#openstack-meeting log

14:01:06 <mriedem> #startmeeting nova
14:01:06 <openstack> Meeting started Thu Jun 16 14:01:06 2016 UTC and is due to finish in 60 minutes.  The chair is mriedem. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:09 <openstack> The meeting name has been set to 'nova'
14:01:10 <claudiub> o/
14:01:12 <jroll> \o
14:01:12 <lxsli> o/
14:01:12 <_diana_> o/
14:01:12 <BobBall> o/~
14:01:14 <PaulMurray> o/
14:01:15 <edleafe> \o
14:01:16 <takashin> o/
14:01:18 <dansmith> yo
14:01:19 <andrearosa> hi
14:01:20 <gibi_> o/
14:01:21 <MikeG451_> o\
14:01:22 <auggy> o/
14:01:23 <gagehugo> o_/
14:01:25 <moshele> hi
14:01:29 <lyarwood> o/
14:01:29 <scottda> hi
14:01:34 <rlrossit_> o/
14:01:44 <cdent> o/
14:01:49 <bauzas> \o
14:01:51 <sahid> o/
14:01:55 <lbeliveau> o/
14:01:55 <johnthetubaguy> o/
14:02:04 <woodster_> o/
14:02:07 <macsz> o/
14:02:10 <mnestratov|2> o/
14:02:12 * bauzas hopes not to see a lot of UK guys here :p
14:02:21 <macsz> :D
14:02:29 <mriedem> alright let's get started
14:02:34 <mriedem> #link meeting agenda https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting
14:02:40 <mriedem> #topic release news
14:02:50 <mriedem> #link Newton release schedule: https://wiki.openstack.org/wiki/Nova/Newton_Release_Schedule
14:02:56 <johnthetubaguy> bauzas: heh, I had to think quite hard what I was meant to be doing
14:03:01 <mriedem> #info June 30: non-priority feature freeze (2 weeks from today)
14:03:11 <mriedem> We have 100 approved blueprints: https://blueprints.launchpad.net/nova/newton - 17 are completed, 16 have not started, 2 are blocked
14:03:25 <mriedem> last week we had 86 approved blueprints, then johnthetubaguy and i did some spec/LP syncing
14:03:51 <mriedem> way more than we can possibly get done, which shouldn't come as a surprise
14:04:10 <mriedem> looking back over mitaka and liberty we approved around the same number, but implemented ~60
14:04:21 <mriedem> just noting that for context
14:04:31 <mriedem> any questions about the release?
14:04:40 <johnthetubaguy> +1 on mriedem's comments there
14:04:58 <mriedem> #topic bugs
14:05:13 <mriedem> i guess the topic thingy isn't working
14:05:22 <mriedem> gate status looks ok
14:05:32 <mriedem> i'm not aware of any blocking issues there
14:05:58 <mriedem> 3rd party CI status - BobBall - looks like xenserver CI is OK again?
14:06:04 <BobBall> I think so :)
14:06:18 <mriedem> mnestratov|2: any updates on the virtuozzo compute CI?
14:06:21 <dansmith> what about
14:06:22 <dansmith> yeah that :)
14:06:30 <mnestratov|2> mriedem: work in progress
14:06:44 <dansmith> mnestratov|2: thanks for being responsive on this!
14:06:48 <mnestratov|2> i hope we will start it again today
14:06:53 <mriedem> mnestratov|2: ok, sounds good
14:07:08 <mriedem> i don't see anyone from vmware around
14:07:18 <mriedem> but tjones was reporting issues in last week's meeting
14:07:40 <mriedem> critical bugs - i'm not aware of any
14:07:55 <mriedem> is auggy here?
14:08:01 <auggy> i am!
14:08:02 <mriedem> she is
14:08:16 <auggy> i'm not aware of any critical bugs, but i haven't been keeping track
14:08:17 <mriedem> aunnam_: sneti: any critical bugs this week?
14:08:27 <mriedem> ^ is on bug skimming duty
14:08:42 <mriedem> alright moving on
14:08:45 <mriedem> #topic reminders
14:08:50 <mriedem> #link Newton review focus list: https://etherpad.openstack.org/p/newton-nova-priorities-tracking
14:08:59 <mriedem> general reminder to please keep that up to date
14:09:06 <mriedem> #help https://wiki.openstack.org/wiki/Nova/BugTriage#Weekly_bug_skimming_duty Volunteers for 1 week of bug skimming duty?
14:09:17 <mriedem> looks like we need people for next week
14:09:34 <mriedem> open/new bugs are creeping up, 27 new in my view this morning
14:10:10 <mriedem> usually when i complain about bug backlog sdague goes wild but he's on vacation
14:10:17 <bauzas> okay, I'll see how I can help
14:10:26 <mriedem> bauzas: thanks, i'll try to triage a few today too
14:10:32 <mriedem> and we can high five later
14:10:36 * bauzas digs for his launchpad cape
14:10:36 <mriedem> #topic stable branch status
14:10:46 <mriedem> nova mitaka 13.1.0 was released this week
14:10:55 <mriedem> which contains a CVE fix
14:11:02 <mriedem> and a bunch of other good fixes
14:11:15 <mriedem> not much else for stable branch news
14:11:22 <mriedem> #topic subteam highlights
14:11:31 <mriedem> there wasn't a cells meeting this week, but there is plenty to review
14:11:40 <mriedem> so check https://etherpad.openstack.org/p/newton-nova-priorities-tracking for cells patches
14:11:47 <mriedem> sounds like the aggregates migration patches are moving along
14:12:02 <mriedem> and melwitt has started on instance groups migration
14:12:08 <mriedem> still need someone to start on quotas migrations though
14:12:17 <mriedem> scheduler,
14:12:24 <mriedem> edleafe: highlights from scheduler meeting?
14:12:29 <mriedem> cdent: ^
14:12:41 <cdent> loads of generic-resource-pools code pending https://review.openstack.org/#/q/topic:bp/generic-resource-pools+status:open
14:12:52 <edleafe> with https://review.openstack.org/#/c/326440
14:12:57 <edleafe> ugh
14:13:07 <edleafe> Still focusing on cdent's series on resource providers / allocation / inventory - starts with https://review.openstack.org/#/c/326440
14:13:31 <edleafe> Live migrate multiple host spec still being reviewed, but now with POC code
14:13:41 <mriedem> so,
14:13:45 <edleafe> that's about it
14:13:48 <mriedem> why are we reviewing that spec?
14:14:06 <edleafe> mriedem: mainly for direction
14:14:18 <mriedem> but are people expecting that to be approved for newton?
14:14:23 <mriedem> or just looking to ocata?
14:14:23 <edleafe> no
14:14:26 <mriedem> ah, ok
14:14:34 <edleafe> it's like the qualitative stuff
14:14:41 <edleafe> we want the discussion on the spec
14:14:51 <edleafe> but we know it's ocata at the earliest
14:14:55 <mriedem> ok, makes sense, plus i'm sure this is fodder for the midcycle
14:15:09 <edleafe> I could move it to backlog if that makes things clearer
14:15:16 <mriedem> no ocata is fine
14:15:16 <edleafe> mriedem: yeah, that too
14:15:31 <bauzas> ocata looks honestly the most reasonable
14:15:48 <mriedem> 50% of newton is going to be ocata right now :)
14:15:48 <mriedem> so yeah
14:15:57 <mriedem> moving on - thanks guys
14:16:02 <bauzas> either way, we are on track with what we agreed
14:16:03 <edleafe> bauzas: but if we don't discuss now, ocata won't be possible
14:16:07 <mriedem> PaulMurray: live migration meeting highlights?
14:16:07 <cdent> on the resource providers stuff, near the end of that stack is the API WIP, which might benefit from some early review as, by design, it is not using nova.wsgi, and will probably rub some people funky, so we need to reach agreement sooner than later
14:16:20 <PaulMurray> most of the issues are around CI
14:16:36 <PaulMurray> mriedem, probably has a better grasp of that than me
14:16:49 <mriedem> we have a patch for testing the stack with imagebackend 'flat'
14:16:50 <PaulMurray> NFS and ceph CI broken on xenial
14:17:00 <mriedem> yeah we're excluding nfs/ceph in the live migration job for a week
14:17:10 * PaulMurray xenial = 16.04
14:17:11 <mriedem> to see how stable the 'normal' scenario is
14:17:32 <mriedem> tdurakov has a fix for the nfs setup issue, but we're holding off o nit
14:18:16 <mriedem> PaulMurray: anything else?
14:18:18 <PaulMurray> Storage pools work is progressing but worries around testing due to CI
14:18:21 <bauzas> I thought infra wasn't supporting xenial ?
14:18:25 <PaulMurray> same as above
14:18:31 <mriedem> bauzas: they have been for a few weeks now
14:18:43 <PaulMurray> manual testing being done as well
14:18:51 <mriedem> we moved the live migration multinode job to xenial in the hopes that newer libvirt/qemu would solve some of the bugs
14:19:08 <PaulMurray> not to mention being able to test new features.....
14:19:42 <PaulMurray> not much else to report on
14:19:58 <mriedem> yeah, i'd like to see a tempest test for live migrating a rescue instance before that code lands, for example
14:20:27 <PaulMurray> mriedem, paul_carlton is working on it
14:20:38 <bauzas> mriedem: k
14:20:40 <mriedem> but if we can figure out that the live migration job in it's basic config is stable(ish) then we can talk about adding that to the check queue, at least non-voting
14:20:43 * PaulMurray that's all the paul_carltons
14:21:00 <mriedem> ok, moving on, thanks Paul
14:21:09 <mriedem> api, sdague is out
14:21:12 <mriedem> alex_xu: around?
14:21:27 <mriedem> i was at that one,
14:21:40 <mriedem> we mostly talked about the api proxy deprecation work, gmann has started on that
14:21:59 <mriedem> and we talked about the extension removal spec, which sdague also posted in the ops and dev ML
14:22:26 <mriedem> sdague has also started working on folding the various extensions back into the main paths
14:22:45 <mriedem> e.g. https://review.openstack.org/#/c/327188/
14:23:21 <mriedem> the idea being agree on how that's done with one extension (disk-config in this case), then others can start doing other extensions following that pattern
14:23:47 <mriedem> api-ref conversion is still happening too, albeit somewhat slowly
14:24:03 <mriedem> moshele: sriov/pci meeting highlights?
14:24:20 <moshele> yes, so regrading CI
14:24:22 <rbradfor> mriedem, I try to give api-ref a few hours a week just to see move
14:24:31 <moshele> Mellanox CI with containers didn't work for us
14:24:43 <mriedem> rbradfor: <3
14:25:06 <moshele> many issues with containers so now we just looking on SR-IOV CI with multinode
14:25:18 <moshele> for the pci migration
14:26:03 <moshele> we several patches for fixing pci migration and revert migration   pci: Update binding:profile for SR-IOV ports - https://review.openstack.org/#/c/242573/21  Fix revert on migration with SR-IOV - https://review.openstack.org/#/c/326174 Allocate PCI devices on migration: https://review.openstack.org/#/c/328983/4
14:27:22 <moshele> also we found that resize and migration are racy because we don't update the resource usage and we wait for periodic task and this tix should solve it Update available resources before confirm stage  https://review.openstack.org/#/c/327356/
14:27:59 <moshele> we need to so more manual testing on migration and migration revert I plan to do it next week
14:28:58 <mriedem> ok, sounds good. multinode CI for this stuff would be amazing.
14:29:19 <mriedem> moshele: anything else?
14:29:22 <moshele> I update the etherpad with patches that are ready for core review   https://etherpad.openstack.org/p/newton-nova-priorities-tracking
14:30:18 <mriedem> ok, thanks moshele
14:30:26 <mriedem> gibi_: was there a notifications meeting this week?
14:30:32 <gibi_> yes, there was
14:30:43 <gibi_> notification transformation work is ongoing
14:30:55 <gibi_> the first step, moving the notification code to separate package, has been merged
14:31:16 <gibi_> the example notification transformations are up on review
14:31:23 <gibi_> some of them is already waiting for core review
14:31:31 <gibi_> see https://etherpad.openstack.org/p/newton-nova-priorities-tracking L211
14:31:38 <gibi_> besides that
14:31:56 <gibi_> people are jumped on the low hangig fruit to transform similar notifications
14:32:08 <gibi_> we have about 10 transformation patch proposed
14:32:26 <gibi_> subteam is reviewing those
14:32:37 <gibi_> but they are quite easy patches
14:33:01 <mriedem> i'll be honest that the wrap_exception one confused me
14:33:08 <mriedem> but i didn't stare at it too long
14:33:25 <rlrossit_> mriedem: just ask, I'm more than willing to explain :)
14:33:30 <gibi_> me too
14:33:31 <gibi_> :)
14:33:51 <mriedem> #action mriedem to have a heart to heart with rlrossit_ this afternoon about that wrap_exception change
14:34:00 <mriedem> anything else?
14:34:03 <rlrossit_> <3
14:34:14 <gibi_> that is all I think
14:34:20 <mriedem> ok, thanks
14:34:26 <mriedem> #topic stuck reviews
14:34:30 <mriedem> there was nothing in the agenda
14:34:37 <mriedem> anything someone needs to bring up there?
14:34:52 <mriedem> #topic open discussion
14:34:58 <mriedem> nothing on the agenda so i'll open it up
14:35:02 <auggy> o/
14:35:11 <mriedem> \o
14:35:21 <auggy> i've been pointing new contributors at the api-ref project, so just a heads up if you see folks asking in #nova
14:35:32 <auggy> that's all :)
14:35:35 <mriedem> they have
14:35:45 <mriedem> and i point them at the wiki :)
14:35:46 <mriedem> or at you
14:36:01 <auggy> ha, yeah thanks :)
14:36:07 <auggy> well i meant about api-ref stuff
14:36:10 <mriedem> mostly questions like, how will i know i've done it correctly and it's perfect?
14:36:35 <mriedem> the -2 hammer of sass will let you konw
14:36:37 <mriedem> that's how
14:36:49 <mriedem> any other items for open discussion
14:36:50 <mriedem> ?
14:36:51 <andrearosa> implementation for non priority specs is ready with no reviews (actually just one): https://review.openstack.org/#/q/topic:bp/detach-boot-volume  and the code-freeze is coming.
14:36:52 <auggy> ah ok, I will update the Mentoring wiki page with a reminder to review the dev process and a link to what reviews are etc
14:37:19 <mriedem> andrearosa: is there a tempest test for that?
14:37:28 <andrearosa> in the last change yes
14:37:33 <andrearosa> nope
14:37:35 <andrearosa> sorry
14:37:40 <andrearosa> I have to check
14:37:49 <mriedem> andrearosa: i want a tempest test on something like that before we land it
14:37:56 <andrearosa> mriedem: ack
14:38:14 <mriedem> honestly i've spent most of the last week ironing out the tempest test for the get-me-a-network series and already flushed out race bugs in neutron because of it
14:38:43 <mriedem> so when we say in specs that we'll add tempest tests for things, i really want to stick to doing that before actually landing code
14:38:49 <mriedem> because i want to see tempest passing on a change first
14:38:53 <andrearosa> makes sense
14:39:03 <mriedem> andrearosa: i'm not trying to pick on you
14:39:07 <mriedem> i'm just beating a drum
14:39:40 <mriedem> ok, anything else?
14:40:07 <mriedem> then let's end early, thanks everyone
14:40:09 <mriedem> #endmeeting