14:00:07 <johnthetubaguy> #startmeeting Nova
14:00:08 <openstack> Meeting started Thu Jun 26 14:00:07 2014 UTC and is due to finish in 60 minutes.  The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:12 <openstack> The meeting name has been set to 'nova'
14:00:13 <dansmith> o/
14:00:19 <johnthetubaguy> #link https://wiki.openstack.org/wiki/Meetings/Nova
14:00:21 <erw> o/
14:00:23 <johnthetubaguy> hows around today then?
14:00:25 <bauzas> o/
14:00:27 <n0ano> o/
14:00:34 <alaski> o/
14:00:35 <mriedem> hi
14:00:46 <johnthetubaguy> #topic Juno-2 is 24th July
14:00:49 <oomichi> hi
14:00:54 <johnthetubaguy> so lets get going, loads on the agenda today
14:01:10 <johnthetubaguy> first thanks for the spec review push
14:01:30 <johnthetubaguy> made a good dent, still not home and dry, bug a good dent is better than nothing
14:01:42 <johnthetubaguy> #link http://lists.openstack.org/pipermail/openstack-dev/2014-June/038475.html
14:01:53 <johnthetubaguy> there was much discussion on the proposed nova-spec freeze
14:01:57 <bauzas> sorry, missed the agenda, do you have one for today? :)
14:02:14 <johnthetubaguy> bauzas: same place as always: https://wiki.openstack.org/wiki/Meetings/Nova
14:02:27 <johnthetubaguy> so, any thoughts on the freeze
14:02:41 <johnthetubaguy> I think people are happy with us slowing down what goes into Juno right?
14:02:49 <johnthetubaguy> blueprint proposal freeze date of 2nd July
14:03:02 <sgordon> right
14:03:02 <johnthetubaguy> well, 3rd
14:03:06 <sgordon> the main source of disagreement
14:03:08 <johnthetubaguy> Jul 10 (-8): Spec approval freeze
14:03:13 <johnthetubaguy> its the K spec stuff
14:03:14 <sgordon> was whether to close spec review for K off
14:03:35 <johnthetubaguy> so I vote we just open K on Jul 10 (-8): Spec approval freeze
14:03:45 <johnthetubaguy> and we talk more at the mid cycle summit
14:03:53 <johnthetubaguy> that seems OK with everyone?
14:03:55 <dansmith> really?
14:03:58 <dansmith> open K already?
14:04:10 <alaski> I don't think we can approve anything for K, right?
14:04:16 <johnthetubaguy> right
14:04:26 <johnthetubaguy> we hold off approve till after the midcycle
14:04:35 <johnthetubaguy> we may just −2 everything after the midcycle
14:04:44 <dansmith> why would we approve things for K even as early as after the meetup?
14:04:59 * dansmith is confused
14:05:12 <johnthetubaguy> dansmith: I don't think we should, but others seemed to disagree, so I figure we just delay approves till after the meet up at the earliest
14:05:23 <johnthetubaguy> sorry, wasn't very clera
14:05:27 <johnthetubaguy> clear^
14:05:33 <dansmith> man, I'm a huge -2 on that
14:05:39 <alaski> me too
14:05:43 <sgordon> did anyone actually want to approve them?
14:05:43 <johnthetubaguy> on what, opening it at all?
14:05:53 <sgordon> i thought it was just people wanting to be able to submit
14:06:05 <mriedem> i thought part of the point of the freeze was to avoid spec review distractions?
14:06:06 <dansmith> well, I suppose anyone can submit anything they want, but it's just going to muddy the waters, stats-wise and dashboard-wise
14:06:12 <mriedem> opneing it up for K defeats that
14:06:16 <dansmith> right, that will massively increase distraction
14:06:36 <johnthetubaguy> dansmith: OK, so we keep K closed till the midcylce where we decide what to do I guess?
14:06:44 <mriedem> dansmith: when you get a chance, can you review my spec for L? thanks!
14:06:53 <johnthetubaguy> (I was hoping phil was here to argue the other side)
14:06:55 <dansmith> johnthetubaguy: as long as at the meetup we decide to punt until much later :P
14:07:07 <johnthetubaguy> dansmith: right, thats what I am hoping for
14:07:31 <dansmith> I should go catch up on that thread I guess
14:07:33 <alaski> patches can still be submitted, and if people want to review them they can.  But there's not going to be any real attention from cores/drivers
14:07:39 <n0ano> so that means we freeze new specs after 7/3 and don't accept new ones until when?
14:08:00 <dansmith> alaski: well, like I said anyone can submit whatever they want, but I think we should try to discourage it
14:08:18 <johnthetubaguy> n0ano: we haven't decided when, we proposed  Sep 25 (+3): RC 1 build expected, K spec review approvals start
14:08:21 <dansmith> alaski: although, I think maybe we can filter by path with new gerrit
14:08:25 <johnthetubaguy> but there was not really agreement on that
14:08:37 <dansmith> alaski: so maybe we can create a dashboard that is only J thing and then it's not as much of an issue, I dunno
14:08:43 <n0ano> that's an awful long hiatus on specs
14:08:44 <dansmith> but ffs, lets at least wait and discuss at the meetup :)
14:08:48 <johnthetubaguy> dansmith: oh thats a good idea possibly...
14:08:50 <johnthetubaguy> yeah
14:09:03 <johnthetubaguy> we are probably freezing Juno though in a few weeks
14:09:06 <alaski> dansmith: sure.  we should definitely punt unti the meetup
14:09:11 <johnthetubaguy> please moan on the ML if you don't want that
14:09:36 <johnthetubaguy> anyways, lets not argue all day on this
14:09:44 <johnthetubaguy> #topic Bugs
14:10:28 <johnthetubaguy> OK, so not much info here
14:10:35 <johnthetubaguy> lets cover gate bugs in the gate bit...
14:10:41 <johnthetubaguy> #topic Gate
14:10:46 <johnthetubaguy> right...
14:11:01 <johnthetubaguy> do people want to talk about anything here?
14:11:04 <mriedem> the live snapshot workaround should be going through the gate now
14:11:11 <dansmith> I think the recent major problem is worked around
14:11:12 <johnthetubaguy> #link https://review.openstack.org/#/c/102643/
14:11:12 <dansmith> yeah
14:11:21 <mriedem> i posted a few bugs
14:11:30 <mriedem> two are image related, so could be dupes now
14:11:34 <mriedem> looking at logstash they spike at the same time
14:11:44 <mriedem> the other is the ssh timeout bug we've had for awhile
14:11:47 <johnthetubaguy> yeah, just wanted to say, we think we have that sorted now, hopefully
14:11:53 <mriedem> theory is still a boto leak in network resources
14:11:54 <johnthetubaguy> mriedem: cool fire away
14:11:59 <johnthetubaguy> do people need help on that?
14:12:11 <mriedem> there is a link with a series of ec2 reviews that we should focus on
14:12:31 <mriedem> basically there are people working on upstreaming ec2 fixes and want to submit better tempest tests for ec2, but they need the nova bugs fixed first
14:12:35 <johnthetubaguy> mriedem: did you say this might be a dup? https://bugs.launchpad.net/nova/+bug/1320617
14:12:37 <uvirtbot> Launchpad bug 1320617 in nova "[Image] failed to reach ACTIVE status within the required time (196 s). Current status: SAVING" [Undecided,New]
14:12:47 <mriedem> johnthetubaguy: yup
14:12:56 <mriedem> not sure though, haven't dug into the fails on that one in awhile
14:13:10 <mriedem> but the logstash trends were similar after the move to trusty
14:13:11 <johnthetubaguy> OK, so help welcome there I guess
14:13:16 <mriedem> http://status.openstack.org/elastic-recheck/
14:13:30 <johnthetubaguy> looks like you are digging into this one: https://bugs.launchpad.net/tempest/+bug/1298472
14:13:36 <uvirtbot> Launchpad bug 1298472 in tempest "SSHTimeout in tempest scenario tests using nova-network" [Critical,Fix committed]
14:13:45 <johnthetubaguy> cool, any more for any more on this?
14:14:00 <mriedem> nope
14:14:15 <johnthetubaguy> #help gate is not a happy bunny some more help digging would be good
14:14:29 <johnthetubaguy> #topic Review status
14:14:43 <johnthetubaguy> so this could be a section we drop, but I wanted to look at our review queue
14:15:00 <johnthetubaguy> #link nova-specs queue http://russellbryant.net/openstack-stats/nova-specs-openreviews.html
14:15:20 <johnthetubaguy> wait times have been going quite high, so we need to look at bit further down the list
14:15:28 <johnthetubaguy> I have started clicking abandon on patches
14:15:43 <johnthetubaguy> turns out that bot isn't turned on any more, after gerrit upgrade
14:15:47 <mriedem> right
14:15:52 <johnthetubaguy> we might want to turn that back on at some point...
14:15:55 <mriedem> would be nice to turn that on again.....or auto-WIP
14:16:02 <johnthetubaguy> but the list in the stats is really good, and I forgot about that
14:16:14 <johnthetubaguy> mriedem: hmm, i do like that idea, seems kinder
14:16:23 <dansmith> I'd rather auto-abandon
14:16:34 <dansmith> auto-wip just means we have a long tail, when we know what it really means
14:16:34 <mriedem> yeah, auto-abandon is better for bugs
14:16:45 <dansmith> on the spec review stats,
14:16:46 <johnthetubaguy> dansmith: it is cleaner… and we used to do that
14:16:59 <dansmith> if you look at it, there is actually a relatively small set that have been waiting longer than three days
14:17:11 <dansmith> like six out of 106 open reviews
14:17:16 <dansmith> I think that's pretty damned good
14:17:16 <johnthetubaguy> #link nova stats http://russellbryant.net/openstack-stats/nova-openreviews.html
14:17:19 <johnthetubaguy> so...
14:17:28 <dansmith> granted, we just had a review day, but.. I'm not too concerned
14:17:35 <johnthetubaguy> Total Open Reviews: 631
14:17:35 <johnthetubaguy> Waiting on Submitter: 387
14:17:35 <johnthetubaguy> Waiting on Reviewer: 244
14:18:05 <johnthetubaguy> yeah, its not the worst its been
14:18:13 <johnthetubaguy> seems I got a bit laggy for a bit there
14:18:34 * mriedem isn't too concerned honestly
14:18:35 <johnthetubaguy> the waiting on submitter is probably really high due to the lack of auto-abandon
14:19:02 <johnthetubaguy> well waiting 134 days since last −1 or −2 seems bad, but yeah, generally not too bad overall
14:19:16 <johnthetubaguy> moving on… unless there are more ideas
14:19:27 <mriedem> if you're waiting 6 months for a response, and it's important,
14:19:30 <mriedem> go to IRC :)
14:19:38 <johnthetubaguy> well… that is fair
14:19:51 <mriedem> idk, be more active people
14:19:59 <dansmith> for serious..
14:20:03 <dansmith> next...
14:20:12 <johnthetubaguy> #topic Sub team reports
14:20:31 <johnthetubaguy> (yeah, thats a little pointless in that section , will try improve that one...)
14:20:34 <johnthetubaguy> so...
14:20:47 <johnthetubaguy> not sure we have everyone here, so hands up to talk?
14:20:53 <erw> Docker - we’re working on Cinder support
14:20:53 <bauzas> o/
14:20:54 * n0ano gantt
14:21:18 <johnthetubaguy> erw: you were first, any more news?
14:21:27 <erw> slower and funzo have it working, but we need upstream patches in docker itself
14:21:45 <johnthetubaguy> ah, well sounds like a good step forwards, thanks for the update
14:21:48 <erw> a spec is incoming for merging back in
14:21:52 <erw> and we’re working on the the CI
14:21:56 <erw> that’s it
14:22:18 <johnthetubaguy> erw: that could well get pushed to K at this point, as a heads up, but maybe we will see people at the mid cycle meet up
14:22:35 <johnthetubaguy> bauzas: n0ano: scheduler stuff?
14:22:36 <dansmith> yeah, seems most likely for K to me
14:22:48 <funzo> our intent was to make the july 2 spec deadline for Nova
14:22:51 <erw> johnthetubaguy: acknowledged. I’m okay with that, personally - although I know there are users that would prefer otherwise.
14:22:54 <n0ano> mostly work in progress, bauzas can give more details if you want
14:22:59 <bauzas> johnthetubaguy: well, good progress on the scheduler client so far
14:23:00 <funzo> and if it didn't get in for Juno, we would be in good shape for K
14:23:22 <bauzas> johnthetubaguy: we're still waiting reviews for a crucial spec
14:23:38 <bauzas> johnthetubaguy: https://review.openstack.org/89893
14:23:46 <johnthetubaguy> erw: funzo: I think ironic should be our goal in Juno, so probably best to be sure of being ready on day 1 of K
14:24:02 <bauzas> johnthetubaguy: implementation drafts began re: 89893
14:24:03 <johnthetubaguy> bauzas: its in my browser, I will take look
14:24:12 <erw> anyway - we can offline that conversation.
14:24:12 <bauzas> johnthetubaguy: thanks
14:24:26 <johnthetubaguy> erw: sure, sounds good :)
14:24:37 <bauzas> johnthetubaguy: and as you know, for the records, https://review.openstack.org/82778 is ready for reviewing
14:24:48 <johnthetubaguy> bauzas: :)
14:24:59 <bauzas> that's it for me
14:25:06 <johnthetubaguy> bauzas: thanks
14:25:11 <johnthetubaguy> any more sub team stuff?
14:25:25 <johnthetubaguy> ironic, NFV, etc?
14:25:31 <johnthetubaguy> no update from the XenAPI land really
14:25:36 <johnthetubaguy> OK...
14:25:43 <johnthetubaguy> #topic Open discussion
14:25:54 <johnthetubaguy> so first topic, unit tests to trusty only
14:26:01 <dansmith> anyone opposed to that?
14:26:02 <johnthetubaguy> anyone want to talk to that?
14:26:09 <dansmith> seems like it should be all good at this point
14:26:13 <dansmith> tempest/devstack not so much,
14:26:20 <johnthetubaguy> yeah, seems sound for unit tests
14:26:21 <dansmith> but infra would like to stop running unit tests on precise,
14:26:23 <dansmith> and I think that seems fine
14:26:39 <dansmith> mriedem: any concerns?
14:26:55 <mriedem> unit tests shouldn't make much of a difference on trusty right?
14:27:00 <mriedem> mostly py27
14:27:06 <mriedem> there are libvirt tests that use a real connection
14:27:18 <dansmith> but they seem to be working, according to clarkb
14:27:19 <johnthetubaguy> I was going to ask, anyone remember finding bugs by running unit tests on two distros
14:27:40 <mriedem> no major concerns from me, if we hit racy things in the libvirt driver unit tests we'd have to work those out anyway
14:27:43 <mriedem> like yesterday
14:27:46 <dansmith> johnthetubaguy: sometimes, but probably not worth the cost of duplicate tests
14:27:59 <johnthetubaguy> #info no objections to running unit tests on trusty only in the meeting
14:28:03 <dansmith> okay, well, I'll give the green light then
14:28:09 <johnthetubaguy> dansmith: yeah, that seems fair
14:28:11 <johnthetubaguy> cools
14:28:27 <johnthetubaguy> nova network APIs: http://lists.openstack.org/pipermail/openstack-dev/2014-June/037707.html
14:28:32 <johnthetubaguy> although you probably covered that last time
14:29:03 <dansmith> my comment to that is "patches accepted"
14:29:04 <johnthetubaguy> request for rbd patch reviews
14:29:29 <johnthetubaguy> dansmith: right, I think one should be admin only, the other not
14:29:30 <johnthetubaguy> yep
14:29:50 <johnthetubaguy> so, adding a barbican dependency: https://review.openstack.org/#/c/94370/
14:30:20 <dansmith> barbican has been discussed, and I think we're okay with it for an optional feature
14:30:23 <dansmith> dogfooding goals, and all
14:30:24 <dansmith> right?
14:30:32 <johnthetubaguy> I was going to say, that sounds right
14:30:38 <johnthetubaguy> as long as its optional, cools
14:30:40 <dansmith> I know mikal and russellb are on board at least
14:30:41 <dansmith> yeah
14:30:57 <johnthetubaguy> I suspect that got left on the agenda
14:31:05 <johnthetubaguy> cool, so we are done in 30mins on the agenda
14:31:11 <johnthetubaguy> any more topics from people?
14:31:29 <johnthetubaguy> oh, one more from me
14:31:38 <johnthetubaguy> please register for the mid cycle meet up if you are going
14:32:02 <johnthetubaguy> #link please register https://www.eventbrite.com.au/e/openstack-nova-juno-mid-cycle-developer-meetup-tickets-11878128803
14:32:14 <johnthetubaguy> well, tumbleweed goes past
14:32:15 <sgordon> johnthetubaguy, i missed my queue!
14:32:16 <johnthetubaguy> so thanks all
14:32:19 <stpierre> johnthetubaguy: you suggested bringing up https://review.openstack.org/#/c/97909/ in the meeting, although i wasn't sure what specifically you had in mind to discuss.  sounds like you maybe wanted a broader discussion
14:32:23 <sgordon> on the NFV side we created a review dash: http://bit.ly/1iFdldx
14:32:36 <johnthetubaguy> sgordon: fire away
14:32:40 <sgordon> working on automating regeneration of it on a regular basis but the list is up to date as of yesterday
14:32:55 <sgordon> just if people are interested in following which proposals are believed to relate to that area
14:33:03 <bauzas> sgordon: well the list is up to date until the wiki page is updated :)
14:33:10 <sgordon> bauzas, right :)
14:33:22 <johnthetubaguy> OK thanks good to know
14:33:24 <bauzas> sgordon: so if we know the page has changed I can run the script by hand
14:33:33 <sgordon> bauzas, it hasn't - i set notifications on that page
14:33:45 <bauzas> sgordon: cool, will do too
14:34:05 * sgordon kicks the tumbleweed back into the street
14:34:06 <johnthetubaguy> stpierre: so, its a nova-network feature, I am just thinking you should try and follow neturon's lead rather than add something different if possible
14:34:34 <johnthetubaguy> stpierre: was curious what you where thinking about that comment? does that sound OK?
14:35:18 * johnthetubaguy is thinking he is only half connected to IRC at this point...
14:35:18 <stpierre> yeah, i looked into it a bit and i looks like neutron has support for arbitrary dhcp options on a per-port basis, which obviously isn't something we can do in nova
14:35:57 <stpierre> it also looks like the options are set in the database rather than in the config, which is definitely something we can do if we decide that we want per-network dnsmasq options instead of/in addition to global dnsmasq options
14:36:08 <stpierre> so neutron is similar but different, for good reasons
14:36:25 <johnthetubaguy> dansmith: curious about your thoughts here, I kinda feel like we want something in neutron that matches it first, before we take it in nova?
14:36:36 <johnthetubaguy> but maybe thats too inflexible
14:36:42 <dansmith> johnthetubaguy: I'll take a look
14:36:49 <johnthetubaguy> dansmith: thanks
14:36:54 <stpierre> ty both
14:37:05 <johnthetubaguy> stpierre: cool, we will try get back to you on that one
14:37:11 <johnthetubaguy> awesome, so we are all done now I guess
14:37:17 <dansmith> +1
14:37:23 <johnthetubaguy> you get 20 mins back of your day, with any luck...
14:37:27 <leifz> awesome
14:37:29 <johnthetubaguy> thanks all
14:37:31 <stpierre> i think this may just need to be a point of discontinuity between neutron and nova because of their different architectures and whatnot, but i'll let the experts decide :)
14:37:32 <johnthetubaguy> #endmeeting