21:01:07 #startmeeting nova 21:01:08 Meeting started Thu Jun 27 21:01:07 2013 UTC. The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:11 The meeting name has been set to 'nova' 21:01:14 Hello! 21:01:17 who's around to talk about nova? 21:01:19 hello 21:01:19 \o 21:01:22 russellb: hi 21:01:22 ji 21:01:23 me 21:01:25 hi 21:01:28 yo, quickly brb, but i'll be there 21:01:28 me 21:01:38 o/ 21:01:43 me 21:01:47 welcome all 21:01:53 \o 21:01:54 #topic blueprints 21:02:07 #link https://launchpad.net/nova/+milestone/havana-2 21:02:12 havana-2 is set to be released in 3 weeks 21:02:20 which means the merge deadline is about 2.5 weeks away 21:02:38 i started moving blueprints to havana-3, starting with ones still marked as "Not Started" 21:02:42 we still have 74 :-) 21:03:01 so, let's spend some time going through some of them and talk about their status 21:03:21 cyeoh: hi! we've got lots of v3 api blueprints. i believe we were originally targeting to finish it by havana-2 21:03:26 cyeoh: how is it looking? 21:03:54 well not too bad, I don't think that all will get merged by H2 21:04:03 but certainly most will be in review 21:04:18 ok. so if it'll all be in review, then feel very confident about havana-3 at least? 21:04:23 I think we have probably around 60 or so changesets in the review queue at the moment? I haven't had a close look today 21:04:28 wow 21:04:50 yea, that was my main reason for setting a fairly aggressive H2 goal - to make sure we make H3 21:04:52 #help biggest hold-up on v3 api blueprints' progress is getting reviews. close to 60 in the queue 21:05:15 at least most of the conversions are pretty easy to review IMO 21:05:25 the things at the end that are more uncertain are the increased tempest testing and documentation 21:05:48 but we had that discussion around xml/json - and it will be a matter of prioritising what is most important 21:05:51 so, how about novaclient support? 21:05:53 is that on the radar at all? 21:06:16 we don't really have anyone explicitly assigned for novaclient work. 21:06:32 Will need to see how things look post H2 21:06:36 #action need to create a v3 api novaclient blueprint and find someone to work on it 21:07:13 as for docs, primarily the api docs part right? 21:07:37 yes, and smaller things like guides on how to write a v3 extension 21:07:50 hopefully much of the api part can be autogenerated like we do for v2, but it will be some work 21:08:05 at least it could be worked on late if needed 21:08:09 yes, have to sort out exactly what we need and the changesets will be big if anything like v2 21:08:31 so biggest thing you need is reviews? 21:08:33 anything else? 21:08:35 is this an ok time to bring up the status of nova-network? It does sort of impact exactly what needs to be ported for v3 21:08:43 (yes reviews is the biggest thing) 21:08:49 vishy: you around? 21:08:56 vishy has the "deprecate-nova-network" blueprint 21:08:56 yeah 21:09:13 multitasking 21:09:45 so there hasn't been much progress so far 21:10:00 afaik, not a ton of progress there, so it's a little up in the air ... 21:10:02 i'm still of the opinion that we should assume quantum for v3 21:10:03 v2 isn't going away yet 21:10:04 and hopefully we can kill it when we kill nova-network 21:10:05 happy to hear other opinions if someone wants to argue otherwise ... 21:10:12 sorry, had a network blip 21:10:12 it looks like there are some extensions around which don't support quantum and only nova-network. If we don't have nova-network in Havana then do they need to be ported? 21:10:17 russellb: +1 21:10:25 vishy: ok cool 21:10:37 i would leave all the network stuff out of v3 21:10:47 ok thanks 21:10:49 ok great 21:10:56 cyeoh: anything else? 21:10:59 I have a status update on the Cinder volume encryption blueprint https://blueprints.launchpad.net/nova/+spec/encrypt-cinder-volumes 21:11:09 We (APL) are currently working with Cinder folks to store encryption metadata (e.g., an encryption key id) with the volume 21:11:20 This is needed to support special Cinder features 21:11:24 russellb: just that re: the proposed in/out of core - anything without a -1 (which I think is everything now) we're just going ahead with 21:11:29 We expect to submit an updated version of our code next week, but it will be tight for getting it accepted into Havana-2 21:11:29 other than that nothing else 21:11:43 cyeoh: ok, i'd like to review each of those if i can, so tag me on the reviews 21:11:58 bpb: ok, cool. sounds like good progress, though! 21:12:04 russellb: ok 21:12:24 bpb: should we leave it on the havana-2 list, or go ahead and bump it? 21:13:24 bpb: well just let me know 21:13:27 comstud: you around? 21:13:35 comstud: you have a high prio blueprint on havana-2 for the native mysql db driver 21:13:43 i am 21:13:46 Russel: We're still shooting for Havana-2 - sorry for the spamming 21:13:47 comstud: that going to show up? or should we bump it? 21:13:54 bpb: cool, sounds good, thanks! 21:14:03 well 21:14:19 bpb: i'm afraid a batch of stuff will get bumped because of review bandwidth in the last days ... so we'll see 21:14:19 no, it's not going to make 2 i guess 21:14:24 ok 21:14:47 still havana-3? 21:14:55 yeah, i'll make it happen. 21:15:00 ok cool 21:15:22 dansmith: you have a ton of objects patches coming through, want to give an update? 21:15:23 i wanted objects to be stablized a bit before doing the work 21:15:28 even though it's not targeted at havana-2 21:15:28 well, committing the work... 21:15:30 russellb: yeah 21:15:34 comstud: makes sense 21:15:56 russellb: I was actually going to say, I'm pretty sure "objectify all the things" isn't going to make all of H anyway, 21:16:03 but to satisfy some people, I probably should also prove sqlalchemy is a huge problem. 21:16:06 heh. 21:16:12 (more than I already have) 21:16:16 so maybe we should cut up some smaller sub-blueprints to go underneath the main one 21:16:34 dansmith: yeah, or rescope the main one you have now to what you think is achievable for havana 21:16:41 "make all of H" ? "make H" 21:16:46 and then we can have icehouse-objectification or whatever 21:16:52 russellb: sure, I can do that 21:16:58 ok cool, sounds good to me 21:17:01 russellb: I'd like to try to get something small targeted for H2, 21:17:04 to show progress, 21:17:09 but I'm not sure what that would be right now 21:17:16 dansmith: ok, feel free to create a blueprint if you can think of something 21:17:19 okay 21:17:20 gosh i really hope I can step up my game here soon wrt objects 21:17:21 convert instance or something? 21:17:29 comstud: me too, sheesh man! :P 21:17:34 i know, right? 21:17:42 russellb: nah, that's too big I think 21:17:48 dansmith: k 21:17:56 well, just let me know what you come up with :) 21:17:57 russellb: "nova-api uses objects" maybe 21:17:59 yeah 21:18:04 but i like the idea of having something in havana-2 21:18:05 comstud: just kidding <3 21:18:10 comstud: he's not kidding 21:18:10 i know 21:18:13 i'm not 21:18:13 :) 21:18:18 hehe 21:18:18 haha, nah, I am, I promise 21:18:19 ok, on to something else 21:18:20 we've got a ton 21:18:37 alaski: hey, you've got a high prio one 21:18:49 making scheduler queried instead of a proxy service 21:19:15 There's one review up for a new scheduler call 21:19:39 Working on making conductor use it now. 21:19:53 But all the work should be under review soon 21:19:56 nice 21:19:59 sounds good then! 21:20:04 need anything? 21:20:08 (other than review on what you have up..) 21:20:08 money 21:20:12 i'd take money too 21:20:17 same as everyone else, reviews 21:20:24 yeah, i feel like we're behind on reviews 21:20:34 sort of a different topic though ... maybe we can come back to it 21:21:14 timello: hey, i saw you join 21:21:22 timello: how about your cold migrations to conductor refactor? 21:21:31 * johnthetubaguy waves 21:21:39 johnthetubaguy: hi! 21:21:56 johnthetubaguy: and your live migrations to conductor refactor? 21:22:16 in a related front, live-migrate has got some stuff in, other bits are in reviews, little bits not done yet 21:22:33 main sorry is the orchestration with the TaskFlow lib, can't see that getting in too soon 21:22:48 i wasn't considering that part of this personally 21:22:57 yeah, its probably next time sadly 21:23:03 could reconsider it in the Icehouse release 21:23:16 I was hoping to get tasks that detected the service got restarted, but yes, that should be I now 21:23:31 its looking more like H-3 I guess 21:23:33 still havana-2 achievable? 21:23:36 ok, i can move it 21:23:42 well, its damm close at least 21:23:53 leaving H-2 for now may help the reviews :P 21:24:11 there is stuff to review now at least, and bits to finish off 21:24:19 anyways, was wondering about the cold migration 21:24:42 yeah, pinged timello 21:24:51 timello: speak up if you're back and we can come back to cold migrations 21:24:59 cool 21:25:00 anyone else have a blueprint you'd like to give an update on? 21:25:05 or one you'd like to ask about? 21:25:23 in general, please help me out by making sure your blueprint status is accurate 21:25:24 https://blueprints.launchpad.net/nova/+spec/user-locale-api is "pretty much done" 21:25:38 just waiting on reviews and a change to merge in Oslo 21:25:44 johnthetubaguy let me know about how i can help make sure taskflow helps u :) 21:25:51 mrodden: cool, so need review ... 21:25:53 at your service ;) 21:26:01 looks like we have it marked properly 21:26:10 harlowja: will do, just looking like I now 21:26:10 and hopefully we can get it reviewed.. 21:26:29 yea. if it gets bumped to H3 if we don't have enough review bandwidth 21:26:40 could we bump the priority up in that case? 21:26:47 johnthetubaguy agreed 21:26:59 mrodden: yeah, that happened in G, right? that's fine 21:27:03 yep 21:27:08 mrodden: you'll have to remind me 21:27:17 np, i have people reminding me daily... 21:27:19 :) 21:27:29 heh 21:27:38 anyone else? 21:28:27 alright, next topic 21:28:28 #topic bugs 21:28:36 #link https://wiki.openstack.org/wiki/Nova/BugTriage 21:28:48 we have about 60 new nova bugs, and 47 novaclient bugs 21:29:06 we really need someone to take on triage of novaclient, kind of like we have owners for tags 21:29:08 dan and I filed a lot of bugs for work that needs to happen for objects 21:29:08 anyone willing? 21:29:17 not sure how many we left 'new' 21:29:24 (fyi) 21:29:25 comstud: cool, have a tag for them? 21:29:27 yes 21:29:29 comstud: objects? 21:29:29 unified-objects 21:29:33 k 21:29:44 so, i also noticed that my filter i've been using for untagged bugs is silly 21:29:59 it misses stuff if someone added their own tags 21:30:06 doh 21:30:15 like grizzly-backport-potential, or whatever random tags they come up with 21:30:17 russellb: yeah, do you know how to get that tag to show up in the list on the right? 21:30:18 so some cleanup to do ... 21:30:27 dansmith: i do, i have to make it an official tag 21:30:36 oh, pretty please? 21:30:36 dansmith: ping me later and i can 21:30:39 okay 21:30:47 so, in addition to really needing someone to own novaclient 21:30:56 network could use some work ... 17 network untriaged bugs 21:30:58 arosen: ^^^ 21:31:02 russellb: I'll help with novaclient bugs 21:31:07 s/work/help/ 21:31:17 alaski: awesome!!! much appreciated, it's something we've long neglected 21:31:25 alaski: can you update the wiki page? 21:31:30 sure 21:31:34 cool. 21:31:41 another is libvirt, 9 new bugs to triage there 21:31:56 api has 6 to triage 21:31:56 cyeoh: ^ 21:32:01 and compute has 5 21:32:15 melwitt1: ^ 21:32:23 hartsocks: 4 on vmware 21:32:28 I see 'em 21:32:37 do we want a bug day post H-2 to try and tidy up a bit? 21:32:37 4 for volume integration ... nobody has taken that tag yet 21:32:50 or should that be, pre H-2, maybe 21:32:53 johnthetubaguy: not a bad idea ... bug days seem to be helpful ... as long as they're not too often 21:33:04 +1 for not too often 21:33:07 we had a really good one, then another the next week and nobody did anything :) 21:33:14 heh 21:33:28 so we stopped, and then came up with this tagged approach to split up the work 21:33:35 so anyway, triage! 21:33:38 yeah, I remember now 21:33:41 any specific bugs worth talking about? 21:33:54 i found this one while triaging today: https://bugs.launchpad.net/nova/+bug/1194093 21:34:06 have a fix for master written i think, just need to add tests 21:34:25 either i'm totally missing something, or this hasn't worked in ages 21:34:32 guessing the latter 21:34:51 Quick question on triaging: if it's a feature request what's the best way to "close" it? 21:34:54 hmm, that was the feature no one remembered getting added right 21:35:12 alaski: good question ... typically put it as Confirmed+Wishlist if it seems reasonable 21:35:28 johnthetubaguy: has been there since the dark ages of 2011 at least 21:35:34 ah 21:35:42 alaski: more detailed notes on that kind of thing on the main BugTriage wiki page 21:35:43 Ok. I've been using Invalid+wishlist, but confirmed is nicer :) 21:36:03 alaski: yeah ... and then we have a BugTriage task on the main page to go back and close out old wishlist bugs 21:36:13 or, mark them Opinion+Wishlist i think 21:36:22 which is effectively closed 21:36:22 cool, that answers my question 21:36:23 yeah, I think we said invalid means like never look at again 21:36:38 Opinion is like ... "maybe" 21:36:48 "it's not you, it's me" 21:37:07 any other bugs? 21:37:07 haha 21:37:38 (i know we have plenty of them, but rather, any that you want to discuss?) 21:37:59 alright, subteam time them 21:38:01 #topic subteams 21:38:09 hartsocks: sup! 21:38:12 yo 21:38:24 o/ 21:38:33 So I'll spam y'all again with our reviews on Friday… on the mailing list. Fun right? 21:38:46 fine with me 21:38:48 \o/ 21:39:02 groovy. 21:39:11 so you need reviews :) 21:39:14 need anything else? 21:39:22 Some of us need to clean up our acts… after the last reviews :-) 21:39:48 heh 21:39:53 We're hashing out some of our BP's and such. H2 is going to be optimistic at the rate we're moving. 21:40:04 hartsocks: ok. let me know what should be bumped 21:40:10 I'll keep beating the drums. 21:40:24 i like drums 21:40:36 Our highest priority BP is: https://blueprints.launchpad.net/nova/+spec/multiple-clusters-managed-by-one-service 21:40:45 But we're still hashing out how to pull it off. 21:41:05 thought there was a patch up already? 21:41:12 ah, we were thinking cells at one point 21:41:14 just still WIP? 21:41:22 johnthetubaguy: yeah i'm still thinking that long term... 21:41:34 russellb: gotcha 21:41:49 johnthetubaguy: still some design work to do there to define a driver layer 21:42:02 The review is up. IIRC there's still one point to work out. 21:42:09 hartsocks: ok 21:42:29 johnthetubaguy: i'm interested in it, but haven't made much progress ... more just looking around and thinking about it some 21:43:39 russellb: cool, its a tricky one, think comstud mentioned federated cloud stuff too, probably worth trying for both 21:43:49 yeah 21:44:00 so ... may end up a design summit brainstorm session 21:44:06 we'll see 21:44:10 anyways, yes, lets not get distracted 21:44:14 k :) 21:44:18 hartsocks: anything else? 21:44:27 Nope. 21:44:31 kthx 21:44:34 n0ano: hi! 21:44:37 scheduler! 21:45:15 hi, not much to say, we had a very quiet week, I just came here to tell everyone nothing to tell :-) 21:45:36 k :) 21:45:40 harlowja: what's up 21:45:44 hey hey! 21:45:45 devananda: preemptive ping, you're up next 21:46:18 so mainly just continuing stablizing, and hoping to get 0.1 release out for h2 so that cinder can integrate, keeping in touch with heat folks about back and forth on heat usage 21:46:41 ok, and sounding like we won't be ready to look at using it soon ... 21:46:42 thinking it'd be nice to get more nova folks that have some ideas on what or how nova could use it involved 21:46:58 been busy getting the code into the right places where it could be considered 21:47:02 agreed 21:47:09 which is step #1 i think 21:47:12 harlowja: hopeing to get some time to help with that 21:47:19 thx much johnthetubaguy :) 21:47:42 so thats about it :) 21:47:45 cool 21:47:47 devananda: around? 21:47:55 russellb: hi! 21:47:59 \o/ 21:48:03 how goes ironic / baremetal? 21:48:23 so, not much to report progress wise. I'll be out next week (EuroPython) and we're probably not going to have anything ready for H3 21:48:36 until h3? 21:48:38 kind of in that middle-slump period right now :( 21:48:39 or not anything for h3? 21:48:44 not anything for H3 21:48:50 iow, not until early "I" at this point 21:48:54 ok 21:49:07 so, need to look over these baremetal havana blueprints that were kind of pending what happens with ironic 21:49:23 no rush, but when you have some time, let me know what we should do with them (if you can't update them yourself) 21:49:30 nova baremetal continues to get worked on by a few folks, particularly they have been adding better redhat support 21:49:45 yeah, saw a patch go up today from dprince i think 21:49:47 ack. i'll try to get to the BPs in the next week 21:49:52 cool 21:50:06 need anything? 21:50:21 not from nova 21:50:26 i have a ping out to glance folks for some reviews :) 21:50:31 k 21:50:37 is the db team still meeting? 21:50:50 dripton ostensibly took that over, but i haven't heard anything in a bit 21:50:53 ok 21:50:56 he's out this week 21:50:59 if it's not meeting and folks need me to restart it, let me know 21:51:05 boris-42 (and his team) has been doing a bunch 21:51:14 code wise i mean 21:51:15 hiii 21:51:16 ) 21:51:22 yea! i've been trying (and failing) to keep up with their flood of reviews :) 21:51:27 I will write today a big mail 21:51:30 in mailing list 21:51:33 boris-42: on status? 21:51:42 where I would like to describe not only status=) 21:51:47 but also future plans=) 21:51:50 cool 21:51:52 sounds good 21:51:52 and not only in nova=) 21:52:02 I think that IRC is not good enough for it=) 21:52:06 heh, just be careful not to make plans so big that you can't make progress :) 21:52:08 baby steps! 21:52:20 yeah=) I will write our baby steps!=) 21:52:25 cool 21:52:29 and while you're here 21:52:32 PCI passthrough? 21:52:35 yeah 21:52:35 working on that? 21:52:41 I publish yesterday 21:52:44 oh, cool 21:52:44 bunch of botch+) 21:52:50 didn't sllep for 2 days=) 21:53:07 to get all work + meeting + this stuff=) 21:53:07 hardcore 21:53:18 And now I am not able to sleep 21:53:35 and I drink already a lot of alcohol but it seems doesn't help=) 21:53:42 lol 21:53:43 to fat=) 21:53:51 for alchol=) 21:54:02 yeah ummm 21:54:13 hope you're able to sleep soon then :) 21:54:18 =) 21:54:33 #topic open discussion 21:54:48 so one thing that was mentioned a bunch was needing reviews 21:54:54 so, xenapi: no update really anyways 21:54:57 I hope we will have enough time in Havana to provide UC and DB Archiving in Cinder and Glance* 21:55:01 johnthetubaguy: sorry! 21:55:13 i came up with these stats to help determine if we were keeping up: http://russellbryant.net/openstack-stats/nova-openreviews.txt 21:55:17 russellb: no worries, its quite quiet, I got a few bits in 21:55:23 and have a goal of having average of < 4 days 21:55:30 according to the stats, we're "keeping up" 21:55:41 nice 21:55:45 but, 296 open 21:55:48 lol=) 21:55:58 I have about 7=) 21:56:04 some probably keep getting rebased and stuff, and that resets the counter 21:56:04 do we track anything from the last non-WIP or draft to merge? 21:56:21 I get lots of forced rebases, which might make it look like I was waiting less? 21:56:26 johnthetubaguy: the numbers are the wait time on the current patch rev 21:56:32 update it, restarts the counter 21:56:39 right. 21:56:45 by the way Russell 21:56:47 yeah, just wonder if its worth checking the other one too? 21:56:49 so ... not sure how to account for that 21:56:50 test-db-api is almost finish! 21:56:55 boris-42: awesome! 21:57:08