05 I don't think there is anything to raise here except we're chugging along nicely ATM 15:09:20 always a good sign 15:09:25 We're still nova-network and 6.2 based, but I've not had any time to change either of those things 15:09:43 yes, that could get tricky soon 15:09:53 Soon when? 15:10:03 I suspect liberty might be when you need neutron tests to be class B, if I were a betting man 15:10:06 Are we yet at the stage where nova-network might be deprecated in L? 15:10:22 Oh yes, that's well within our plan 15:10:38 Of course, the biggest issue with neutron support is that we suspect recent changesets have broken the XenAPI integration 15:10:42 in reality, its looking dodgy of L, its possible though 15:10:49 but without a CI it's not reasonable to ask the cores to prevent breakages 15:10:59 agreed 15:11:07 OK - well then we should still be in plenty of time then 15:11:18 We can move on to... 15:11:25 #topic Recruiting 15:11:35 So - Citrix is recruiting OpenStack developers in Nanjing, China 15:12:11 http://jobs.citrix.com/job/Nanjing-Sr_Software-Development-Engineer%28OpenStack%29/253527100/ 15:12:42 One of the first items this team will be doing is working on the Neutron support 15:12:54 So if anyone wants to move to Nanjing... :) 15:13:15 Or, more likely, if you know anyone in Nanjing looking for a new opportunity 15:13:20 Worth applying. 15:13:22 * johnthetubaguy looks other direction 15:13:40 its quite a log commute 15:14:19 Hopefully we'll have a team ready to start even before the summit - in which case we might be able to fly a few of them over to Vancouver, but we'll see how that goes. 15:14:51 Oh - I forgot to mention on libvirt+xen CI... there's a change we desperately need 15:15:10 ah, interesting... 15:15:10 https://review.openstack.org/#/c/159106/ 15:15:37 libvirt+xen support was broken by a change a few weeks ago which dependended on a libvirt function the xen driver does not provide 15:15:50 The above change correctly works around this missing function 15:16:27 ah… doh 15:16:43 yes so currently I've got our zuul merging that for every change 15:16:48 * BobBall loves cross-project dependencies 15:17:03 But I think that's all I've got for today 15:17:06 #topic AOB 15:17:10 Do you have anything johnthetubaguy? 15:17:55 BobBall: I just targeted that on RC1 15:18:17 Ah perfect - thanks 15:18:30 Hoepfully that'll mean that Kilo doesn't ship with broken libvirt+xen support 15:19:05 Are we done? 15:19:26 I have one final point - I'm on vacation in a fortnight, so I suggest the next meeting is in 1 months time 15:19:40 Specifically on the 15th April 15:20:29 Any objections? 15:21:13 OK 15:21:17 all good 15:21:22 #info Next Meeting 15th April 15:21:27 sorry, got pulled into another channel for a sec 15:21:33 catch you later! 15:21:35 I think we're done then 15:21:44 #endmeeting