21:04:19 <russellb> #startmeeting nova 21:04:20 <openstack> Meeting started Thu Mar 14 21:04:19 2013 UTC. The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:04:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:04:23 <openstack> The meeting name has been set to 'nova' 21:04:26 <russellb> #chair vishy 21:04:26 <openstack> Current chairs: russellb vishy 21:04:33 <russellb> who's here to talk nova?! 21:04:37 <sdague> o/ 21:04:38 <dansmith> \o/ 21:04:56 * jog0 waves 21:05:10 <cyeoh> hi 21:05:26 <russellb> soooooooooo 21:05:42 <russellb> https://wiki.openstack.org/wiki/Meetings/Nova 21:05:45 <russellb> #link https://wiki.openstack.org/wiki/Meetings/Nova 21:05:47 <vishy> hi 21:05:51 <cburgess> greetings 21:06:04 <devananda> \o 21:06:05 <dansmith> cripes, what am I doing adding agenda items? 21:06:07 <vishy> russellb: fyi the traceback happens with local exceptions too 21:06:12 <russellb> vishy: really? 21:06:22 <dansmith> vishy: hmm, I'm confused then 21:06:27 <russellb> same here, what adds it 21:06:43 <russellb> so, we're talking about https://review.openstack.org/#/c/24475/ 21:06:54 <russellb> and specifically how the traceback gets into the exception message 21:07:34 <mikal> Hi 21:07:55 <russellb> but in the meantime we can discuss rc1 really quick 21:08:07 <vishy> i could be wrong 21:08:09 <russellb> in the release meeting, an ETA for grizzly-rc1 for nova was set for March 20th 21:08:12 <russellb> Wednesday 21:08:21 <russellb> vishy: yeah, we should get to the bottom of it 21:08:36 <russellb> to do rc1 on Wednesday, that means have everything done by Tuesday night 21:08:38 <vishy> russellb: you are correct 21:08:44 <alaski> o/ 21:08:51 <russellb> vishy: remote only? 21:08:56 <vishy> yes 21:09:01 <russellb> vishy: huzzah 21:09:01 <dansmith> heh 21:09:10 <russellb> #link http://old-wiki.openstack.org/release/rc/ 21:09:21 <russellb> our graph for the road to RCs looks pretty good :) 21:09:36 <dansmith> yeah, we're not losing anymore! 21:09:48 <russellb> #link https://launchpad.net/nova/+milestone/grizzly-rc1 21:10:05 <russellb> so, looking at that, i'm assuming all the In Progress ones are in good shape ... 21:10:15 <russellb> would be good to look at the other 4 21:10:16 <vishy> russellb: message = failure.get('message', "") + "\n" + "\n".join(trace) 21:10:46 <russellb> arosen: you around? 21:11:02 <russellb> would be good to look at the 4 bugs not listed as in progress yet 21:11:18 <dansmith> russellb: the one of mine on that list is in review now 21:11:19 <russellb> arosen has two bugs assigned, both in Triaged 21:11:39 <russellb> we have one issue not assigned 21:11:50 <russellb> so ideally we should get it assigned ASAP, or remove it from the blockers list 21:11:56 <russellb> and tag with "grizzly-rc-potential" instead 21:12:06 <russellb> dansmith: cool 21:12:18 <russellb> https://bugs.launchpad.net/nova/+bug/1143659 21:12:19 <uvirtbot> Launchpad bug 1143659 in nova "Deleting instance during snapshot leaves snapshot state in saving" [Medium,Triaged] 21:12:22 <russellb> rmk: around? 21:12:27 <rmk> yes 21:12:38 <russellb> rmk: you looking at that? looked like you had it assigned at one point? 21:13:07 <rmk> russellb: I unassigned it for myself because I don't know how to fix it 21:13:30 <rmk> Thought I could handle it with an API restriction but that's not what we want 21:13:43 <rmk> honestly i dont think that bug needs to be fixed for rc1 21:13:50 <dansmith> sdague: surely we can mark migration testing with data done at this point, right? 21:13:52 <russellb> rmk: ok, that was going to be my next question 21:14:01 <sdague> dansmith: I thought it was 21:14:04 <rmk> russellb: It's a corner case and not really a big deal imo 21:14:05 <russellb> if it's not a regression, and not the end of the world kind of bug, we can probably not block on it IMO 21:14:07 <rmk> right 21:14:13 <russellb> rmk: ok cool, i'll untarget then :) 21:14:13 <dansmith> sdague: oh, it is, just targeted so it's showing up.. nevermind 21:15:18 <russellb> https://bugs.launchpad.net/nova/+bug/1100462 21:15:19 <uvirtbot> Launchpad bug 1100462 in nova "Instance task_state being left as migrating after InvalidSharedStorage error" [Medium,Triaged] 21:15:26 <russellb> doesn't appear to be actively workedon ... 21:15:32 <russellb> vishy: thoughts on moving that to grizzly-rc-potential? 21:15:56 <vishy> seems fine 21:16:22 <russellb> k, sounds like it affected folsom too, and we've lived this long without fixing it, heh 21:16:25 * russellb untargets 21:16:32 <vishy> arosen is looking at the other two but the second one looks super easy with a patch already in place if someone want sto steal it 21:16:34 <russellb> vishy: how about the other 2 assigned to arosen, i think i saw you guys talking earlier ... 21:18:01 <russellb> vishy: ack, ok, i guess i can steal that 21:18:34 <russellb> luckily the patch is trivial, otherwise the lack of being able to submit to gerrit officially would be a problem 21:18:56 <russellb> (happy to let someone else steal it if someone else wants to) 21:19:15 <russellb> any other bugs / rc1 talk? otherwise looks like dansmith had a topic for today 21:20:03 <russellb> alrighty, then.... dansmith, what's up. 21:20:06 <dansmith> so, 21:20:13 <dansmith> I think I put this on last week out of sheer frustration, 21:20:25 <dansmith> and because we had gone a couple weeks without a lot of meeting fodder 21:20:40 <dansmith> we have a bunch of tests that fake a lot of instance_type stuff, 21:20:51 <dansmith> lots of modifying individual attributes of a type that we pass in and such 21:21:22 <dansmith> and there are a ton of places where we use inconsistent types, and where the instance has memory_mb=1024 but the type we pass to some call with it shows memory_mb=512 21:22:00 <dansmith> it makes it hugely problematic when we go to unify some of the instance_type stuff and have to patch up a bunch of tests to change all these strange assumptions that are made 21:22:21 <dansmith> I had a lot of rant material last week, but they're not as fresh for me now 21:22:27 <russellb> :) 21:22:30 <russellb> go go go 21:22:36 <russellb> but yeah, sounds like a good thing to clean up to me. 21:22:38 <dansmith> but basically, I think we need a BP for havana to seriously clean up some of that stuff 21:23:19 <mikal> Sounds like a good idea to me 21:23:24 <dansmith> </rant> 21:23:33 <russellb> anyone else have a rant? 21:23:37 <dansmith> heh 21:23:38 <jog0> dansmith: ++ to Havana cleanup 21:23:48 <russellb> i think mikal does 21:23:56 <mikal> I hate everything 21:24:00 <mikal> Is that specific enough? 21:24:03 <sdague> not a rant, but another thing dansmith and I talked about was having a havana blueprint to get rid of all stack traces in a successful tempest run 21:24:04 <russellb> yep, thanks 21:24:04 <dims> lol 21:24:10 <mikal> We really really need a security review gerrit 21:24:15 <russellb> sdague: +1 21:24:17 <mikal> This whole current arrangement is bonkers 21:24:19 <sdague> because if everything worked, there shouldn't be stack traces in our logs 21:24:26 <sdague> mikal: +1000 21:24:27 <dims> sdague, need to revisit the SKIP'ped tests too 21:24:31 <russellb> sdague: absolutely 21:24:57 <russellb> should never see a stack trace unless something very bad or unexpected happened 21:25:00 <sdague> dims: yeh, we actually have a tempest tool to query launchpad for closed bugs 21:25:08 <sdague> we run it from time to time 21:25:13 <russellb> as in, every traceback you see in a log == bug IMO 21:25:21 <sdague> russellb: yes, definitely 21:25:32 <jog0> russellb: ++ 21:25:34 <dansmith> no way, I have to abandon my periodlic_task_log_a_stack_trace() patch? 21:25:43 <sdague> I'm working up some tooling to make the stack trace identification easier 21:25:53 <dripton> Just make tempest grep the output for Traceback and fail the test if it's there. That'll stop all work for a couple of days then it'll be fixed. 21:26:09 <sdague> dripton: well, eventually 21:26:18 <sdague> we have to get rid of the current ones first 21:26:28 <dims> right 21:26:40 <russellb> sdague: i was pretty surprised when you linked to that log the other day 21:27:01 <sdague> russellb: it's nothing new 21:27:12 <sdague> I just decided to get beligerent about it :) 21:27:28 <russellb> sdague: and i just don't look unless tests failed, heh 21:28:17 <russellb> so speaking of blueprints and such ... have people been filing design summit sessions? 21:28:39 <dansmith> yeah, I have 21:28:40 <mikal> There are a fair few nova ones 21:28:43 <devananda> lots 21:28:46 <mikal> Not sure how the number compares to last time though 21:28:57 <sdague> there are a couple more api ones cyeoh and I need to get in 21:28:59 <russellb> 29 for nova, in a quick count 21:29:02 <mikal> If we have a gap, we could always go get a beer... 21:29:15 <mikal> Is cyeoh going to be able to make the summit? 21:29:18 <russellb> i don't think we'll struggle to fill the nova track ... 21:29:21 <sdague> sadly no 21:29:26 <mikal> sdague: :( 21:29:59 <dansmith> mikal: he talks funny anyway 21:30:08 <dims> speaking of which, how can i help when i am at the summit? 21:30:08 <cyeoh> mikal: hopefully the next one! 21:31:05 <mikal> cyeoh: yay! 21:31:15 <mikal> dims: attend sessions and have sane ideas? 21:31:34 <sdague> mikal: that's crazy talk... 21:31:35 <russellb> and if your ideas aren't sane, keep them to yourself? lol. 21:31:39 <dims> mikal, that's a give :) hopefully sane ideas 21:31:58 <jog0> are we planning on revving the API in Havana? 21:32:09 <russellb> i was thinking of writing a blog rant about design summit best practices / etiquette / things that annoy me :) 21:32:10 <devananda> definitely share the crazy ideas ;) 21:32:18 <mikal> Thre seem to be a lot of EC2 API tweaks listed in bugs at the moment 21:32:24 <mikal> It would be good to know what the plan is with those 21:32:32 <russellb> mikal: "plan" ? 21:32:39 <mikal> i.e. incorrect return codes, wrong results, etc 21:32:44 <russellb> for them to sit there until someone cares enough about ec2 to work on them? 21:32:50 <dansmith> Plan: assign to mikal 21:32:54 <mikal> I get the feeling theres an IBM tem who intends to fix them all 21:32:56 <russellb> i've noticed a significant lack of people caring about EC2 this dev cycle ... 21:32:59 <mikal> I tried to fix one and was told off 21:33:13 <russellb> mikal: I just -2 anything from IBM 21:33:19 <dansmith> russellb: me too 21:33:28 <mikal> Heh 21:33:33 <mikal> Are still logging this meeting? 21:33:35 <russellb> but really, how much EC2 work happened in grizzly? 21:33:36 <vishy> i'm surprised there isn't more help 21:33:42 <russellb> mikal: yes. 21:33:45 <rmk> I want to rant too 21:33:46 <russellb> vishy: me too 21:33:49 <rmk> My rant is about dan's rant 21:33:53 <vishy> jog0: no love from you guys on ec2 fixes ? 21:33:54 <rmk> jk 21:33:55 <russellb> vishy: people scream for it, but it's been dead this cycle 21:34:01 <mikal> russellb: I get the feeling people walked away around the folsom timeline 21:34:17 <dims> russellb, means it's working perfectly! 21:34:26 <russellb> dims: see bug list, heh 21:34:27 <mikal> Let me rephrase -- we should try and kill off those bugs in Havana 21:34:28 <vishy> we need to really get on v3 api early in h 21:34:36 <vishy> so we can work out the kinks 21:34:42 <dims> russellb, yea, i have been tagging them 21:34:52 <vishy> after watching what happened recently with keystone v3 21:34:54 <russellb> dims: cool 21:34:56 <vishy> i'm a bit scared 21:35:04 <sdague> vishy: yes, agreed 21:35:09 <russellb> vishy: oh dear, yeah, that's an interesting situation 21:35:11 <vishy> also we should spend some effort cleaning up novaclient 21:35:22 <jog0> vishy: its on my list but I never seem to get that for down it 21:35:28 <rmk> nova-client focus.. yes 21:35:29 <vishy> like have it detect which extensions are enabled a 21:35:40 <devananda> ^ ++ 21:35:41 <russellb> would be nice for someone to just decide to make novaclient their focus 21:35:41 <uvirtbot> devananda: Error: "++" is not a valid command. 21:35:43 <rmk> like get-vnc-console <type> 21:35:49 <rmk> or the net and network commands 21:35:49 <vishy> right now it is somewhat annoying 21:35:56 <russellb> oh oh oh, yes what rmk said 21:36:01 <mikal> We need to close bugs in general 21:36:02 <russellb> network-create and net-create >_< 21:36:04 <mikal> ITs a bit embarrssing 21:36:07 <russellb> which both say "Create a network" 21:36:23 <rmk> all this great work happens in nova but we don't make it too easy to consume with the client 21:36:32 * dansmith notes that he started a group rant-about-everything 21:36:38 <rmk> i blame dan for this 21:36:43 <russellb> dansmith: yes! 21:36:45 <mikal> dansmith: good work 21:36:49 <jog0> with the V3 api for H, the client will need to detect that too 21:36:53 <dansmith> =D 21:37:36 <sdague> jog0: yep, good fodder for the session 21:37:56 <mikal> We should have a rant session at the summit 21:38:00 <rmk> nova get-spice-console be3968fd-b5a3-484e-a38a-25d6e64d7edb spice-html5 <--- weird 21:38:01 <jog0> btw if anyone is interested in working on the CLI we now have integration tests for it, well very simple ones https://github.com/openstack/tempest/tree/master/cli 21:38:05 <mikal> Try and keep the ranting contained to that one sacraficial session 21:38:17 <rmk> oh oh can I rant about imagebackend 21:38:43 <sdague> isn't that called "rant and drink at the nebula party"? 21:39:58 <russellb> rmk: don't ask, just rant :) 21:40:33 <russellb> OH 21:40:46 <russellb> 1 week ago today we had 547 open bugs for nova 21:40:50 <russellb> right now we have 478 21:40:51 <russellb> \o/ 21:41:19 <russellb> 100 of which are marked as in progress. 21:41:31 <dripton> At this rate they'll all be fixed in 7 weeks! 21:41:52 <vishy> we should have a buglist zero effort 21:41:56 <vishy> that would be fun 21:42:20 <vishy> especially if we had a giant wallclock of the bug count somewhere 21:42:41 <sdague> i've thought about building something like that with a raspberry pi 21:43:12 <russellb> also, 160 in Fix Committed 21:43:19 <russellb> (not counted in the 478) 21:43:38 <russellb> but that means we've merged 160 fixes that actually had a launchpad bug since grizzly-3 21:43:39 <vishy> h-1: no new features until bugcount 0 21:43:40 <sdague> russellb: you driving bug friday again tomorrow? 21:43:42 <russellb> so, yay. 21:43:45 <russellb> vishy: lol 21:43:48 <vishy> :) 21:43:56 <russellb> sdague: i'm going to be out tomorrow, so i asked mikal if he'd like to 21:44:00 <sdague> cool 21:44:02 <mikal> vishy: there are still diablo bugs in there... 21:44:11 <russellb> vishy: should propose that on the -dev list 21:44:14 <sdague> there are bugs before diable in there 21:44:15 <mikal> Yes, I will be ranting about bugs tomorrow 21:44:21 <sdague> diablo 21:44:28 <sdague> there are some crazy things at the end of the bug queue 21:44:33 <mikal> Although, its a bit hard when we have a rc coming out 21:44:39 <russellb> some of it is still correct though, heh 21:44:53 <sdague> mikal: honestly, just triage would help 21:45:09 <mikal> sdague: well, I suspect we need a round of re-triage 21:45:12 <russellb> mikal: doesn't hurt to sneak in more fixes 21:45:14 <mikal> Some were triaged a very long time ago 21:45:19 <sdague> russellb: some of it refers to directories we don't have any more :) 21:45:27 <mikal> russellb: depends on how safe those fixes are 21:45:30 <russellb> sdague: haha yes, i close those as Invalid :) 21:45:34 <russellb> mikal: true 21:45:48 <mikal> Oh yes, please please please close invalid bugs if you find them 21:46:39 <russellb> mikal pays in goats if you close bugs 21:46:51 <mikal> Heh 21:46:51 <russellb> it's kinda weird 21:47:09 <sdague> :) 21:47:21 <mikal> russellb: http://www.amazon.com/Webkinz-Plush-Stuffed-Animal-Mountain/dp/B001QYPPO0 21:47:26 <russellb> random other reminder ... if you haven't voted in the PTL election, you should, and it ends tonight. 21:47:45 <russellb> mikal: only 14 in stock, what will you do? 21:47:49 <dims> there are 59 in incomplete status, what can we do with them? 21:48:05 <russellb> dims: https://wiki.openstack.org/wiki/BugTriage 21:48:15 <russellb> see Task 4 21:48:36 <dims> cool 21:48:50 <russellb> task 9 is new 21:48:54 <dims> will go thru those 21:49:04 <russellb> there's still a lot that can be done with task 9 21:49:25 <russellb> but in general, the tasks are ordered by priority 21:49:50 <sdague> yeh, typically just chatter away on -nova during triage to ask questions if you have doubts on anything 21:50:27 <dims> sounds good 21:50:36 <russellb> mikal: going to send out a bug day email? 21:51:12 <russellb> i'd send, but i think it makes sense to come from someone who's going to be around and banging the bug drum :) 21:51:36 <mikal> I was going to do it tomorrow morning 21:51:43 <mikal> Do you normally do it the day before? 21:51:43 <russellb> k, cool 21:51:56 <russellb> whichever 21:52:04 * russellb had been doing it day before 21:52:17 <russellb> any other comments / questions / rants ? 21:52:42 <russellb> review plugs? 21:52:45 <russellb> heh 21:52:49 <dansmith> shameless! 21:53:38 <russellb> wait until master opens back up 21:53:43 <russellb> the review plugging will flood back in :) 21:53:56 <russellb> thanks for hanging out everyone 21:54:00 <russellb> #endmeeting