21:00:52 <mikal> #startmeeting nova 21:00:53 <openstack> Meeting started Thu Oct 9 21:00:52 2014 UTC and is due to finish in 60 minutes. The chair is mikal. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:56 <openstack> The meeting name has been set to 'nova' 21:01:07 <mikal> Greetings earth humans 21:01:10 <jogo> o/ 21:01:19 <melwitt> \o 21:01:20 <alaski> hi 21:01:23 <dansmith> foo 21:01:26 <jogo> bar 21:01:27 <angdraug> o/ 21:01:29 <mikal> The agenda is quite short this week, so here's hoping for a short meeting 21:01:42 <mikal> #topic Juno RC2 21:01:47 <mikal> There will be an RC2 for nova 21:01:58 <mikal> When I went ot bed last night we were still waiting for a fix for one last bug 21:02:18 <mikal> Oh, it merged over night... 21:02:29 <dansmith> which was that? 21:02:33 <mikal> One sec 21:02:35 <mikal> Laptop fail 21:02:46 <angdraug> #link https://review.openstack.org/126144 21:02:48 <dansmith> it's a poor workman who blames his tools 21:02:49 <mikal> #link https://review.openstack.org/#/c/126144/ 21:02:57 <dansmith> ah, yeah, good 21:03:06 * mikal double checks the bug list as well 21:03:09 <angdraug> what about the evacuate fix for rbd? 21:03:12 <angdraug> #link https://review.openstack.org/121745 21:03:25 <mriedem> angdraug: don't think so 21:03:26 <mikal> #link https://launchpad.net/nova/+milestone/juno-rc2 21:03:29 <mriedem> it doesn't break upgrades does it? 21:03:34 <dansmith> mriedem: agreed 21:03:41 <mikal> angdraug: its not targetted to rc2? 21:03:50 <mikal> So yeah, that's good 21:03:55 <mikal> We have fixes for all the targetted bugs 21:03:56 <mriedem> it's not 21:04:03 <angdraug> no its not, I don't have the right to target it to anything 21:04:06 <mikal> I will check if we've done the backports after my morning meetings 21:04:07 <mriedem> i've marked it for juno-backport-potential 21:04:13 <angdraug> thanks! 21:04:16 <mriedem> angdraug: join the bug team, it's free 21:04:28 <mikal> I will (or some volunteer) need to re-generate the translations for the rc 21:04:32 <mikal> But that's a mechanical thing 21:04:41 <mriedem> not it 21:04:47 <mikal> Heh, I can do it 21:04:50 <mikal> Its "documented" at least 21:04:58 <mikal> So... Let's move on 21:05:02 <mikal> #topic Kilo 21:05:10 * dansmith has a kilo thing 21:05:14 <mikal> Specs are now open 21:05:21 <mikal> There's a fast track for previously approved stuff 21:05:27 <mikal> And another different fast track for trivial stuff 21:05:36 <mikal> dansmith: what is your thing? 21:05:39 <dansmith> https://review.openstack.org/#/c/126620/ 21:05:41 <dansmith> we 21:05:48 <dansmith> have been talking about doing that for at least two cycles 21:05:57 <dansmith> and instance_extra is a good place to put it, 21:06:10 <dansmith> and some of the stuff we have planned really needs extra_specs for flavors to be persisted 21:06:15 <dansmith> so eff it, lets get it done! 21:06:20 <mikal> Heh 21:06:23 <dansmith> I have patches up for the code I'm working on, 21:06:24 <mikal> But, dansmith -1'ed it 21:06:25 <dansmith> but that's the spec 21:06:38 <dansmith> yeah, that guy is an ass, ignore his -1 21:06:43 <mikal> Heh 21:06:55 <mikal> Ok, so I guess that's a request for specs-core's to review that spec please 21:06:57 <dansmith> anyway, hoping to get that one in the queue soon here 21:07:00 <dansmith> cha, please 21:07:04 <mikal> #link https://review.openstack.org/#/c/126620/ 21:07:11 <dansmith> I haven't created the blueprint yet in case anyone wants to argue over the horrible name I picked 21:07:18 <mikal> Ok 21:07:20 <mikal> Fair enough 21:07:25 <mikal> So... 21:07:37 <mikal> We said we would ask about trivial things in each meeting 21:07:46 <mikal> So, has anyone proposed a trivial BP yet that they need approved? 21:07:48 <mriedem> jogo had one 21:07:54 <mikal> linky? 21:08:02 <mriedem> http://lists.openstack.org/pipermail/openstack-dev/2014-October/048199.html 21:08:04 <mriedem> pp[d 21:08:07 <mriedem> oops 21:08:07 <mikal> Actually, let's do this 21:08:08 <dansmith> we already approved the kilo-objects one 21:08:09 <mriedem> finger fail 21:08:17 <mikal> #topic Kilo trivial BPs approval 21:08:25 <mriedem> https://review.openstack.org/#/c/127123/ 21:08:39 <mriedem> that's the one jogo thought was trivial 21:09:08 <jogo> yup https://blueprints.launchpad.net/nova/+spec/io-ops-weight 21:09:20 <mikal> Ahhh, cool 21:09:31 <mikal> Does anyone disagree with it being trivial? 21:09:35 <dansmith> it was already a spec though 21:09:44 <mikal> dansmith: the spec was only 10 lines 21:09:45 <dansmith> maybe it's worth just keeping it since it's done? 21:09:56 <jogo> dansmith: but the delta in the spec is 10 lines 21:10:05 <alaski> yeah, I wouldn't ask for a spec now but since it's there... why not keep it 21:10:26 <jogo> alaski: that sounds reasonable 21:10:28 <dansmith> um 21:10:35 <boden> question -- I've seen some patches coming through for what I'd consider trivial BPs, but the BPs are not approved.. should we be -1|2'ing these until the BP is approved? 21:10:44 <alaski> perhaps the precedent of a spec for a weigher isn't good, dunno 21:10:48 <jogo> boden: this is only for specs 21:10:50 <dansmith> the spec has a delta too, but it's actually just moving the spec from juno right? 21:11:06 <jogo> boden: ahh sorry misread your question, I think you should WIP them 21:11:15 <mikal> Oh I see 21:11:32 <mikal> Yeah, its a move from Juno with a small tweak post move? 21:11:37 <dansmith> right 21:11:43 <mikal> ITs a juno-approved spec which wasn't implemented 21:11:45 <jogo> dansmith: yeah, if we want to keep the juno copy in tree he needs a rebase 21:11:52 <dansmith> mikal: right 21:11:57 <dansmith> mikal: seems silly to throw it away 21:12:02 <mikal> jogo: I think we should keep the juno one in place 21:12:12 <mikal> But why don't we just fix it for him? 21:12:15 <mikal> And then approve it? 21:12:16 <dansmith> I mean, it looks pretty minor but.. not sure why we'd throw away info 21:12:21 <jogo> mikal: that works 21:12:23 <dansmith> sure 21:12:30 <jogo> so this would count as trivial but he already wrote it, so why not 21:12:32 <mikal> dansmith: I didn't realize it was a move, it just looked like a really small diff at first 21:12:37 <dansmith> jogo: right 21:12:47 <mikal> jogo: yeah, and it gets the "previously-approved" fast track anyways 21:12:55 <mikal> Ok, let's do that 21:12:57 <mikal> A volunteer? 21:12:58 <dansmith> mikal: yeah, the previously-approved tag is the key I think 21:13:11 <dansmith> unfortunately, moves are really obscure in gerrit :( 21:13:25 <dansmith> I volunteer... jogo 21:13:29 <mikal> Heh 21:13:42 <mikal> I'll do it after my meetings if jogo doesn't get to it before then 21:13:43 <jogo> dansmith: already on it 21:13:49 <mikal> Ok, I shall stand down... 21:13:52 <dansmith> jogo is like one of the best spec rebasers I know 21:13:57 * dansmith blows smoke 21:13:59 <mikal> So, any other trivial Kilo BPs? 21:14:25 <mikal> No, moving on 21:14:33 <mikal> #topic Code for unapproved blueprints 21:14:41 <mikal> Bringing this up because we were just asked 21:14:55 <mikal> My thoughts are that we should do a procedural -2 on code where the BP isn't approved yet 21:14:59 <mikal> Like we did in Juno 21:15:09 <mikal> The biggest wart being that means that the -2'er needs to be around to remove their -2 later 21:15:32 <mikal> Anyone got a better idea? 21:15:51 <mriedem> could also -WIP but that gets lost on new patch set 21:15:52 * dansmith hears crickets 21:16:00 <mriedem> so don't -2 and die 21:16:01 <mikal> Ok, the plan stands 21:16:08 <mikal> I think in Juno it was mostly John and I who did it 21:16:12 <mikal> But I think any core can 21:16:17 <mikal> As long as they promise to be around later to remove it 21:16:21 <mikal> No vacations, ever 21:16:41 <mikal> #topic Gate status 21:16:47 <mikal> Anything we need to know with the gate? 21:16:51 <mriedem> naw 21:16:54 <dansmith> things aren't happy right? 21:16:57 <mriedem> we have quota bugs 21:17:03 * mriedem looks up link 21:17:06 <dansmith> lots of failures because of mirror timeouts it seems 21:17:08 <jogo> well we have confirmed we have quota bugs 21:17:12 <jogo> we knew we had them for a long time 21:17:17 <mriedem> the timeouts are resolved i think 21:17:20 <mikal> Could we perhaps fix them? 21:17:21 <dansmith> ah, okay 21:17:27 <mriedem> https://bugs.launchpad.net/nova/+bug/1353962 21:17:29 <mikal> The quota bugs that is 21:17:32 <uvirtbot> Launchpad bug 1353962 in nova "Test job failes with FixedIpLimitExceeded with nova network" [Undecided,Confirmed] 21:17:38 <mriedem> so i got a few patches up for more logging, 21:17:46 <mriedem> and we're seeing some really f'ing weird things with quotas in the logs re: fixed_ips 21:17:51 <mriedem> like overquota on 0 usage 21:17:54 <mriedem> and -1 usage 21:18:14 <mriedem> i've refactored most of the quota_reserve method in the db api so it's not so terrible to look at 21:18:22 <mikal> mriedem: the bug is unassigned, but it sounds like you're working it? 21:18:25 <mriedem> next step is figuring out what in there is buggy b/c of project vs user specific quotas 21:18:38 <mriedem> mikal: well, i'm losing sleep over it... 21:18:44 <mriedem> so i guess 21:18:53 <mikal> mriedem: do you need volunteers to help? 21:18:56 <mriedem> yes 21:18:58 <mriedem> people that know quotas 21:18:59 <mikal> mriedem: or are we blocking on more debugging info? 21:19:02 <mriedem> and nova-network? 21:19:06 <mikal> Jay Pipes perhaps? 21:19:07 <mriedem> so vishy, comstud and Vek 21:19:10 <mikal> Or Vishy 21:19:31 <mriedem> it's really fixed_ips only i think b/c that's the only thing we default to unlimited quotas 21:19:47 <mriedem> and i think that plus the fact it's a per-project quota resource is showing the bug somewhere 21:19:56 <mikal> mriedem: do you want me to ping vishy, or shall you? 21:20:02 <mriedem> i just did i think 21:20:08 <mikal> Ok 21:20:09 <mriedem> oh and this might help https://review.openstack.org/#/c/76413/ 21:20:17 <mriedem> i'm going to clean that up if the author doens't first 21:20:48 <mikal> The author has been awol on that patch for what, two weeks? 21:20:50 <mriedem> otherwise the gate issues are mostly infra http://status.openstack.org/elastic-recheck/gate.html 21:20:52 <mikal> I think you're safe to clean it up 21:21:00 <mriedem> mikal: it's one of our oldest patches, that's how i saw it 21:21:04 <mriedem> from the ML shame this morning 21:21:13 <mriedem> so i'll clean it up when i get 5 minutes 21:21:16 <mikal> There was ML shame? 21:21:21 <mikal> I haven't read my email yet today 21:21:22 <mriedem> well, indirectly 21:21:29 <mriedem> review stats 21:21:37 <mikal> Ahhh, ok 21:21:43 <mikal> So, nothing else for the gate? 21:21:45 <mriedem> anyway, that's it 21:21:45 <mikal> Move on again? 21:21:45 <mriedem> no 21:21:47 <mriedem> yes 21:21:55 <mikal> #topic Open Discussion 21:22:00 <mikal> Ok, what else do people have? 21:22:03 <dansmith> objects 21:22:13 <mikal> We should have them 21:22:14 <jogo> start planning the next midcycle? 21:22:22 <jogo> other teams are doing it, I feel left out 21:22:24 <mikal> Ok, let's do one at a time though 21:22:33 <mikal> dansmith: as in you want help with objects? 21:22:43 <dansmith> well, always, but that's not what I was going to say: 21:22:57 <dansmith> first, I have some patches up for cleaning up how we manage backports of dependent objects, 21:23:01 <dansmith> so people should look at those 21:23:14 <dansmith> secondly, it looks like we're going to try to push the core infrastructure into an oslo library in kilo, 21:23:34 <mriedem> ooo 21:23:51 <mriedem> that's gantt's wet dream right? 21:23:52 <yjiang5> dansmith: cool 21:23:54 <dansmith> which means (a) some cleanups need to go into nova before the forklift, so please be on the lookout for those and (b) at some point, we'll need to cut over to the library, which will be some more 21:24:06 <mikal> Cool 21:24:11 <dansmith> mriedem: heh, I dunno what they dream about :) 21:24:15 <mikal> dansmith: do we want a spec for it? Or a trivial BP? 21:24:15 <mriedem> objects 21:24:20 <mikal> Or we can see what to review to get it done? 21:24:28 <dansmith> mikal: there will be a spec for oslo soon 21:24:31 <mikal> /Or/So/ 21:24:41 <dansmith> mikal: I dunno that we really need another one for "delete code from nova when appropriate" 21:24:42 <mriedem> oslo.objects 21:24:44 <dansmith> but whatever you want 21:24:48 <mikal> I think a nova trivial BP to cover the refactoring might help reviewers 21:24:49 <yjiang5> dansmith: devananda: I remember ironic is using the object also, does it sync with nova side? 21:24:51 <mriedem> we don't have specs for moving to oslo libs 21:24:51 <dansmith> okay 21:24:55 <mriedem> otherwise dims' head would explode 21:25:07 <dansmith> yjiang5: very occasionally, this should help 21:25:18 <mikal> Ok, next was the mid-cycle 21:25:26 <mikal> So now is probably a good time to confess that I totally have a plan 21:25:33 <yjiang5> dansmith: cool. 21:25:38 <jogo> mikal: woot 21:25:45 <dansmith> hawaii? 21:25:48 <mikal> jogo: the plan is current California, in January 21:25:51 <devananda> yjiang5: yes, we are interested in consolidating objects and moving them to oslo 21:25:55 <angdraug> yay! 21:25:56 <mikal> IBM wanted it, but there's snow 21:25:59 <mikal> And I don't like snow 21:26:01 <jogo> not Hawaii? 21:26:13 <mikal> I haven't written an email about it yet, but expect that in the next day or so 21:26:13 <tjones> almost hawaii 21:26:17 <jogo> mikal: southern California 21:26:21 <mikal> jogo: HP failed to step up to offer a Hawaiian location 21:26:24 <dansmith> in no way is california "almost hawaii" 21:26:30 <tjones> LOL 21:26:31 <mikal> jogo: think Palo Alto 21:26:39 <mikal> dansmith: it has in-and-out 21:26:42 <jogo> mikal: :( damn 21:26:43 <mriedem> ibm would do the L mid cycle 21:26:44 <mikal> dansmith: that's close enough for me 21:26:51 <dansmith> pfft 21:26:52 <devananda> mikal: just not sunnyvale 21:26:56 <mriedem> june is much better in MN than feb 21:27:03 <mikal> mriedem: yes, I think we should seriously consider IBM for the Lemming mid-cycle 21:27:13 <dansmith> but wait 21:27:21 <dansmith> if I step back onto an IBM campus, I burst into flame or something 21:27:28 <mikal> dansmith: heh 21:27:31 <edleafe> mikal: oh snap 21:27:34 <mikal> dansmith: I will bring a bucket of water then 21:27:37 <dansmith> okay 21:27:43 <mikal> So, yes 21:27:49 <mikal> I think we're actually ahead of the curve with mid-cycle 21:27:56 <mikal> I just haven't communicated well 21:28:03 <mikal> So, any other topics? 21:28:09 <mikal> Or do people want 30 minutes of their day back? 21:28:14 <dansmith> early mark! 21:28:21 <mikal> Its a thing 21:28:23 <jogo> I have one last quick thing 21:28:27 <mikal> jogo: go 21:28:56 <jogo> please look at https://etherpad.openstack.org/p/kilo-nova-priorities 21:29:02 <mikal> Oh yes 21:29:09 <mikal> And continue to look at the summit ideas ehterpad too 21:29:11 <jogo> #note please look at and contribute to https://etherpad.openstack.org/p/kilo-nova-priorities 21:29:18 <mikal> I want to try and get those two nailed down in the next week or so 21:29:23 <mikal> Once rc2 is out the door 21:30:07 <mikal> https://blueprints.launchpad.net/nova/+spec/io-ops-weight for reference 21:30:19 <mikal> So, we're done? 21:30:34 <jogo> #link https://etherpad.openstack.org/p/kilo-nova-priorities 21:30:46 <jogo> I can never remember the meeting bot commands 21:31:03 <jogo> early mark! 21:31:11 <mikal> Going... 21:31:14 <mikal> Going... 21:31:18 <jaypipes> gone. 21:31:21 <mikal> Gone 21:31:26 <mikal> #endmeeting