17:06:00 #startmeeting VMwareAPI 17:06:01 Meeting started Wed Oct 16 17:06:00 2013 UTC and is due to finish in 60 minutes. The chair is tjones. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:06:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:06:03 I'm here to listen 17:06:06 The meeting name has been set to 'vmwareapi' 17:06:09 hi dan 17:06:23 anyone else lurking about? 17:06:55 ok looks like this will be a short meeting ;-) Lets get started 17:07:04 :) 17:07:38 * dims lurking 17:07:59 1st i'd like to mention that i've submitted a proposal for the icehouse design summit for vmware roadmap. The link is here http://summit.openstack.org/cfp/details/247 17:08:03 hi dims 17:08:28 Thanks for doing that Tracy 17:08:46 feel free to take a look at this and at the etherpad that gives a little more info. The deadline for proposals is 10/17 BTW 17:08:56 I am interested in the improvement for live migration 17:09:54 AFAIK live migrations cann't works with latest Havana + VCenter 17:09:55 great yaguang - we can use some help in getting more detailed info on this 17:10:14 garyk mentioned that as well. 17:10:21 a bug report is here, https://bugs.launchpad.net/nova/+bug/1192192 17:10:23 Launchpad bug 1192192 in nova "Nova initiated Live Migration regression for vmware VCDriver" [Medium,Confirmed] 17:10:44 *looking* 17:10:48 tjones, i would love some programmable pluggability to select the datastore and compute node (specific ESXi inside vCenter) for deployment and migration 17:12:10 dims - what is the use case you are thinking of here? 17:13:36 say someone wants to evacuate a ESXi host for maintenance, i'd like them to be able to use nova cli to move vm(s) to other compute nodes 17:13:58 yaguang - yes that is the bug for live migration. Can you help with a BP on this? 17:14:12 tjones, sure 17:14:47 dims: if we exposed one nova compute per ESX host (like we should be doing, IMHO) then this would be natural, right? 17:15:13 dims: i.e., you can do this on kvm/xen because they use one compute per hypervisor, as nova was intended 17:15:33 dims: when you put a host in maint mode - VC should move the VMs already though - am i missing something? 17:15:53 dansmith, right. just like (http://docs.openstack.org/grizzly/openstack-compute/admin/content/specify-host-to-boot-instances-on.html) 17:16:11 dims: exactly, which is why I don't like the hide-a-cluster-behind-a-compute model :) 17:16:32 dansmith - hence your other question earlier to me ;-) 17:16:46 tjones: yes, actually, although I was going somewhere else with that :) 17:16:56 tjones, will draw pictures when we get near a board :) 17:17:16 dims - ok that's in HK then :-D 17:17:46 dims: please go ahead and add that to the etherpad and we can discuss further 17:18:01 thanks tjones 17:18:03 https://etherpad.openstack.org/p/T4tQMQf5uS 17:18:22 ok any other topics on the design summit? 17:19:21 I think there are also some topics around Glance/image handling 17:20:05 danflorea: should we add that to our etherpad as well ? 17:20:39 and if so - will you? 17:20:41 I would think that there should be a separate etherpad for them 17:20:50 http://summit.openstack.org/cfp/details/236 17:20:58 http://summit.openstack.org/cfp/details/235 17:21:17 If we want a combined etherpad I guess that works too? 17:22:13 are we covering that in our nova discussion or will there be a separate glance topic? Perhaps it would be helpful to just link to that. I don't want to get too much and i also don't want to miss anything. We need to prioritize 17:23:02 They should be separate glance topics 17:23:38 ok good - anything else on the design summit? 17:24:23 hi, just checked in 17:24:36 hi viul: 17:24:50 Ok - anyone have a BP for discussion today? 17:26:47 for this week - lets try to flesh out our ideas in the etherpad for the design summit and we can discuss them next week. Moving on to bugs. 17:26:53 #topic bugs 17:28:47 I don't see anything critical here 17:29:18 I just filed one yesterday about sparse disk spawn issue. 17:29:33 https://bugs.launchpad.net/nova/+bug/1240373 17:29:34 Launchpad bug 1240373 in nova "VMware: Sparse glance vmdk's size property is mistaken for capacity" [High,New] 17:30:55 if glance lies about the image's capacity, we will incorrect attempt to extend a disk to flavor size when we can't 17:31:31 The are reasonable workarounds 17:31:35 viul: nice ;-) 17:32:06 but there should be something we can do on driver side to better handle this as well. 17:32:23 ok im going to confirm that one 17:32:36 was going to suggest just that. 17:32:47 any bugs we want to review before triaging the new bugs? 17:33:05 i see that we still have quite a few that need +1 reviews 17:33:55 http://paste.openstack.org/show/48574/ 17:34:37 nothing is waiting for core - so this is on us to get ready for them. Please spend some time this week reviewing. 17:34:53 Several need an update too 17:35:15 shall i move on to bug triage? 17:35:27 sure 17:35:37 #topic new bugs 17:36:38 http://goo.gl/5emePr 17:37:18 #link https://bugs.launchpad.net/nova/+bug/1237795 17:37:19 Launchpad bug 1237795 in nova "VMware: restarting nova compute reports invalid instances" [High,New] 17:38:00 this one was opened by garyk (who is in barcelona today) 17:38:11 he said it was on grizzly and havana both 17:38:13 wanted to find out more from Gary on this one 17:38:38 ok i'll spend some time looking at this today 17:38:55 #link https://bugs.launchpad.net/nova/+bug/1237795 17:38:57 Launchpad bug 1237795 in nova "VMware: restarting nova compute reports invalid instances" [High,New] 17:40:54 i'm assuming his cluster is containing both VC created and OS created VMs. Is that a valid use case? I mean - of course you can do it - but is it a likely senario? 17:41:20 Yeah I think it should be 17:41:43 I see the review posted fixes the instance list reporting 17:41:55 just wondering if there is more to the bug. 17:41:55 I think this would completely break HA 17:42:16 hee hee - im looking at the wrong bug 17:42:53 #link https://bugs.launchpad.net/nova/+bug/1240200 17:42:54 how so, dan? 17:42:55 Launchpad bug 1240200 in nova "VMware ESX driver reports incorrect metrics for memory, disk and cpu usage" [High,New] 17:43:07 #undo 17:43:08 Removing item from minutes: 17:43:39 sorry - moved on too fast. Dan - you have a comment about https://bugs.launchpad.net/nova/+bug/1237795 17:43:40 Launchpad bug 1237795 in nova "VMware: restarting nova compute reports invalid instances" [High,New] 17:44:25 Yes. HA may require us to restart the controller, which includes the nova-compute agent. If that doesn't work correctly, HA will not work. 17:46:32 lets discuss that more in the bug report - i want to get through the other bugs and we are running out of time. 17:46:56 next is #link https://bugs.launchpad.net/nova/+bug/1240200 17:46:57 Launchpad bug 1240200 in nova "VMware ESX driver reports incorrect metrics for memory, disk and cpu usage" [High,New] 17:48:28 The user see s the stats from the hypervisor page, which will report the usage stats from the instances perspective. 17:48:45 the audit logs should report the hypervisor s view of stats 17:48:57 we need to get that information to check if it's right. 17:50:03 ok do you want to confirm that one if its not correct? The reporter is working on it - but it's not confirmed 17:50:39 I will update the bug, it's not a bug with the driver. 17:50:59 next 2 are still waiting for updates from the submitter 17:51:36 last one was reported by our QE guy - sabari i think you were looking at it ? #link https://bugs.launchpad.net/nova/+bug/1240292 17:51:37 Launchpad bug 1240292 in nova "vmware: Error in logging causing vcdriver report 0 resources" [Undecided,New] 17:52:48 Yes, it s an intermittent issue. Mostly a suds error (need to check). Has happened once during one of the periodic reporting 17:53:24 lovely. 17:53:46 we have a little time left - #topic open discussion 17:53:59 irc://chat.freenode.net:6667/#topicĀ open discussion 17:54:24 anything anyone wants to bring up? BTW - Shawn should be back next week 17:55:13 huge thanks for filling in, and great job! 17:55:24 thanks! 17:56:41 ok i guess we are done this week. Thanks all! 17:57:13 #endmeeting