21:00:28 <russellb> #startmeeting nova
21:00:29 <openstack> Meeting started Thu Dec  6 21:00:28 2012 UTC.  The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:32 <openstack> The meeting name has been set to 'nova'
21:00:40 <russellb> Greetings, all.
21:00:43 <russellb> who's around
21:00:44 <dansmith> <- here, with really bad bug stats, fair warning.
21:00:58 <dprince> hi
21:01:00 <comstud> here, but sick.
21:01:02 <markmc> hey
21:01:09 <russellb> comstud: bummer, sorry :(
21:01:14 <russellb> #link http://wiki.openstack.org/Meetings/Nova
21:01:25 <comstud> ya, sux.. i'm out of town too
21:01:32 <russellb> #topic project updates
21:01:34 <mikal> Ahoy!
21:01:35 <sdague> <- here
21:01:41 <russellb> comstud: let's hit cells first
21:01:56 <comstud> ok
21:01:56 <russellb> comstud: what's the latest on cells
21:02:13 <comstud> so close to done, but.. i've done nothing but slept the last 2 days :)
21:02:21 <russellb> yeah, priorities :)
21:02:26 <comstud> i mean, done reworking
21:02:31 <comstud> but it's looking much better to me
21:02:42 <russellb> great
21:02:51 <russellb> grizzly-2 is January 10th, i think we should aim for that
21:02:55 <comstud> tests might be a pain to re-work
21:03:01 <russellb> still a good amount of time, but holidays will eat some of that
21:03:06 <comstud> yea
21:03:10 <comstud> i'm not going anywhere
21:03:13 <comstud> but i might not get reviews
21:03:15 <comstud> :)
21:03:26 <russellb> heh, i'm not going anywhere either, but technically i'll be off
21:03:30 <russellb> i can review anyway
21:03:45 <comstud> so dedicated
21:03:54 <mikal> I'll be around too, not that I've reviewed much cells stuff. The 3,500 line one scares me.
21:03:59 <russellb> and they don't have christmas in the southern hemisphere, so mikal can review too
21:04:05 <mikal> :P
21:04:21 <comstud> the 3500 line one may be more lines when i'm done here
21:04:23 <dansmith> is that because the trees don't stand upside down well?
21:04:23 <comstud> but easier to follow
21:04:27 <comstud> not sure, haven't diffed it yet
21:04:38 <comstud> but hey, that one started out at 4500 i think
21:04:45 <russellb> comstud: cool, i saw you have a lot of reviews on later patches
21:04:46 <mikal> dansmith: don't make me kangaroo you up the billabong
21:04:49 <russellb> some approved already, which is kind of odd
21:04:51 <comstud> yeah
21:04:54 <comstud> haha yeah
21:05:00 <russellb> not sure how you can approve something that depends on something you didn't review :)
21:05:10 <dansmith> mikal: heh, you all have kangaroos as pets, right?
21:05:18 <comstud> i wasn't going to complain about it
21:05:19 <russellb> so next week updated patches perhaps?  if you're feeling better?
21:05:25 <russellb> comstud: heh, i wouldn't either..
21:05:27 <comstud> yeah
21:05:35 <russellb> ok, cool, anything else you want to discuss?
21:05:55 <comstud> nothing on my mind other than baremetal has caused a lot of pain :)
21:05:58 <comstud> bugs
21:06:04 <russellb> aw
21:06:06 <comstud> but that's just a comment.
21:06:12 <russellb> that's the other big project we've been hitting in this meeting
21:06:16 <comstud> yeah
21:06:23 <comstud> continue on
21:06:24 <russellb> anyone care to comment on the status there?
21:06:25 <comstud> i'm good
21:06:33 <dansmith> I'm afraid to go near baremetal because most of them have one green check
21:06:33 <russellb> comstud: what kind of bugs have hit you?
21:06:39 <russellb> so we know what to watch out for better
21:07:26 <russellb> in the context of cells, or you mean in general?
21:07:43 <russellb> i remember a resource tracker issue, delaying setting 'node' until after resources were claimed
21:08:38 <russellb> oookay.  :)
21:09:13 <russellb> so couple other specific projects ...
21:09:18 <russellb> boot from volume improvements
21:09:24 <comstud> russellb: reisze stuff and resource tracker stuff
21:09:29 <russellb> comstud: ah ok
21:09:29 <comstud> both might hav been resource tracker
21:09:33 <comstud> basically...
21:09:36 <comstud> has to do with adding of 'node'
21:09:38 <russellb> sorry :)
21:09:39 <comstud> to compute_nodes
21:09:41 <russellb> right
21:10:01 <russellb> so boot from volume ... https://blueprints.launchpad.net/nova/+spec/improve-boot-from-volume
21:10:04 <russellb> #link https://blueprints.launchpad.net/nova/+spec/improve-boot-from-volume
21:10:10 <russellb> nikola wanted to discuss it, but i don't see him here
21:10:20 <russellb> probably because his first big patch has been merged since asking for it to be on the agenda
21:10:25 <russellb> but i think there is still more work planned there
21:10:33 <russellb> so, if you're interested take a look and respond to the ML thread about it
21:10:40 <comstud> merge and flee
21:10:46 <russellb> heh
21:10:52 <comstud> (that's what i'm going to do when cells lands)
21:10:59 <russellb> alexpilotti requested discussing get-password (vishy's branch)
21:10:59 <dansmith> drive-by patching?
21:11:02 <comstud> yeah
21:11:06 <russellb> but since vishy is out today, let's push it to next week
21:11:12 <alexpilotti> fine
21:11:34 <russellb> or you can always post to the ML in the meantime
21:11:43 <russellb> lots of stuff in progress on https://launchpad.net/nova/+milestone/grizzly-2
21:12:06 <russellb> only 2 not started
21:12:11 <russellb> 1 is a lie, because i started it this week
21:12:22 <russellb> i think the block device handling one may still be looking for a volunteer
21:12:32 <russellb> so, take a look if you're not busy enough already :)
21:12:57 <russellb> #topic nova bugs
21:13:12 <russellb> #link http://webnumbr.com/untouched-nova-bugs
21:13:15 <russellb> 27
21:13:19 <mikal> http://www.stillhq.com/openstack/nova/triage-20121207.txt
21:13:19 <russellb> about where we've been hovering
21:13:35 * russellb awards mikal with the nova-bugs gold star of the week
21:13:37 <markmc> ah come on, I triaged 2!
21:13:39 <russellb> hurrah mikal
21:13:46 <mikal> LOL
21:13:49 <markmc> before making the tragic mistaking of fixing one :)
21:13:50 <russellb> markmc: bug in the script!
21:13:52 <russellb> it's rigged!
21:13:55 <markmc> which ended the triaging session
21:13:56 <mikal> There are a bunch of bugs I haven't had a chance to get to
21:14:08 <sdague> heh
21:14:11 <dansmith> russellb: I have something for grizzly-2 I think,
21:14:15 <mikal> russellb: its on github! Fix it!
21:14:27 <russellb> dansmith: ok one sec
21:14:34 <dansmith> russellb: would it be okay if we have a blueprint for removing all direct db calls from manager? presumably to be complete around then...
21:14:45 <dprince> man. I feel like a slacker.
21:14:55 <russellb> dansmith: oh, um ... yeah, as a child blueprint under no-db-compute?
21:14:59 <russellb> dansmith: that makes sense.
21:15:13 <russellb> dansmith: that's going to be tight, but a good goal, and we better make it if we want to finish by grizzly-3 ...
21:15:16 <dansmith> russellb: just trying to have enough steps in the process so we can point to public blueprints to show status
21:15:18 <sdague> on the bugs topic, https://bugs.launchpad.net/nova/+bug/1079687 is the last thing which is blocking tempest gate on nova. And given that a nova bug that tempest would have caught slipped in last week, it would be nice to get more eyes on it.
21:15:19 <uvirtbot> Launchpad bug 1079687 in nova "Flaky failures of instances to reach BUILD and ACTIVE states" [Undecided,New]
21:15:25 <russellb> dansmith: good idea, file away
21:15:29 <dansmith> russellb: okay, thanks
21:15:34 <russellb> dansmith: and link me so i can press buttons
21:16:02 <russellb> i hate "flaky" bugs, heh
21:16:21 <sdague> yes, agreed
21:16:32 <dansmith> russellb: do I need to do something to make it a proper child?
21:16:39 <sdague> it appears nova has to be a very specific kind of slow to trigger it
21:16:39 <dansmith> no-db-virt never got to be related in that way
21:16:54 <russellb> dansmith: there is a dependencies section somewhere ... down lower on the page once it's filed I think
21:16:59 <dansmith> okay
21:17:14 <russellb> may have to be done from no-db-compute?  i can't remember
21:17:55 <russellb> #help need eyes on this bug, the last thing blocking the tempest gate on nova - https://bugs.launchpad.net/nova/+bug/1079687
21:17:56 <uvirtbot> Launchpad bug 1079687 in nova "Flaky failures of instances to reach BUILD and ACTIVE states" [Undecided,New]
21:18:12 <dansmith> https://blueprints.launchpad.net/nova/+spec/no-db-compute-manager
21:18:28 <russellb> dansmith: thanks, i'll get it in grizzly-2 after meeting
21:18:44 <dansmith> russellb: I think you need to edit the deps on no-db-compute, because I can edit deps for the one I filed, but not the main one
21:18:49 <dansmith> russellb: and if you do, can you add no-db-virt as well?
21:18:53 <russellb> sdague: let's push through remaining discussion, and if there's time left we can group-analyze the bug some :)
21:19:01 <russellb> dansmith: ack, fine with me
21:19:06 <sdague> I'm fine with that
21:19:07 <dansmith> thanks
21:19:08 <russellb> anything else on bugs?
21:19:30 <markmc> bug day next week?
21:19:34 <markmc> squashing day
21:19:40 <mikal> That sounds like a good idea
21:19:43 <dansmith> dang, I was hoping for a bug-writing day
21:19:46 <russellb> yeah, i like that
21:20:00 <markmc> I think ttx is organizing one
21:20:01 <sdague> sure, we're due for one
21:20:23 <russellb> markmc: oh?  ok great ... yeah, we've got a bunch we can work through
21:20:44 <russellb> i'm seeing good legit stuff every time i go triage some
21:20:49 <markmc> http://eavesdrop.openstack.org/meetings/project/2012/project.2012-12-04-21.02.html
21:20:56 <markmc> #info Doc Bug day December 6, Regular Bug day December 13
21:21:00 <markmc> #link Doc Bug day December 6, Regular Bug day December 13
21:21:06 <markmc> #undo
21:21:11 <markmc> #link http://eavesdrop.openstack.org/meetings/project/2012/project.2012-12-04-21.02.html
21:21:35 <russellb> nice, so we can check out our progress in the next nova meeting :)
21:22:13 <russellb> so everyone put it on your calendar to block out unnecessary meetings :)
21:22:19 <sdague> and mikal will be asleep while the rest of us are working, so someone might beat him
21:22:35 <mikal> Heh
21:23:03 <russellb> so, as usual, if everyone could triage something this week, that'd be a great help
21:23:07 <russellb> still not enough names on the bug triage reports
21:23:21 <russellb> especially important this week if we're going to squash next thursday
21:23:29 <russellb> the better the bug list, the more productive we can be
21:23:52 <russellb> and with that ...
21:23:53 <comstud> yes sir
21:23:53 <russellb> #topic open discussion
21:24:41 <russellb> open discussion until we have nothing left, and then whoever wants to stick around, we can spend some time trying to troubleshoot this tempest gate blocker
21:25:21 <russellb> awesome.
21:25:33 <russellb> everyone is happy!
21:25:36 <dprince> what is the tempest gate blocker?
21:25:37 <comstud> welllll
21:26:02 <dprince> I'm not. Python-keystoneclient ruined my day.
21:26:04 <russellb> dprince: sdague brought up this bug earlier, said it was the last bug blocking being able to gate on the whole tempest suite - https://bugs.launchpad.net/nova/+bug/1079687
21:26:06 <uvirtbot> Launchpad bug 1079687 in nova "Flaky failures of instances to reach BUILD and ACTIVE states" [Undecided,New]
21:26:08 <russellb> aw
21:26:26 <comstud> i don't understand 'to reach BUILD' part of that
21:26:31 <comstud> instances are created in BUILD state.
21:26:54 <sdague> dprince: it's a race in tempest which gets triggered by exactly how slow the ci nodes are about 15% of the time
21:27:23 <dprince> sdague: I run similar tests on the hour on bare metal and get fairly consistent results.
21:27:28 <dprince> sdague: passes.
21:27:50 <sdague> davidkranz has been running hourly gate jobs on ci, and it fails about 15% of the time
21:27:51 <dprince> sdague: How do we know this isn't some sort of IO bound... public cloud type failure?
21:28:22 <sdague> dprince: we don't, however, we can't light the gate on CI if it isn't repeatable on CI
21:28:35 <dprince> sdague: I'm happy to take a closer look at it offline.
21:29:11 <sdague> cool
21:29:23 <russellb> thanks dprince
21:29:59 <dprince> Yep. Sunil might also be able to help look into this as well (I'll point it out to him here shortly)
21:30:51 <russellb> well, sounds like we have some eyes on it then ...
21:31:10 <russellb> anything else before we #endmeeting?
21:31:37 <dansmith> do it!
21:31:40 <russellb> anyone want to shamelessly plug their review of choice?
21:32:03 <sdague> oh mtreinish's coverage patch should get a review
21:32:05 <dprince> what about ones we don't like?
21:32:15 <russellb> dprince: that's fair too
21:32:26 <sdague> #link https://review.openstack.org/#/c/14468/
21:32:30 <sdague> plug for review
21:32:31 <russellb> dprince: you can have on the record that you really don't like something :)
21:32:47 <dansmith> now I'm curious...
21:32:48 <russellb> ah yes, that's a good one
21:32:59 <mtreinish> sdague: thanks for the plug
21:33:09 <mtreinish> too fast
21:33:12 <dprince> I'll stay quiet... :) for now.
21:33:19 <russellb> dprince: oh come onnnn
21:33:22 <russellb> stir the pot
21:33:24 <sdague> dprince: plug for -2 ?
21:33:41 <dansmith> dprince: come on, I wanna see if it's the same one I'm thinking of :)
21:34:03 <dprince> shhhh.
21:34:10 <dprince> it is a secret
21:34:16 <russellb> heh, ok..
21:34:31 <sdague> I'd also plug clarb's testr - https://review.openstack.org/#/c/15078/
21:34:42 <sdague> which is close, he might have the last fix in a different review now
21:34:43 <dansmith> sdague: you've had enough
21:34:55 <sdague> hey, I like running the unit tests in 42 seconds
21:35:13 <russellb> sdague: +1
21:35:31 <russellb> i'm very happy about that ...
21:35:34 <russellb> the time part anyway
21:35:37 <comstud> i don't like that our clients do caching and use keyring by default.
21:35:44 <comstud> it's extremely annoying and has bitten me a number of times.
21:36:06 <comstud> (just noticing some of dprince's patches to fix that somewhat for keystoneclient)
21:36:27 <sdague> actually, specifically other eyes on this one - https://review.openstack.org/#/c/17639/ which is the fix, would be good
21:36:33 <sdague> then the testr one should pass
21:36:43 <comstud> was trying to find a patch of dprince's to pick on
21:36:46 <comstud> but can't find one :-/
21:36:55 <comstud> :)
21:37:00 <dprince> hold on... I'l give them to you:
21:37:05 <dprince> https://review.openstack.org/#/c/17630/
21:37:11 <dprince> https://review.openstack.org/#/c/17634/
21:37:27 <comstud> oh no, i found those
21:37:28 <dprince> And... FTR that is why python-keystoneclient ruined my day.
21:37:30 <comstud> i meant i can't pick on them
21:37:40 <comstud> because i agree
21:39:15 <comstud> i want novaclient to have cache/keyring off by default.
21:39:16 <russellb> alright, well guess we're wrapped up
21:39:18 <comstud> :-/
21:39:20 <russellb> or not
21:39:33 <comstud> it's not important
21:39:47 <russellb> comstud: propose it and see what happens?  :)
21:39:55 <comstud> hehe ya
21:40:26 <russellb> we can continue ranting over in -nova :)
21:40:29 <russellb> thanks for coming all
21:40:31 <russellb> #endmeeting