14:00:33 #startmeeting Nova Live Migration 14:00:35 Meeting started Tue Mar 28 14:00:33 2017 UTC and is due to finish in 60 minutes. The chair is tdurakov. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:38 The meeting name has been set to 'nova_live_migration' 14:01:19 #link agenda https://wiki.openstack.org/wiki/Meetings/NovaLiveMigration 14:01:31 #topic bugs 14:02:47 there is a new bug filed 14:02:54 o/ 14:03:12 #link https://bugs.launchpad.net/nova/+bug/1672597 14:03:12 Launchpad bug 1672597 in OpenStack Compute (nova) "[live migration] The instance directory on the destination host is not clean up" [Undecided,Incomplete] 14:03:45 according to bug desciption ceph based storage in case of failure doesn't clean up destination node 14:04:19 I've asked question to bug reporter to clarify release version 14:04:32 once got response - will work on that 14:05:15 o/ 14:05:18 siva_krishnan left a comment on the https://bugs.launchpad.net/nova/+bug/1671379 - so waiting for the results 14:05:18 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] 14:05:32 this is for the new bugs. 14:05:46 any other new? 14:07:41 for the existing ones, I've wrote to pkoniszewski on the claims fix 14:07:56 so will address comments and resubmit patch-set 14:08:33 #action to review https://review.openstack.org/#/c/434870/ 14:08:47 anything I missed? 14:09:34 tudurakov sorry late for the meeting 14:10:14 siva_krishnan: no problem, I saw your comment on the bug, we've discussed yesterday, thank you 14:10:28 let's wait for the response on that first 14:10:34 moving on 14:10:40 #topic pike specs 14:11:20 config part for completion timeout is waiting for +2 as for me 14:11:47 raj_singh: https://review.openstack.org/#/c/442809/ - I've +1 this one 14:11:55 but left a comment inline 14:12:06 could you please take a look 14:12:21 tdurakov: I just saw that, will reply on the patch 14:12:44 I'm just interested if you did some benchmarking or it's kind of intuitive guess 14:13:08 overall all 3 specs looks good 14:13:27 anything on implementation? 14:13:52 yea I have a backlog spec for that, my goal is to gather data around LM to have proper values for config options 14:14:14 ++ 14:14:26 rescue implementation is in progress, while other 2 I will kick off soon 14:14:38 cool 14:15:14 once ready fill free to ping or add to review them 14:15:22 s/feel 14:15:55 alright 14:17:00 going next 14:17:03 #topic CI 14:17:25 I sam markus_z's patches for serial console were updated 14:17:30 please review them 14:17:53 #link https://review.openstack.org/#/c/346815/ 14:19:20 do you have updates? 14:19:41 tdurakov: Do you update Nova review etherpad once things are ready to be reviewed by core team? 14:20:03 raj_singh: good question, need to update it 14:20:24 yea I was wondering if my specs can be on that list now 14:20:33 raj_singh: yep 14:20:37 it got +1's from John and you 14:20:54 and siva_krishnan 14:21:11 tdurakov: thx 14:21:56 ok 14:22:05 #topic open discussion 14:22:24 * davidgiluk is just working on some qemu patches to allow the RARP/ARP section to be longer/configurable 14:22:50 davidgiluk: do you have a patch on gerrit? 14:23:10 tdurakov: No, not posted yet, and anyway these are QEMU patches so they'll go to qemu-devel 14:23:32 but they may be useful to those who have sluggish network switchovers 14:23:50 davidgiluk: I think it would be useful anyway to take a look 14:23:56 talking about the underlay 14:24:01 nod 14:24:19 don't you know, are there any activity to expose more information on jobinfo? 14:24:33 e.g. disk copy progress during block-migration? 14:25:14 tdurakov: Is this using the older qemu block migration or the new drive-mirror/copy stuff over nbd ? 14:25:33 nbd I expect 14:25:49 * davidgiluk will ask 14:26:05 anyway, afair there are no info about the progress in libvirt job info 14:26:22 so, block migration is threated as a no progress for while 14:26:30 until disk copy is finished 14:26:30 Yes, much thanks to davidgiluk for helping me debug with the RARP / ARP stuff 14:26:48 tdurakov: Which one is this that you're talking about? Sorry, I wasn't paying full attention 14:26:49 tdurakov: OK, I'll figure out if we've got that info anywhere 14:27:23 kashyap: I'm talking about more details exposure on the live-migration job info 14:27:25 You can actually verify by looking at libvirt debug logs what _kind_ of migration is being used -- old-style "migrate" or the NBD-based on 14:29:11 tdurakov: The existing libvirt API to monitor job progress isn't sufficient? (`virsh blockjob --info VM`) 14:29:27 But yeah, we can hash out the details on #openstack-nova 14:30:01 And let the rest of the discussion proceed here 14:30:30 * tdurakov trying to find out link 14:30:37 are there any other open questions? 14:31:28 tudurakov, we do log these info while monitoring live migration right ? 14:32:04 siva_krishnan: yes 14:34:01 if there noting else 14:34:06 let's finish the meeting 14:34:12 thanks for joining 14:34:15 #endmeeting