14:00:56 <tdurakov> #startmeeting Nova Live Migration 14:00:57 <openstack> Meeting started Tue Mar 21 14:00:56 2017 UTC and is due to finish in 60 minutes. The chair is tdurakov. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:02 <openstack> The meeting name has been set to 'nova_live_migration' 14:01:38 <davidgiluk> o/ 14:01:41 <siva_krishnan> o/ 14:01:42 <tdurakov> #link agenda: https://wiki.openstack.org/wiki/Meetings/NovaLiveMigration#Agenda_for_next_meeting 14:02:20 * tdurakov happy to get back, since the last week was a cold-time 14:02:29 <tdurakov> so 14:02:40 <mriedem> o/ 14:02:50 <tdurakov> let's start, I'd reorder topics in agend 14:02:54 <tdurakov> s/agenda 14:03:02 <tdurakov> #topic Pike specs 14:03:26 <tdurakov> first one: https://review.openstack.org/#/c/438467/ 14:04:41 <tdurakov> I think it's ok to merge it 14:04:54 <tdurakov> johnthetubaguy: any updates on that? 14:06:13 <siva_krishnan> tudurakov I dont see him online 14:06:26 <tdurakov> oh 14:07:01 <tdurakov> I'd add it to the weekly highlights, hope nova cores review/merge it 14:07:15 <tdurakov> next one: https://review.openstack.org/#/c/442809/ 14:07:40 <tdurakov> actually it's a second part of the topic, been discussed on the ptg 14:08:18 <siva_krishnan> yeah, its REST microversion API changes for first one 14:09:02 <tdurakov> I've +1 first one, but need to read second once again 14:09:29 <tdurakov> anyway, there is +2 from johnthetubaguy already, so looks good 14:10:43 <tdurakov> the third one: https://review.openstack.org/#/c/347161/ 14:10:58 <tdurakov> siva_krishnan: how's the progress with the implementation? 14:11:56 <siva_krishnan> tdurakov: patches are up. Probably need to do some more testing and add some test cases 14:12:13 <tdurakov> link please? 14:12:42 <mriedem> siva_krishnan: are you also picking up the tempest change for that api? 14:12:50 <siva_krishnan> https://review.openstack.org/#/c/308198/ 14:13:12 <mriedem> https://review.openstack.org/#/c/336414/ 14:13:14 <mriedem> siva_krishnan: ^ 14:13:25 <siva_krishnan> mriedem: yeah need to add tempest test as well. Need changes on python-novaclient as well 14:13:33 <mriedem> ^ is the tempest change 14:13:48 <mriedem> i'll get it restored 14:13:54 <siva_krishnan> thanks mriedem :) 14:14:16 <tdurakov> siva_krishnan: novaclient change, just to add new microversion? 14:14:37 <siva_krishnan> tdurakov: yeah 14:14:49 <johnthetubaguy> its a raise the max supported version thing, I think 14:15:08 <tdurakov> ok 14:15:16 <tdurakov> anything else on the specs? 14:16:09 <johnthetubaguy> I think thats all the ones raj_singh and siva_krishnan are working on, but I could have missed one 14:16:48 <tdurakov> ok 14:16:58 <tdurakov> next 14:17:02 <tdurakov> #topic bugs 14:17:55 <tdurakov> there is a new one for the l-m https://bugs.launchpad.net/nova/+bug/1671379 14:17:55 <openstack> Launchpad bug 1671379 in OpenStack Compute (nova) "The first VM of one network in one compute node cannot send RARP packets during KVM's live-migration in a neutron ML2 hierachical port binding environment whose second mechanism driver was configured as the existing OVS driver "openvswitch"" [Undecided,New] 14:20:02 <davidgiluk> tdurakov: I wish I understood the details; I do get quite a few reports of openstack migrations losing the RARPs 14:21:21 <tdurakov> afair there were one related to external events and nova-neutron communication, but it should be already fixed 14:22:35 <davidgiluk> tdurakov: My understanding is that openstack doesn't guarantee the networking is setup by the point qemu completes the migration 14:22:51 <davidgiluk> tdurakov: whether that's related to the detail of that bug is a different matter 14:25:21 <tdurakov> who is interested to take a look at the problem? 14:26:06 <tdurakov> if it's about earlier neutron notifications it should be partially addressed by https://review.openstack.org/#/c/434870/ 14:26:16 <tdurakov> but this one is about pre-copy 14:27:03 <siva_krishnan> tdurakov: i think above patch takes care of both post copy and pre copy as well 14:27:34 <tdurakov> siva_krishnan: please left a comment on the bug 14:28:02 <raj_singh> siva_krishnan: Maybe good to test the above bug with your fix? 14:28:09 <davidgiluk> siva_krishnan: At what point does that series cause the destination to wire in the networking? 14:28:25 <siva_krishnan> tdurakov: will do it. 14:28:55 <raj_singh> davidgiluk: I believe when we get domain pause event from source VM 14:29:38 <siva_krishnan> davidgiluk: 14:29:46 <davidgiluk> raj_singh: ok, and how long does it then take OS to do the rewiring? 14:30:42 <raj_singh> siva_krishnan: I think you did some testing to measure the downtime, right? 14:31:08 <siva_krishnan> davidgiluk: based on following events VIR_DOMAIN_EVENT_SUSPENDED_POSTCOPY- post copy and libvirt.VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED- precopy 14:31:58 <siva_krishnan> I could just see 0-2% packet loss 14:32:16 <davidgiluk> siva_krishnan/ raj_singh: OK, so just check because I'd heard neutron can take a few seconds to reconfigure and that would still miss the RARPs 14:32:40 <tdurakov> siva_krishnan: what's the total # of packets? 14:33:32 <siva_krishnan> tdurakov: http://paste.openstack.org/show/601673/- Here are my test results 14:34:32 <tdurakov> ok, thank you 14:35:27 <tdurakov> to sum up, it would be great to check the bug with the patch been applied 14:36:05 <tdurakov> next one 14:36:22 <tdurakov> https://review.openstack.org/#/c/244489/ 14:37:09 <tdurakov> review/merge is needed 14:38:43 <tdurakov> the last one we already mentioned is about events for live-migration 14:39:02 <tdurakov> https://review.openstack.org/#/c/434870/ 14:39:50 <tdurakov> review it please too 14:39:56 <tdurakov> any other bugs? 14:40:41 <tdurakov> next 14:40:46 <tdurakov> #topic CI 14:41:14 <tdurakov> any updates on this? 14:42:16 <tdurakov> ok 14:42:23 <tdurakov> #topic open discussions? 14:44:42 <tdurakov> thanks everyone 14:44:52 <tdurakov> have a good day 14:44:57 <tdurakov> #endmeeting