#openstack-meeting log

21:02:04 <russellb> #startmeeting nova
21:02:05 <openstack> Meeting started Thu Mar 21 21:02:04 2013 UTC.  The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:02:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:02:10 <openstack> The meeting name has been set to 'nova'
21:02:11 <russellb> #chair vishy
21:02:12 <openstack> Current chairs: russellb vishy
21:02:21 <russellb> Who's around to talk about Nova?
21:02:25 * dansmith gurgles
21:02:36 <boris-42> Hi
21:02:39 <cyeoh> hi
21:02:58 <mikal> Ahoy
21:03:05 <sdague> o/
21:03:06 <alaski> hi
21:03:20 <russellb> 7 so far :)
21:03:24 <vishy> hi
21:03:29 <russellb> 8!
21:03:37 <russellb> #link https://wiki.openstack.org/wiki/Meetings/Nova
21:03:58 <russellb> #topic grizzly-rc1 and maybe rc2
21:04:11 <russellb> #link https://launchpad.net/nova/+milestone/grizzly-rc1
21:04:15 <russellb> 200 bugs closed in 4 weeks :)
21:04:21 <russellb> bugs fixed*
21:04:24 <mikal> That's pretty cool
21:04:33 <sdague> nice
21:04:34 <russellb> even more closed
21:04:53 <russellb> yeah, i thought that was pretty awesome
21:05:11 <mikal> Still a pretty big collection of ancient bugs to close though...
21:05:20 <russellb> There is the possibility of grizzly-rc2
21:05:30 <russellb> vishy pointed out this bug earlier today: https://bugs.launchpad.net/oslo/+bug/1158179
21:05:31 <uvirtbot> Launchpad bug 1158179 in oslo "Lockutils - deletes directory rendering locks invalid" [Undecided,In progress]
21:05:55 <russellb> so we should be on the lookout for other bugs that may be rc2 worthy
21:06:07 <boris-42> Btw
21:06:09 <boris-42> https://answers.launchpad.net/nova/+question/224818
21:06:40 <russellb> We have 24 new bugs, need to triage those, hopefully nothing release blocker worthy
21:06:41 <boris-42> I don't know is this only caused by wrong settings or something is not ok with service update method…
21:07:35 <boris-42> But I wrote tests for db.service methods recent.. so it should work as expected..
21:08:08 <russellb> grizzly-3?  ancient!
21:08:29 <russellb> anything else on grizzly rc status?
21:08:49 <russellb> #topic design summit
21:09:01 <russellb> #link http://summit.openstack.org
21:09:10 <russellb> we already have more proposals than time slots :-)
21:09:23 <russellb> official deadline is end of March, but if you know of something you want to propose, please do it as soon as you're able
21:09:35 <russellb> i'm trying to get as much of the planning done early as possible
21:09:50 <russellb> as i'll be on vacation for the beginning of april, for most of the time before summit
21:10:11 <russellb> also take a look over what's proposed and see what you think is missing
21:10:16 <russellb> as far as big topics that you'd expect to be discussed
21:10:47 <russellb> any questions / comments about design summit planning?
21:11:20 <russellb> alrighty
21:11:25 <russellb> #topic backportable db migrations
21:11:32 <russellb> #link http://lists.openstack.org/pipermail/openstack-dev/2013-March/006827.html
21:11:36 <sdague> only question, did you get quantum folks to buy into the cross sessions?
21:11:53 <russellb> sdague: yes, I spoke to Mark yesterday about it and he's totally on board
21:11:58 <sdague> cool
21:12:04 <russellb> our idea was to reserve 2 sessions for Nova + Quantum people
21:12:10 <russellb> one at the beginning of the week, and one near the end
21:12:38 <russellb> one to outline some goals, identify quantum as a priority for nova, discuss what it would take to become the default ...
21:12:55 <russellb> and then again to get back together, identify concrete tasks, and discuss anything that changed in our thinking as we went through the rest of our sessions
21:13:19 <russellb> hopefully we'll fit in a room :)
21:13:30 <russellb> sdague: how does that sound to you?
21:13:53 <sdague> russellb: sounds good.
21:13:58 <russellb> cool
21:14:24 <russellb> so on this db migrations thing ... i mainly wanted nova-core to be aware to make sure everyone kept it in mind when reviewing migrations
21:14:45 <russellb> the placeholders should not be replaced with a real migration in havana (i saw that in one patch today)
21:14:47 <mikal> We expect new migrations to consume one of those reserved numbers? Or to start after those?
21:14:53 <mikal> And the reseved ones are for grizzly?
21:14:53 <russellb> start after those
21:14:58 <russellb> correct
21:14:59 <mikal> Ahhh, ok, yeah
21:15:07 <mikal> There was a review which did the wrong thing yesterday then
21:15:16 <russellb> didn't go in did it?
21:15:32 <boris-42> russelb
21:15:39 <boris-42> Probably we should add big comment?
21:15:45 <boris-42> DO NOT REPLACE THIS?)
21:15:47 <mikal> russellb: no -- https://review.openstack.org/#/c/21185/
21:16:00 <russellb> boris-42: sure :-)
21:16:00 <mikal> boris-42: I like idiot proofing
21:16:03 <dripton> I'll add comments to the placeholder migrations to make it clear.
21:16:04 <russellb> mikal: k thx
21:16:13 <russellb> good idea
21:16:22 <boris-42> =)
21:16:31 <russellb> other thing to make sure of is that migrations are idempotent
21:16:38 <mikal> russellb: I'll make sure matiu knows to not replace that
21:16:46 <russellb> in theory that makes them safe to backport if someone chooses to ...
21:17:12 <russellb> and let this be motivation for someone to work on alembic again :-)
21:17:26 <dripton> We have test_migrations.py now which makes it somewhat easier to test idempotence.
21:17:26 <boris-42> I will be glad=)
21:17:30 * matiu reads back
21:17:35 <dripton> At least it tests that they go up and back down correctly.
21:17:36 <boris-42> to avoid sqlalchemy-migration at all=)
21:17:43 <dripton> If there's sufficient interest in Alembic for backportable migrations, then I will bring https://review.openstack.org/#/c/15196/ up to date.
21:17:47 <boris-42> sqlalchemy-migrate*
21:18:04 <sdague> except, doesn't alembic cause the other issue where you can't backport unless you backport exactly the same migration?
21:18:11 <sdague> I thought that came up in summit last time
21:18:16 <matiu> yeah I fixed that review already, I put the placeholder back :)
21:18:26 <sdague> or is the a concept of a merge migration to handle that?
21:18:38 <russellb> sdague: i don't know, i'd like to get some concrete info on how it would work in that case
21:19:07 <dripton> sdague: alembic supports arbitrary ordering of migrations (like git with commits), so you could theoretically rework the migration into almost-the-same for your backport, if needed.
21:19:46 <dripton> it's the static ordering that kills us with migrate.  As far as the APIs go I don't care.
21:19:59 <sdague> dripton: ok, well it's probably worth chatting about in portland
21:20:09 <russellb> would be good to write up some documentation with a proposal to migrate to it
21:20:39 <sdague> just to make sure all the edge cases are covered, like a long running grizzly tree with 5 backported migrations that aren't exactly the same as havana, and then trying to roll forward to havana release
21:20:40 <russellb> to cover the work needed to migrate, and how it would (or wouldn't) help backportability, etc.
21:20:57 <dripton> I'll write up a blueprint for switching to alembic.
21:21:03 <sdague> dripton: awesome
21:21:04 <russellb> that would be great
21:21:11 <russellb> blueprints also have a "link to specification"
21:21:17 <russellb> would be good to write up details on the wiki and link to it
21:21:40 <boris-42> dription +1
21:21:50 * russellb notes that quantum uses alembic
21:22:08 <russellb> so we wouldn't be breaking new ground for openstack in general
21:22:15 <dripton> but they just started using it so I don't think they have much to tell us.  It just means OpenStack doesn't gain another dependency.
21:22:19 <russellb> but they started fresh IIRC, as opposed to having to migrate to it
21:23:04 <sdague> yeh, the cuttover will be "fun" :)
21:23:23 <mikal> Well, we do love a good painful migration...
21:24:06 <russellb> cool, anything more on this topic?
21:24:28 <russellb> #topic review criteria
21:24:44 <russellb> So, nova-core is pretty big, and consistency in the way we review is something we should always be striving for
21:24:52 <russellb> so I think it's a good topic to bring up any week
21:25:08 <russellb> one thing I thought about today was around OpenStack API changes
21:25:14 <russellb> we discussed before about requiring API samples, which is great
21:25:25 <russellb> i was thinking we should also require a corresponding novaclient patch be ready to go
21:26:13 <dripton> How does that work with gerrit since nova and novaclient are in different repos?
21:26:14 <russellb> seem reasonable?
21:26:18 <sdague> sounds reasonable, might make people look at novaclient more often if nothing else
21:26:19 <russellb> just convention
21:26:25 <sdague> dripton: yeh, just like adding tempest tests
21:26:29 <alaski> I like it, though I'm one of the recent offenders
21:26:33 <sdague> you can't enforce it on a single commit
21:26:39 <russellb> alaski: cool ;-)
21:26:42 <sdague> but you can shame people if they don't do it
21:26:48 <dripton> ok, we're good at shame
21:27:25 <sdague> well, we can make it possitive and have mikal hand out cookies :)
21:27:32 <sdague> for those that do it right
21:27:38 <mikal> We certainly need to get better at paying attention to novaclient
21:27:39 <russellb> and gold stars!
21:27:44 <mikal> sdague: I can do that...
21:28:04 <russellb> so i just have a review queue bookmark that shows me all the stuff i want to be reviewing regularly
21:28:12 <russellb> but to each their own, whatever works for you
21:28:13 <sdague> I wonder if there is a way to trick gerrit into showing novaclient as well when you click nova in the ui
21:28:23 <russellb> but using that bookmark the last few weeks has been very helpful
21:28:27 <mikal> There is a project search param
21:28:28 <dripton> russellb: can you paste your bookmark again (mine is probably worse)
21:28:36 <mikal> But I'm not sure it can do two projects at once
21:28:46 <russellb> https://review.openstack.org/#/q/status:open+(project:openstack/nova+OR+project:openstack/python-novaclient+OR+project:openstack/oslo-incubator+OR+project:openstack/oslo.config)+(branch:master+OR+branch:stable/grizzly),n,z
21:28:53 <russellb> that's what i'm using ... you can edit to your liking
21:28:57 <mikal> Oh, nice
21:29:04 <mikal> That's better than mine
21:29:18 <dripton> better than mine too.  Stealing.
21:29:23 <mikal> Heh
21:29:29 <russellb> heh, steal away
21:29:34 <sdague> russellb: that should go somewhere, it's super useful
21:29:45 <russellb> sdague: it's going in the meeting log!
21:29:54 <russellb> sdague: but yeah, dunno where to put it
21:29:59 <dripton> wiki page?
21:30:04 <russellb> which one
21:30:10 <russellb> ReviewWorkflowTips
21:30:20 <russellb> i just made that up.
21:31:15 <russellb> #topic open discussion
21:31:19 <russellb> that's all i had
21:31:26 <sdague> #link https://review.openstack.org/#/q/status:open+(project:openstack/nova+OR+project:openstack/python-novaclient+OR+project:openstack/oslo-incubator+OR+project:openstack/oslo.config)+(branch:master+OR+branch:stable/grizzly),n,z
21:31:31 <mikal> I want to know if anyone is using moniker at all
21:31:34 <russellb> sdague: ah good idea :)
21:31:35 <sdague> now it will be in the highlights :)
21:31:37 <mikal> Given I discovered it this morning
21:31:59 <mikal> And if people are, should I rip the DNS driver layer out of nova-network
21:32:10 <mikal> Especially given it only supports LDAP backed DNS
21:32:21 <dripton> https://wiki.openstack.org/wiki/ReviewWorkflowTips exists now
21:32:33 <russellb> i don't think we should be considering ripping anything out in favor of something that's not even incubated yet
21:32:55 <mikal> russellb: I think that depends on if _anyone_ uses the current implementation
21:32:57 <russellb> and i think it'd be worth asking the quantum folks what they think about the direction of DNS support
21:33:01 <mikal> I am accussing it of being dead code
21:33:08 <russellb> heh
21:33:16 <russellb> presumably who added it uses it?
21:33:28 <mikal> That would be wikimedia
21:33:37 * devananda shows up late (packing for a 10-day trip), reads scrollback
21:33:50 <russellb> devananda: we signed you up to fix everything
21:33:54 <mikal> Heh
21:34:01 <sdague> any idea if monkier is applying for incubation?
21:34:01 <devananda> russellb: perfect. I'll have plenty of time on the plane :)
21:34:13 <sdague> just to even know if it's fruitful
21:34:16 <russellb> devananda: enjoy your trip :)
21:34:18 <dripton> devananda: if you can fix all 400 bugs on the plane trip, I'll buy your plane ticket.
21:34:18 <mikal> They're on stackforge, which I think means yes?
21:34:40 <russellb> mikal: recent activity?
21:35:00 <mikal> March was quiet for them
21:35:09 <mikal> There's basically one developer as well, which is a concern
21:35:44 <mikal> For people's benefit, the back story is I'm sitting one a dyndns driver for the dns layer in nova-network
21:35:45 <russellb> reach out to the dev, ask him how he views the status/future
21:35:53 <mikal> And I'm trying to decide if I send it off for review, or redo with moniker
21:37:23 <sdague> mikal: yeh, one dev doesn't make an incubated project
21:37:29 <russellb> mikal: meh, send for review, you've already done the work
21:37:38 <mikal> russellb: fair enough
21:37:45 <mikal> russellb: I'm annoyed with myself mostly
21:37:54 <mikal> sdague: well, I could fix that by helping him...
21:38:01 <russellb> may have to get redone someday for quantum, if that stuff ends up in quantum
21:38:12 <dripton> there are actually at least 3 devs. It's just 1 has done most of the recent work.
21:38:13 <mikal> The other factor is moniker is biggish for what it does, and another daemon to setup
21:38:20 <mikal> So maybe that makes having a DNS driver in nova more exciting
21:38:32 <sdague> or quantum?
21:38:35 <mikal> Top authors (mikal is 0 lines):
21:38:37 <mikal> 4: andrewbogott
21:38:37 <mikal> 10: Davide Guerri
21:38:37 <mikal> 124: Andrey Perminov
21:38:37 <mikal> 154: JC Martin
21:38:39 <mikal> 1713: Patrick Galbraith
21:38:42 <mikal> 6690: Endre Karlson
21:38:44 <mikal> 30395: Kiall Mac Innes
21:38:47 <mikal> (Says my crappy script)
21:38:52 <mikal> So its mostly one person
21:39:03 <sdague> what about by changesets?
21:39:12 <sdague> lines are always skewed
21:39:20 <russellb> how about email openstack-dev and ask about project status/future?
21:39:23 <russellb> we're just guessing here
21:39:28 <dripton> +1
21:39:35 <mikal> Yeah, I'll send an email
21:39:39 <russellb> k
21:39:45 <mikal> I might as -operators if anyone has the current DNS driver deployed as well
21:39:55 <russellb> cool
21:40:13 <mikal> sdague: 309 changesets from Kiall, the next closest is Endre with 30.
21:40:40 <mikal> So yeah, this is mostly me picking y'alls brains
21:40:52 <mikal> It seems there's some hesitation with moniker, so I'll go and ask questions
21:40:56 <russellb> ow that hurts
21:41:14 <russellb> any other comments / questions / rants ?
21:43:08 <russellb> allllllrighty then, thanks!
21:43:24 <russellb> #endmeeting