21:01:04 #startmeeting nova 21:01:05 Meeting started Thu Jan 10 21:01:04 2013 UTC. The chair is vishy. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:06 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:08 The meeting name has been set to 'nova' 21:01:14 hi! 21:01:20 #chair russellb 21:01:21 Current chairs: russellb vishy 21:01:43 we don't really have anything on the agenda 21:01:45 #link http://wiki.openstack.org/Meetings/Nova 21:01:49 :) 21:01:55 heh, yeah. 21:01:59 thanks for the blueprint approval! 21:02:03 I guess we can talk about bugs for a moment 21:02:10 russellb: sounds good 21:02:17 #topic bugs 21:02:17 #link http://webnumbr.com/untouched-nova-bugs 21:02:26 untouched bugs is up to 48 21:02:32 creeped up over the holidays, not surprising 21:02:54 so, just need some folks to put that back on their weekly to-do to help triage a few 21:03:29 that's pretty much all i wanted to point out ... 21:04:35 any other bug discussion? 21:04:42 russellb: yeah 21:04:53 was hoping to get https://bugs.launchpad.net/nova/+bug/1075716 backported to Folsom 21:04:54 Launchpad bug 1075716 in nova "Soft delete does update quotas correctly" [Undecided,Fix released] 21:05:11 have you proposed it? 21:05:14 as its user facing bug that is easy to hit in production. 21:05:32 russellb: working on it now, its not a completely trivial backport though 21:05:35 i don't think you need to wait for someone else to backport it if it's important to ou 21:05:37 gotcha 21:06:05 if johannes or someone else who worked on quota is around, help would be appreceated 21:06:30 so, want me to target to folsom and assign to you? 21:06:36 russellb: sure 21:07:33 jog0: k done 21:07:41 russellb: thanks 21:08:43 I guess another bug note ... since we're now in the grizzly-3 cycle, we should definitely be watching out for bugs that we want to make sure get fixed in grizzly (and target them to grizzly-3 for now?) 21:09:15 need to start narrowing in on what the final release is going to look like 21:09:19 comstud: do you remember anyone else that looked at quotas? 21:10:59 any other bug discussion? 21:11:39 ok then. 21:11:43 #topic grizzly-3 21:12:03 grizzly-3 is Feb 21, that'll be here faster than we think :) 21:12:09 that's also feature freeze, in theory 21:12:26 we have way more blueprints than will get done, realistically ... https://blueprints.launchpad.net/nova/+milestone/grizzly-3 21:12:33 if you have cycles, take a look for something to volunteer for 21:12:37 russellb, so only bugs after 21st? 21:12:49 dims: that's the idea, only bugs without an exception 21:13:06 k 21:13:10 you can ask for a feature freeze exception, but don't count on it 21:13:26 especially with nova, we'll be busy enough with bugs that overloading ourselves with feature exceptions isn't ideal 21:14:01 understood 21:14:24 looks like all blueprints are assigned, technically 21:14:37 but if something looks interesting, don't hestitate to contact whoever is assigned and offer to help 21:14:54 i guess that's it on that ... 21:15:06 #topic Open Discussion 21:15:31 So, I feel like attendance has been kind of light to this meeting in general ... 21:16:06 is there something we can be doing to make this more useful? 21:16:30 would like to encourage more attendance and participation. 21:16:50 add a meeting reminder to the ML 21:16:50 russellb: seems like there is more early in the cycle 21:16:57 russellb: while we are planning things out 21:17:04 yeah, true. 21:17:24 Also, I'm here, I just haven't had anything to say 21:17:26 jog0: yeah, I usually send one. I didn't today (and attendence seems lighter than usual, too) 21:17:26 seems like now it is mostly implementation and reviews 21:17:39 and asking for reviews :) 21:18:05 45 blueprints in g-3 seems totally reasonable :) 21:18:13 cool, well, i'm happy to put in work outside of the meeting if there's something i can do to make it more useful, happy to hear ideas if anyone has them at some point 21:18:23 vishy: really? 21:18:32 sarcasm is hard to detect on IRC, heh. 21:18:35 * vishy is dripping with sarcasm 21:18:37 smell the sarcasm! 21:18:41 excellent. 21:19:04 well I guess as long as everyone is aware of the deadline 21:19:12 this looks interesting https://review.openstack.org/#/c/18462/ 21:19:15 and for the love of (*&#$(#@, please don't show up with a massive patch 2 days before the deadline 21:19:20 we had multiple instances of that with folsom 21:19:30 i keep missing stuff since i turned off my notification of every single commit 21:19:34 define massive ;) 21:19:48 beagles: thousands of lines 21:20:08 i can haz merge? 21:20:23 ugh 21:20:25 i know you have 100 other open reviews and i just dropped this, but ... 21:21:10 * russellb reads about utilization-based-scheduling 21:21:30 as a new comer, i end up staring at reviews of my patches to see if there was any movement, i am slowly learning to peel my eyes away from them. any tips? 21:21:48 whoops 21:22:00 comstud: ;) 21:22:15 dims: I'd say it's fair to expect to have to wait a bit but also fair to ping people for reviews if it sits too long 21:22:22 time flies when fixing bugs 21:22:23 comstud highlights on (*&#$(#@ 21:22:31 lol 21:22:38 It's not as glamorous to review which is why they take along time, everyone prefers to write new code :) 21:22:52 :) 21:22:55 sometimes i prefer to review, oddly 21:23:12 though i haven't been reviewing as much as usual lately ... need to kick it back up 21:23:43 dims: also is useful to ping people that you know will be interested 21:23:59 Or always ping vish like I do 21:24:03 jk 21:24:08 rmk: nice. 21:24:12 like if they touched the code around where you are editing 21:24:12 vishy, right. i am trying to get to know folks :) 21:24:58 I like the idea of utilization based scheduling 21:25:02 We have related problems to solve 21:25:31 Say you're using utilization based but have a bunch of suspended VMs all over the place.. we never reschedule 21:26:20 * dims says thanks for the tips 21:26:39 an interesting thing to add to the scheduling mix for sure 21:27:08 i like it 21:27:13 nods 21:27:33 welp ... anything else for open discussion? reviews anyone wants to plug? 21:27:46 light meeting today. 21:27:58 who wants to help comstud get the rest of cells in? 21:27:59 :) 21:28:08 I haven't looked closely at this patch but how much historical utilization data does it take into account? 21:28:18 is it just a point-in-time utilization snapshot? 21:28:55 At first glance it looks like it is 21:29:37 nah, it has a window of samples 21:29:52 see AveragingProfiler in the compute resource_tracker 21:30:18 everyone wants to help me 21:30:22 right!? 21:30:40 technically there's still more I don't have up for review yet ;) 21:30:57 heh, ok but seriously, grizzly-3, we have *got* to have it all in 21:31:03 comstud: start raging at people to review 21:31:19 * russellb should review the rest i guess 21:31:31 russellb: yeah, i rage at you first. 21:31:41 ()!*$()!@*$ 21:31:43 rmk: i see that it's keeping 10 samples by default, but not sure how often that is 21:32:17 interesting 21:35:21 comstud: is there going to be a new rev of cells for devstack? 21:35:49 rmk: every run of periodic tasks, so roughly once a minute 21:35:57 sdague: not sure, i wasn't the one working on it 21:36:04 i assume the answer is yes at some point here 21:36:24 rmk: took me a while to dig for that, heh 21:36:54 I wrote the first version of cells on devstack. I haven't looked into what's changed with cells recently but I can work on getting together an update 21:37:08 russellb: cool thanks.. 21:37:10 doesn't seem like the usage has changed, really 21:37:14 Going to look at this some more 21:37:19 rmk: cool :) 21:37:36 comstud cleaned up the code, but didn't seem like the config and such changed, so hopefully it just still works 21:37:39 alaski: shouldn't be much externally.. except the API extension 21:38:04 * dansmith staggers in 21:38:37 anyone want to poke at 19077, adding a REST API for baremetal? 21:38:42 dansmith: slacker! 21:38:46 dansmith: jk <3 21:38:48 cool, I'll test it out and update if necessary. And push up a review at some point. 21:39:12 russellb: my lotus notes calendar doesn't understand UTC, so I correct in my head.. and lost track today :D 21:39:30 zomg time zones 21:41:49 dansmith: not sure how much you saw in #openstack-nova ... i started an rpc based servicegroup driver, but realized there's still a lot of direct service db access scattered around, so it's going to be more difficult than i thought 21:42:12 russellb: yeah, not surprised.. like the thing comstud is working on right now 21:42:12 dansmith: i may just do a conductor db proxy approach as the first pass like we did everything else ... 21:42:21 russellb: okay 21:42:56 I've been trying to start digging into finding some of those hidden dragons for two days, but keep getting distracted 21:43:08 all good, i know how it goes 21:43:44 nova party at dansmith's house in April? :-p 21:44:08 heh 21:44:30 never been to portland, so looking forward to it 21:44:39 dansmith: it seems silly to pass an insanely large message with full instance model when the DB only needs the uuid 21:44:48 I try to avoid it at all costs, so don't expect me to know my way around or anything :D 21:44:48 dansmith: but I see other calls do this 21:44:55 oops, wrong chan 21:44:59 :) 21:45:02 heh 21:45:04 * comstud forgot we were having a meeting 21:45:08 comstud: yeah we've been pretty much doing that everywhere 21:45:14 comstud: well the meeting doesn't really have a topic right now 21:45:22 it's like ... #openstack-nova ... but logged? 21:45:27 haha 21:45:37 i guess we can move back over :) 21:45:44 any last meeting worthy topics? 21:46:08 if you have any agenda requests, update the wiki before next week 21:46:18 or ping me and i'll add it if you don't feel like it, heh 21:46:34 kthxbye. 21:46:42 #endmeeting