21:00:34 <russellb> #startmeeting nova
21:00:35 <openstack> Meeting started Thu Jul 11 21:00:34 2013 UTC.  The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:38 <openstack> The meeting name has been set to 'nova'
21:00:42 <russellb> hello, everyone
21:00:44 <russellb> who's here?
21:00:49 <dansmith> \o
21:00:50 <dripton> hi
21:00:52 * johnthetubaguy waves
21:00:53 <cyeoh> hi
21:00:54 <n0ano> o/
21:00:55 <sdague> o/
21:00:55 <bnemec> o/
21:00:56 <mikal> Morning
21:00:57 <mrodden> heyo
21:01:06 <russellb> #link http://wiki.openstack.org/Meetings/Nova
21:01:13 <russellb> #topic havana-2
21:01:19 <russellb> havana-2 release is next week
21:01:20 <alaski> o/
21:01:29 <russellb> merge deadline is EOD Tuesday, July 16
21:01:42 <russellb> I've already deferred a *bunch* of stuff to havana-3 that didn't look like it could make it
21:01:47 <russellb> let's review the list of what's left
21:01:56 <russellb> #link https://launchpad.net/nova/+milestone/havana-2
21:02:00 <harlowja> \o/
21:02:12 <russellb> first up, let's cover v3 API work
21:02:14 <russellb> cyeoh: hi!
21:02:25 <cyeoh> heh ok.
21:02:28 <russellb> a lot has merged in the last week
21:02:31 <devananda> \o
21:02:36 <cyeoh> so we have about 40 or so extensions merged so far
21:02:46 <cyeoh> around 15-20 more to go
21:03:05 <cyeoh> I don't think all will make it in - merge conflicts and also we need a newer version of stevedore for 1 of them
21:03:07 <russellb> ok, so of these blueprints, are any that aren't marked as Implemented feasible for h2?
21:03:19 <cyeoh> but they should all flow in not long after
21:04:05 <russellb> as in, which ones should we go ahead and set for havana-3?
21:04:17 <cyeoh> sorry just doing a quick check
21:04:21 <russellb> i guess about all of them
21:04:26 <russellb> except maybe "convert core API" ?
21:04:30 <cyeoh> https://blueprints.launchpad.net/nova/+spec/v3-api-core-as-extensions might actually be done
21:04:37 <cyeoh> there is one more to go and that's it
21:04:37 <russellb> nice
21:04:44 <russellb> ok, so let's keep that one in h2
21:04:45 <russellb> and move the rest
21:04:57 <russellb> can you update them?
21:05:07 <cyeoh> yea the rest are all pretty much dependent on all the extensions being merged before considered done
21:05:09 <cyeoh> ok, will do
21:05:12 <russellb> yeah makes sense
21:05:15 <russellb> really nice progress though
21:05:32 <russellb> anything else on v3 API you'd like to discuss?
21:05:36 <cyeoh> thx - have had lots of help getting them merged - thank you to all who have pushed them through with all the merge conflicts...
21:05:49 <cyeoh> nothing else on v3 API.
21:05:52 <russellb> sure np, and now you're on the hook to help push the ones through that others write :)
21:05:59 <cyeoh> :-)
21:06:08 <russellb> k
21:06:19 <russellb> alaski: query-scheduler
21:06:34 <russellb> alaski: looks like some core bits are in, but not the part where we change the main instance boot code path to use it
21:06:44 <alaski> Right.
21:06:51 <alaski> From a dependency standpoint it's not in the way
21:07:03 <alaski> but there's a little more to do, which probably won't make h-2
21:07:21 <russellb> ok, wanted to clarify that part was in scope for this blueprint
21:07:22 <russellb> sounds like it is
21:07:25 <russellb> and won't make it
21:07:27 <russellb> so defer to h-3?
21:07:46 <alaski> yep.  That part turned out to be more nuanced than I had hoped.
21:08:12 <russellb> cool, there are often dragons when you go trying to redo something that has been there for ages :)
21:08:35 <russellb> dansmith: compute-api-objects
21:08:46 <russellb> dansmith: have a chance for h-2?
21:08:46 <dansmith> russellb: not gonna make h2,
21:08:50 <russellb> ack, will defer
21:08:53 <dansmith> comstud: disagree?
21:09:03 <comstud> no
21:09:15 <dansmith> I wish it were different, but, it ain't.
21:09:16 <comstud> Too many methods left
21:09:29 <russellb> boris-42: db-api-tests
21:09:37 <russellb> boris-42: looks really close to done
21:10:25 <russellb> according to status on the blueprint, only a few spots left to cover, and 2/3 have patches up for review
21:10:36 <russellb> so we'll leave this on for h-2 and hope we can get it done by the deadline
21:10:38 <russellb> boris-42: sound good?
21:10:46 <russellb> boris may be asleep.  :-)
21:11:11 <dripton> boris-42 would ask everyone to please review those patches.
21:11:28 <mikal> Heh, so many db reviews
21:11:44 <russellb> next is cross-service-request-id
21:11:47 <russellb> anyone know status of that one?
21:12:15 <russellb> has 3 +2s and a -2
21:12:18 <devananda> i blocked one of the pacthes for it
21:12:32 <devananda> because it exposed the cross serive id in a way that any user could set any value for it
21:12:39 <devananda> and tht would get passed all over the place
21:12:45 <devananda> kinda defeats the point, IMO
21:13:03 <devananda> i really like the concept, just not that it's exposed in that way
21:13:18 <russellb> yeah, seems like users should have no ability to impact
21:13:21 <russellb> and no knowledge of
21:13:25 <devananda> right
21:13:27 <russellb> just an admin thing
21:13:42 <russellb> ok, well, if it's blocked by a fundamental design issue, i'm deferring it
21:14:00 <johnthetubaguy> seems fair
21:14:36 <russellb> alaski: shelve-instance
21:14:46 <russellb> just needs review?
21:15:04 <alaski> yes.  two need rebase once the first is in, but everythings up
21:15:50 <russellb> ok cool
21:15:55 <russellb> need help reviewing that one then!
21:16:09 <russellb> Vek has eliminate-clear-passwords-from-cells-table
21:16:13 <russellb> I think that one is close to ready
21:16:22 <russellb> i had a problem with the data migration, but we've talked about it and he should have it updated soon
21:16:26 <johnthetubaguy> alaski: will take a closer look tomorrow
21:16:34 <russellb> johnthetubaguy: awesome thanks!
21:16:38 <alaski> johnthetubaguy: thanks.
21:16:55 <russellb> next is ec2-error-codes
21:16:59 <russellb> don't think the author is around
21:17:13 <russellb> I -2'd that one, because it looked like it needed to be broken up a bunch, but that was after a quick review
21:17:22 <russellb> he basically asked me to reconsider, and I haven't had a chance yet
21:17:27 <russellb> so I would appreciate a 2nd opinion
21:17:35 <dansmith> link?
21:17:56 <russellb> https://review.openstack.org/#/c/28330/
21:18:11 <russellb> probably need to give him guidance on how to break it up, that was going to be my response when i get back to it
21:18:19 <russellb> unless someone else looks and tells me they agree with him that it's fine as-is
21:18:47 <dansmith> I'll have a look when we're done
21:18:52 <russellb> ok cool thanks
21:19:02 <mrodden> yeah i think that definitely needs to be broken up...
21:19:10 <johnthetubaguy> at first glance it looks huge
21:19:11 <russellb> mrodden: ok, feedback appreciated on the review then
21:19:18 <boris-42> russellb I am here
21:19:21 <russellb> ok last medium one is from johnthetubaguy
21:19:32 <johnthetubaguy> ah, its in review
21:19:38 <russellb> xenapi large ephemeral disks
21:19:43 <russellb> so just need review?
21:19:49 <johnthetubaguy> got almost approved, just added an extra log
21:19:51 <johnthetubaguy> yeah
21:20:04 <russellb> ok cool, so hopefully we can knock that out in the next few days.
21:20:12 <johnthetubaguy> hopefully
21:20:22 <russellb> the rest still on the list are Low priority ones, anyone around that wants to talk about them?
21:20:44 <russellb> Low priority ones aren't worried about quite as much from a release management perspective, just nice to haves
21:20:56 <russellb> if you have time to help review stuff, please focus on the > Low priority stuff first
21:21:26 <russellb> boris-42: i was reviewing blueprints, covered db-api-tests, looks like there's a good chance we can have it completed for h-2, correct?
21:21:36 <boris-42> yes correct
21:21:45 <boris-42> I will force guys to finish tomorrow all patches
21:21:57 <mikal> boris-42: that sounds... concerning
21:21:59 <boris-42> if instance_* tests will be merged
21:22:09 <boris-42> then there is just a little bit work
21:22:24 <boris-42> I think we will finish before you wake up-)
21:22:34 <russellb> mikal: lol, the forcing part?
21:22:38 <johnthetubaguy> BTW, anyone know why http://status.openstack.org/reviews/ is not refreshing, maybe I need to check my email?
21:22:50 <mikal> russellb: yeah, I imagine a drum being beaten at the back of the room or something
21:23:08 <russellb> mikal: i imagine boris-42 as a russian mob boss
21:23:23 <russellb> "you WILL finish these patches.... or else"
21:23:41 <russellb> johnthetubaguy: i do not ... i would ask in -infra.  i haven't seen any email about it
21:24:09 <boris-42> =))
21:24:12 <boris-42> lol
21:24:28 <johnthetubaguy> russellb: yeah, I should head over there
21:24:31 <russellb> ok
21:24:34 <russellb> so last comment before we move on
21:24:39 <mrodden> user-locale-api is still waiting for reviews, but it also has a prereq. that I am waiting to be merged in oslo, which is actually +2CR right now just waiting for an approval to throw it to Zuul for merge
21:24:55 <russellb> mrodden: ok, will leave it on the list for now then
21:25:02 <russellb> so ... havana-3 is going to be insane
21:25:04 <russellb> https://launchpad.net/nova/+milestone/havana-3
21:25:14 <boris-42> pci passthrough ....
21:25:18 <boris-42> ouch
21:25:24 <russellb> if you want your stuff to actually land, PLEASE have it up for review as early as possible
21:25:26 <boris-42> it makes m cry
21:25:47 <russellb> because if even half of that gets implemented, that's going to be pushing our review bandwidth
21:26:07 <russellb> it also means prioritizing reviews is very important, so we focus on getting the most important stuff in
21:26:11 <boris-42> there are 2 implementations my and Inetels based on my patch
21:26:14 <boris-42> DB patch
21:26:32 <boris-42> that should be rewritten in unified models way
21:26:42 <russellb> your db patch should be rewritten?
21:26:50 <russellb> yes, please use the unified-objects stuff
21:27:02 <russellb> speaking of which, we need to start blocking patches that should be using those, but aren't
21:27:10 <russellb> basically, creating new DB models and then exposing them to the rest of the code
21:27:14 <boris-42> absolutely agree
21:27:14 <russellb> needs to be done using the new way.
21:27:30 <boris-42> but this patch is much more older then unified models=)
21:27:38 <boris-42> so it use old way=)
21:27:42 <russellb> i -2'd another that is in the same situation
21:27:51 <russellb> but we really can't afford going backwards on this effort
21:28:07 <boris-42> I agree
21:28:09 <russellb> it's critical for one of the highest priority efforts we have - the long road to solid support for live rolling upgrades
21:28:16 <boris-42> I agree lol=)
21:28:28 <johnthetubaguy> same for v3 things, all changes to both apis?
21:28:28 <russellb> saying it so everyone else understands too
21:28:28 <boris-42> We will help also with this stuff in I cycle
21:28:37 <comstud> Yea, I kinda crapped all over boris's reviews, unfortunately.
21:28:43 <russellb> johnthetubaguy: that's a good point ... we should probably start forcing v3 versions as well
21:29:06 <dansmith> +1 on v3
21:29:07 <boris-42> comstud it's ok
21:29:08 <johnthetubaguy> yeah, makes sense, for the same reasons
21:29:13 <russellb> cool
21:29:15 <cyeoh> yes, definitely for v3
21:29:31 <russellb> alright, onward
21:29:33 <boris-42> ok I have to sleep+)
21:29:33 <russellb> #topic bugs
21:29:38 <russellb> any particular bugs we need to discuss
21:29:41 <russellb> boris-42: ok thanks!
21:30:00 <russellb> 67 new bugs, a little more than i'd like
21:30:11 <boris-42> russellb do you need some help?)
21:30:15 <boris-42> with bugs?)
21:30:20 <russellb> boris-42: with bug triage, always
21:30:22 <russellb> and fixing them :)
21:30:36 <russellb> https://wiki.openstack.org/wiki/Nova/BugTriage
21:30:40 <boris-42> russelb ok I add some rule about fixing bugs=)
21:30:53 <boris-42> minimum 1 bug / week
21:30:56 <russellb> lol
21:30:59 <russellb> sounds ok to me
21:31:02 <boris-42> for each developer=)
21:31:05 <bnemec> I started triaging - feel free to yell at me if I confirmed anything I shouldn't have. :-)
21:31:08 <dripton> Fixing bugs you created yourself doesn't count
21:31:18 <boris-42> dripton ok=)
21:31:38 <boris-42> dription I will say them=)
21:31:41 <boris-42> lol
21:31:52 <russellb> if anyone comes across High/Critical bugs while triaging, it would be good to bring them up here
21:31:55 <dansmith> dripton: that vastly limits the bugs I can fix for credit :(
21:31:58 <russellb> i want to sure at least those get some attention ..
21:32:33 <boris-42> russellb I would like to merge all models vs migrations syncs in H2
21:32:47 <russellb> boris-42: ok, we can try.
21:32:55 <russellb> let's just to sub-teams
21:32:57 <boris-42> it could produce some unexpected bugs in psql and mysql
21:32:57 <russellb> #topic sub-team reports
21:33:08 <russellb> anyone around to give a report?
21:33:12 <russellb> devananda: ?
21:33:13 <johnthetubaguy> yeah
21:33:16 <boris-42> Ok see you later=)
21:33:20 <n0ano> scheduler
21:33:23 <devananda> o/
21:33:25 <harlowja> \o/ but not that much to report on
21:33:38 <russellb> devananda: wanna go first?
21:33:44 <devananda> sure
21:33:45 <russellb> (Ironic / baremetal driver)
21:34:08 <devananda> not much to report on. working on the API service for ironic, definitely not going to be ready for h3 in a comparable-to-baremetal way
21:34:25 <russellb> devananda: i went ahead and deferred the deprecate-baremetal blueprint to post-havana
21:34:38 <devananda> the tripleo side of baremetal is moving alogn well, but i dont think a lot has changed with the baremetal driver itself
21:34:57 <russellb> yeah, saw the proposal to make an official program for it
21:34:59 <russellb> cool stuff.
21:35:05 <devananda> ironic is still waiting for the image utils to move out of nova and into glanceclient
21:35:24 <devananda> once that lands in glanceclient, i think GheRivero will be able to propose a DRY-it-up patch to nova and cinder
21:35:27 <devananda> but that wont be a functional change
21:35:55 <devananda> that's probably it for the nova side of our work latey
21:36:36 <devananda> [eol]
21:36:39 <russellb> ok cool thanks!
21:36:46 <russellb> johnthetubaguy: you're up, for xenapi!
21:36:51 <johnthetubaguy> cool
21:37:03 <johnthetubaguy> so not a lot of news really
21:37:20 <russellb> you need reviews on your h-2 blueprint :-)
21:37:25 <johnthetubaguy> some work towards getting XenAPI gated, but not visible progress
21:37:34 <russellb> oh yeah?  didn't know about that
21:37:35 <johnthetubaguy> +1 for reviews on that blueprint :)
21:37:46 <russellb> that's cool to hear actually.
21:37:53 <johnthetubaguy> a few h-3 blueprints might get pushed out
21:38:02 <russellb> johnthetubaguy: btw, did you see my email a week or two ago about "compute driver criteria" ?
21:38:10 <johnthetubaguy> yeah, Citrix guys have most of the tempest tests going
21:38:16 <russellb> devananda: you too, since it affects baremetal ^^
21:38:25 <johnthetubaguy> russellb: I did, but I should re-read it and check it again!
21:38:31 <russellb> xenapi is already in the clear since it's running in smokestack
21:38:46 <devananda> hm, nope :( /me goes and looks
21:38:46 <johnthetubaguy> :)
21:38:47 <russellb> baremetal needs to have something going in the next 9 months ... but I believe that's being worked on
21:38:56 <russellb> devananda: no worries, easy to miss stuff on the ML ...
21:39:12 <devananda> russellb: tl;dr - we need a smokestack-like test rig, yes?
21:39:14 <russellb> devananda: basically i want all virt drivers being tested against changes, even if just providing info like smokestack does
21:39:17 <russellb> yes
21:39:29 <devananda> russellb: perfect. already talked at length with folks about this, and yea, it's being worked on
21:39:30 <russellb> and i proposed a Icehouse release deadline basically
21:39:34 <russellb> perfect
21:39:35 <devananda> ++
21:39:54 <russellb> and on that note ...
21:39:58 <russellb> hartsocks: hi!
21:39:59 <johnthetubaguy> so, Citrix guys are going to updated their blueprints soon, maybe dropping some from h-3
21:40:08 <johnthetubaguy> I am all done now though
21:40:11 <russellb> johnthetubaguy: ok cool, let me know and i'll update them
21:40:13 <russellb> johnthetubaguy: thanks!
21:40:22 <hartsocks> okay.
21:40:23 <johnthetubaguy> russellb: thanks, will do
21:40:24 <russellb> hartsocks: updates?  also, did you see that email i'm referencing?
21:40:35 <hartsocks> So on VMwareAPI team...
21:40:38 <hartsocks> email?
21:40:45 <russellb> "Compute driver criteria" i think
21:40:55 <russellb> tl;dr ... are you guys working on some sort of CI integration?
21:40:57 <hartsocks> Doesn't ring a bell.
21:41:02 <hartsocks> Yes.
21:41:06 <russellb> ok, then you're safe
21:41:11 <russellb> i'd like to see it by Icehouse
21:41:23 <hartsocks> I was going to mention I heard back from a group working on 'tempest' for vmwareapi stuff
21:41:25 <russellb> which is a good ways out, so shouldn't be a big deal if you're looking into it already
21:41:36 <hartsocks> I'm not going to forecast when they'll have that up
21:41:45 <russellb> but at some point in the future, i'm going to start to push removing things we're not testing :-)
21:41:55 <russellb> that's what the openstack-dev ML post was about
21:42:00 <russellb> kind of a first warning shot about it
21:42:01 <hartsocks> Okay.
21:42:22 <hartsocks> I will feel much better when we get our test coverage up higher.
21:42:26 * russellb nods
21:42:36 <hartsocks> I went around to the H2 blueprint holders...
21:42:41 <hartsocks> looked them deep in the eyes...
21:42:44 <hartsocks> and asked...
21:42:52 <hartsocks> Are you really really going to make H2?
21:42:55 <russellb> i may have already deferred some/most/all of them ...
21:43:03 <russellb> if they didn't look like they could make it
21:43:03 <hartsocks> Yeah.
21:43:13 <russellb> if you feel otherwise, let me know
21:43:41 <hartsocks> We're close but H2 is too close and a lot of our contributors are traveling in the next couple weeks.
21:43:49 <russellb> ack
21:44:02 <russellb> anything else going on?
21:44:04 <hartsocks> It looks like H2 will be all about bug fixes for us.
21:44:23 <hartsocks> That's the big new.
21:44:28 <hartsocks> s/new/news/
21:44:39 <hartsocks> end.
21:44:52 <russellb> cool, thanks!
21:45:01 <russellb> n0ano: scheduling!
21:45:23 <n0ano> like others, not too much this week, we talked about two patches going around...
21:45:50 <n0ano> code refactoring - e.g. moving common scheduler & cinder code to oslo, concensus was that's a good idea..
21:46:08 <harlowja> +1
21:46:13 <russellb> there is some already in oslo
21:46:16 <russellb> that nova needs to get updated to use
21:46:24 <russellb> common filter and weight infrastructure
21:46:30 <n0ano> other was volume affinity filter - concensus there was this was more likely a weighting function rather than a filter
21:46:34 <russellb> was put there when cinder was doing its filter scheduler
21:46:43 <russellb> n0ano: ok i was going to ask about that
21:46:55 <russellb> i'm fine with the weight, but the more i thought about it, the more I'm -2 on the filter as proposed
21:46:58 <russellb> https://review.openstack.org/#/c/29343/ for reference
21:47:20 <n0ano> indeed, we pretty much agreed with you on that, didn't see any compelling use cases for a filter
21:47:27 <russellb> ok
21:47:36 <n0ano> that's about it
21:47:39 <russellb> ok thanks!
21:47:56 <russellb> harlowja: hi!
21:48:02 <harlowja> howday!
21:48:05 <russellb> sup?
21:48:44 <harlowja> so just getting the relase of taskflow ready, shaping up, few more items being worked on, then hopefully can start plugging it in for H3 into cinder at least to start
21:48:55 <harlowja> just mostly chugging away on making that initial release possible
21:49:10 <harlowja> chug chug
21:49:14 <russellb> cool, well thanks for the update, any other details you want to share?
21:49:21 <harlowja> nothing much yet :)
21:49:53 <russellb> k, thanks!
21:49:57 <russellb> #topic open discussion
21:50:06 <russellb> 10 minutes left if anyone wants to chat about ... anything!
21:50:33 <russellb> jenkins is busy.
21:50:38 <russellb> 87 patches total under test right now
21:50:40 <russellb> 30 in the gate.
21:50:46 <russellb> (across all projects)
21:51:18 <russellb> other random interesting number ... in nova we've had 3597 code reviews done by 176 people in the last 30 days.
21:51:26 <timello_> I'm back. Any feedback on patches: https://review.openstack.org/#/c/32040/ and https://review.openstack.org/#/c/32503/ would be nice.
21:51:41 <timello_> They are related to the cold-migrations-to-conductor bp.
21:52:03 <johnthetubaguy> timello_: on my list for tomorrow, a review day for me tomorrow I feel
21:52:14 <harlowja> russellb is there a jenkins monitor that can be used to see how busy it is?
21:52:15 <comstud> i'll be out of town for over a week
21:52:18 <timello_> johnthetubaguy: awesome, I appreciate.
21:52:21 <russellb> johnthetubaguy: cool, appreciate that
21:52:25 <comstud> with limited online-ness
21:52:37 <russellb> comstud: ah yes, thanks for the reminder / heads up
21:52:41 <russellb> comstud: safe travels to you sir
21:52:44 <comstud> ty
21:53:11 <russellb> harlowja: there are a couple of things you can use ... one is http://status.openstack.org/zuul/
21:53:16 <Chinmay_> hey guys i had a quick question about imagebackend caching
21:53:29 <Chinmay_> any reasons why were create it per flavor ?
21:53:33 <russellb> harlowja: the other is https://github.com/kk7ds/openstack-gerrit-dashboard/
21:53:36 <Chinmay_> s/were/we/
21:53:43 <russellb> harlowja: which is just a command line thingy, but it's pretty awesome.
21:53:49 <harlowja> coolness
21:53:51 <dansmith> russellb: JUST a command line thingy?
21:53:53 <russellb> mikal: ^ ?
21:53:57 <russellb> dansmith: hehe
21:54:01 <dansmith> like bash is JUST a command line thingy?
21:54:05 <russellb> yes.
21:54:10 <dansmith> mkay :)
21:54:18 <mikal> russellb: curse you
21:54:30 <harlowja> is it using ncurses?
21:54:39 <harlowja> :)
21:54:39 <mikal> Chinmay_: IIRC, and it might have changed, we cache the glance image, and then the post resize image
21:54:51 <mikal> Chinmay_: we do that to avoid doing the resize over and over
21:54:58 * johnthetubaguy quietly runs away to bed...
21:55:29 <Chinmay_> mikal : ok thanks
21:56:40 <russellb> harlowja: no
21:56:47 <russellb> harlowja: print!
21:57:08 <dansmith> russellb: I hit that review
21:57:12 <dansmith> the ec2 madness one
21:57:34 <russellb> dansmith: cool
21:58:08 <russellb> dansmith: ok so hopefully that's enough to convince him to split it up
21:58:08 <harlowja> russellb hmmm might have to hook in curses :-p
21:58:38 <russellb> harlowja: yeah
21:58:47 <dansmith> are we done here? :)
21:58:51 <russellb> yep!
21:58:55 <russellb> thanks everyone!
21:58:57 <russellb> #endmeeting