14:01:48 <bauzas> #startmeeting nova_scheduler
14:01:49 <openstack> Meeting started Mon Mar 21 14:01:48 2016 UTC and is due to finish in 60 minutes.  The chair is bauzas. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:50 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:52 <openstack> The meeting name has been set to 'nova_scheduler'
14:01:53 <jaypipes> o/
14:01:55 <Yingxin> o/
14:02:50 <bauzas> mornooning
14:03:08 <bauzas> I have no precise agenda in mind because RC2 but lemme try to handle that one
14:03:16 <edleafe> good UGT morning
14:03:30 <bauzas> #topic Specs, features, blueprints
14:03:31 <cdent> o/
14:03:59 <bauzas> #info Newton development start last week, you can resubmit your nova specs for Newton
14:04:04 <bauzas> #undo
14:04:04 <openstack> Removing item from minutes: <ircmeeting.items.Info object at 0x9fb7f10>
14:04:09 <bauzas> #info Newton development started last week, you can resubmit your nova specs for Newton
14:04:26 <bauzas> so, master is now pointing against newton
14:04:39 <bauzas> which means we can start to merge things that look like not as a bug
14:04:59 <jaypipes> ya
14:05:14 <bauzas> for the moment, most of us are eaten by RC2 but previously approved specs can be again submitted
14:05:23 * jaypipes in the process of answering ajo's question on the openstack-dev list around generic-resource-pools and QoS/avaiability zones in Neutron.
14:05:56 <ajo> hi jaypipes
14:06:03 <jaypipes> ahoy ajo
14:06:06 <ajo> :)
14:06:19 <bauzas> #info in case you had a previously approved spec not yet implemented, please create a new blueprint named <blah>-newton and create a new spec change with Previously-Approved tag
14:06:29 <jaypipes> bauzas: so, I resubmitted a bunch of the resource-providers series to newton.
14:06:34 <bauzas> coolness
14:06:50 <jaypipes> bauzas: but I did a git mv unfortunately, which isn't apparently what I should have done :(
14:06:59 <bauzas> jaypipes: I litterally had zero time to work on features last week, and I don't hope to have more time for it this week too
14:07:14 <ajo> jaypipes, I will extend with another email for qos/minimum bandwidth, but I believe it doesn't fit as-is (at least)
14:07:27 <jaypipes> bauzas: that's cool, because I actually *do* have time to do code reviews on cdent's stuff and resubmit specs this week.
14:07:40 <cdent> \o/
14:07:42 <jaypipes> ajo: wait until you read my response pls :)
14:08:21 <bauzas> okay, I'll slowly move up to reviewing features
14:08:40 <bauzas> I think mriedem provided a nova etherpad for newton
14:08:44 <johnthetubaguy> do we have a any list of specs we want to worry about for newton at this point?
14:08:46 <johnthetubaguy> yeah
14:08:52 <mlavalle> bauzas, jaypipes: is this a good moment to ask quations about https://review.openstack.org/#/c/253187 or should I wait to the open discussion section?
14:09:01 <johnthetubaguy> #link https://etherpad.openstack.org/p/newton-nova-priorities-tracking
14:09:10 <bauzas> even if we haven't yet agreed on prios yet, it could be interesting to put changes up there ^
14:09:29 <johnthetubaguy> aslo this one:
14:09:31 <johnthetubaguy> #link https://etherpad.openstack.org/p/newton-nova-spec-review-tracking
14:09:44 <bauzas> johnthetubaguy: I think atm we should be discussing about what we want for newton
14:09:46 <johnthetubaguy> so as a subteam, its still good to recommend what you want
14:10:25 <bauzas> what I'd like is to see people providing their thoughts about which features they want for newton, and put them in the above etherpad(s=
14:10:25 <johnthetubaguy> I would like to jump on those specs jaypipes has up soon ish for the generic resource pools bits
14:10:36 <jaypipes> mlavalle: go for it.
14:10:43 <bauzas> johnthetubaguy: like I said, probably not this week for me
14:10:45 <ajo> jaypipes, : ack
14:11:25 <mlavalle> jaypipes: line 438 mentions a `resource-providers-allocations` spec. Have you pusshed this one to gerrit yet?
14:11:32 <bauzas> ajo, mlavalle, seems like you had questions about the resource-providers specs ?
14:12:03 <jaypipes> mlavalle: yup: https://review.openstack.org/271779
14:13:08 <mlavalle> jaypipes: thanks. I'll be working on routed networks from the Neutron side. I'll follow up your specs very closely. I have some feedback already that I will publish later
14:13:37 <jaypipes> mlavalle: sweet, thank you sir.
14:13:42 <mlavalle> jaypipes: carl_baldwin has aked me to work on ip deferral for routed networks
14:13:46 <bauzas> I should suggest to wait for jaypipes's reply to ajo
14:13:48 <edleafe> jaypipes: I will review once I'm dug out :)
14:14:03 <jaypipes> mlavalle: ++ carl_baldwin and I had good conversations about this in Bristol.
14:14:04 <ajo> bauzas, yes, I was going to explain, but let's wait for that
14:14:06 <bauzas> in case neutron people have questions
14:14:23 <jaypipes> bauzas: sure, though that is not about routed networks. it's about how QoS can fit here.
14:14:32 <ajo> anyway, jaypipes, the other possible use case (QoS/bw guarantees) is quite different to routed networks
14:14:49 <bauzas> jaypipes: ack
14:15:08 <ajo> in our (QoS/bw) case, we could end up spawning a "resource pool" and an aggregate of a single compute by "resource pool".
14:15:40 <ajo> and we may need nova to identify what's the right compute node for an amount of requested bw, not by the "resource pool" id
14:15:56 <ajo> but, let's elaborate on list. sorry.
14:16:26 <bauzas> ok, moving on ?
14:17:59 <jaypipes> ajo: k, responded.
14:18:06 <ajo> jaypipes++ thanks
14:18:19 <bauzas> ok
14:18:22 <bauzas> #topic bugs
14:18:46 <bauzas> #info RC1 is out, feel free to play with it
14:19:04 <cdent> I marked a bunch of ancient and/or wishlist stuff invalid last week. So the count of total bugs tagged scheduler should be down.
14:19:19 <bauzas> #info in case you find regressions, please ping us asap by creating a bug tagged with mitaka-rc2
14:19:42 <bauzas> so, working on chasing down regressions this week
14:20:14 <bauzas> I remind that since master is now newton, bugfixes are provided against N and need to be backported to stable/mitaka in case they are regressions
14:20:37 <bauzas> mitaka will be GA first week of April IIRC
14:20:59 <bauzas> for other bugs that are not regressions, the usual workflow stays the same
14:21:07 <bauzas> cdent: thanks
14:24:09 <bauzas> nothing important to raise ?
14:25:16 <jaypipes> nope
14:25:32 <bauzas> okay
14:25:37 <bauzas> #topic open discussion
14:25:41 <bauzas> mic is yours
14:26:21 <mlavalle> bauzas: I would like to ask the team for feedback for the routed networks spec: https://review.openstack.org/#/c/263898
14:26:30 <bauzas> sure thing
14:27:01 <mlavalle> I would specifucally like feedback from jaypipes regarding the algorithm outlined in lines 201 to 260
14:27:33 <jaypipes> mlavalle: you got it. gimme 30 minutes to review.
14:27:39 <mlavalle> jaypipes: :-)
14:29:18 <mlavalle> carl_baldwin did a good job outlining a first draft. But we need feedback form the expersts :-)
14:29:58 <jaypipes> mlavalle: I'm no expert, but I'll try my best.
14:30:05 * jaypipes stabs Gerrit Web UI.
14:31:13 <jaypipes> guh, can't stand it... switching to gertty.
14:34:06 <edleafe> so sensitive!
14:34:16 <bauzas> okay, it seems we have 25 mins for us ?
14:34:25 <bauzas> bye guys and ty
14:34:27 <bauzas> #endmeeting