14:00:36 #startmeeting Nova Live Migration 14:00:37 Meeting started Tue Jan 31 14:00:36 2017 UTC and is due to finish in 60 minutes. The chair is tdurakov. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:38 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:40 The meeting name has been set to 'nova_live_migration' 14:00:47 hello everyone 14:00:56 o/ 14:00:59 Hello 14:01:09 agenda https://wiki.openstack.org/wiki/Meetings/NovaLiveMigration 14:01:20 am new here 14:02:05 so, let's start 14:02:09 #topic ci 14:02:17 o/ 14:02:17 * kashyap waves 14:02:47 o/ 14:02:49 update on serial console activity: hook is good, tempest patch requires additional updates 14:03:12 I expect markus_z will update it soon 14:03:38 grenade, seems to start failing again 14:03:38 * johnthetubaguy lurks to see what is release critical and needs urgent review 14:04:09 johnthetubaguy: https://review.openstack.org/#/c/347471/ 14:04:49 according to mriedem's thread(see agenda) job start skipping live-migration tests, due to tempest misconfig 14:05:10 are there any volunteers to work on this problem? 14:06:06 also, clarkb left comment with details of back_and_forth testing(seems to be easy one) 14:06:36 yeah, i saw this 14:06:55 but i might not find time to work on this in upcoming weeks 14:07:17 I'm going to write status update in ml this week... 14:07:28 will find anyone to fix this 14:07:46 #action tdurakov to find volunteers 14:08:08 tdurakov: i will try to work on it 14:08:18 siva_krishnan: sold!:) 14:08:36 thanks, https://review.openstack.org/#/c/382456/ - see comment 14:08:40 siva_krishnan: if you need any help just ping me on IRC 14:08:59 anything else on ci? 14:09:17 next then 14:09:30 pkoniszewski: sure 14:09:34 #topic bugs 14:10:02 first, haven't seen any *new* bugs, been filed recently 14:10:27 siva_krishnan: i'm based in UTC+1 (3 PM right now here), usually stay at work till 5 PM 14:11:04 so, claims 14:11:14 https://review.openstack.org/#/c/244489/ 14:12:02 johnthetubaguy: review very-very welcome 14:13:04 pkoniszewski: am I missed smth on https://review.openstack.org/#/c/244489/58/nova/compute/manager.py@5676 ? 14:13:35 i'm gonna push new change soon 14:13:41 but i didn't touch this piece of code 14:13:57 we should focus on passing migration over RPC rather than hardening this 14:14:11 this patch will never be perfect... 14:15:21 i'm focused on fixing mdbooth concerns, as the issue that he found in live migration rollback is critical 14:15:44 pkoniszewski: agree on that, but, what if we start passing at least migrate_data object? 14:16:01 shouldn't be a big change 14:16:34 at least I'd like to see this change merged within one release 14:16:43 its more code to this already very complicated change 14:17:51 is it worth passing the extra data around first, then adding the claims in a patch on top, would that make it easier to get in? 14:18:16 i'd say that it will postpone this 1.5 years old patch even more 14:18:21 johnthetubaguy: it would, but this change is ps-58 14:18:22 and this bit is not critical 14:18:25 pkoniszewski: ++ 14:18:41 it's something we might want fix some time in the future 14:19:20 pkoniszewski: I'm ok then to address mdbooth comments first, then merge this one, and after that polish it 14:19:35 johnthetubaguy: what do you think? 14:19:45 * mdbooth was ok with it otherwise except for the race I misread. 14:20:01 mdbooth: but the rollback is broken as you pointed it out 14:20:06 pkoniszewski: Yeah. 14:20:15 mdbooth: this is what i'm trying to fix currently 14:20:17 I found 2 problems, and one of them wasn't a problem. 14:20:23 breaking rollback sounds bad 14:20:26 since rollback is cast, right? 14:20:33 mdbooth: yeah, misundestood you, your right 14:20:46 not because it is a cast 14:21:12 but it calls save within context manager 14:21:25 ah, that one, ok 14:21:32 to many ps... 14:21:34 mutated_migration_context context manager 14:21:36 s/too 14:22:23 so, let's wrap this up, fixing this up, then merge, then polish, agree? 14:23:12 i'd prefer this way 14:23:22 johnthetubaguy: ^ 14:25:02 I have a question: Does openstack fully supports parallel (concurrent) live migration of VMs ? 14:25:10 let's move on, I think johnthetubaguy will express opinion on review anyway 14:25:43 armstrong: it supports 14:25:57 any other bugs/fixes? 14:26:29 next 14:26:58 #topic Pike specs 14:27:04 tdurakov: yeah, I need more context on that one, everything I read through it my head explodes, sadly 14:27:16 s/everything/every time/ 14:27:23 johnthetubaguy: np, just review:-p 14:27:58 * pkoniszewski going offline 14:28:01 thanks for the meeting! 14:28:05 bb 14:28:14 I will like to fullt involve 14:28:33 fully* 14:30:04 I'd like to ask everyone to review not only l-m related specs, but all migration specs for pike, e.g. cold-migration abort or cold-migration with target hosts, since it will help to align migrations in nova 14:30:06 tdurakov: where can I start contributing? 14:30:15 and make them consistent 14:30:30 #topic open discussions 14:31:24 armstrong: take a look at bugs: https://bugs.launchpad.net/nova/+bugs?field.tag=live-migration+ 14:31:38 ok thanks 14:31:40 also contribution to ci and test coverage is very welcome 14:31:51 e.g. tempest 14:32:07 + review of the l-m related patches 14:32:20 do we have any open topics? 14:33:27 ok 14:33:31 thanks for coming 14:33:35 #endmeeting