21:01:39 #startmeeting project 21:01:40 hi 21:01:41 Meeting started Tue Feb 19 21:01:39 2013 UTC. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:42 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:44 The meeting name has been set to 'project' 21:01:46 Agenda @ http://wiki.openstack.org/Meetings/ProjectMeeting 21:01:49 #topic General announcements 21:01:51 ttx: i have a cross-project gating concern to bring up 21:02:02 sure, just a sec 21:02:02 can you add it to the agenda? 21:02:15 we'll do it at the end of general discussion 21:02:20 If we don't change anything, you have until the end of the day (i.e. tomorrow morning in Europe) to merge features for Grizzly... 21:02:35 But given that the gate was not exactly in stellar condition today, looks like some /could/ benefit from pushing back one day 21:03:00 does anyone thinnk that would make things better ? There is a lot left on the table right now 21:03:02 ttx: my concern relates specifically to that 21:03:17 vishy: ok, then go ahead :) 21:03:21 ttx: i'd rather just see more things get a 1-day extension 21:03:37 ttx: if we leave the door wide open, i'm not sure the gating will be any better tomorrow. 21:03:38 At the very minimum one day would help to get all approved stuff processed. 21:03:38 ttx: i think we should add a day and temporarily turn off the full tempest run on the gate 21:03:45 it won't affect much from Glance's point of view, but I see it being super-helpful to Nova 21:03:47 it is making the merge cycle too slow 21:04:26 I'd be good with that for the last bit here - it's really been grinding quite a bit 21:04:30 it's been a bit problematic since it was enabled, but amplified this week 21:04:32 vishy: the trick is... are you sure you can reenable it afterwards ? i.e. how to be sure you dodn't break anything ? 21:04:53 jeblair, mordred: ping 21:04:57 ttx: we leave it in the check so we will see failures pretty quickly 21:05:04 the gate does seem to be getting really backed up 21:05:10 even aside from the pypi issues 21:05:28 *nod* indeed 21:05:34 sdague: thoughts ? 21:06:03 well, my thoughts are a lot of people showed up out of the woodwork in the last 72 hours to try to get code merged :) 21:06:24 sdague: indeed. Important stuff shoudl have been merged ages ago 21:06:25 I generally think it's not a great idea, but I also understand the concern to get stuff merged 21:06:27 we need a longer term solution for improving the gate time, but I'm thinking a temporary fix to get the current backlog cleared would be helpful. 21:06:27 sdague: I think that happens for deadlines :) 21:06:33 sdague: but others were saying that it has taken 8 hours to a couple days to get stuff merged even before the last days 21:06:41 it hasn't *only* been a problem this week 21:06:54 russellb: +1 21:06:59 russellb: we've been on a big run up the last 2 weeks 21:07:09 it would be good to pull some stats at some point on that 21:07:21 vishy: did you run the idea to the CI folks ? i.e. how doable is it ? 21:07:40 clarkb said it would be pretty easy to temporarily disable / post-run it 21:07:43 Is this because a single Tempest run occasionally takes 8 hours to run or is it because of sequencing of patches or something entirely different? 21:07:49 personally I'd much prefer to postpone stuff to H and reduce the number of things that we ahve to process 21:07:55 overall I don't like the idea, because I feel like it turns off an important net. But I do get we need to balance it with stuff that needs to get merge. 21:07:58 rather thah turning off checks and praying 21:08:00 soren it is the serialization of patches + 1.5 hours per run 21:08:10 soren: single run doesn't take that long, but a single failure causes everything behind it to have to run again 21:08:42 if we leave folks feeling their code didn't get merged in grizzly because of the gate ... 21:08:49 vishy's proposal makes sense to me 21:08:51 well its possible not necessaroly very easy. we would have to stop zuul to reload its config 21:08:53 markmc: right that is my concern 21:09:00 jeblair ^ 21:09:02 russellb: I know it doesn't *usually* take that long. I just wondered if it occasionally did. IOW whether this was a tempest problem. 21:09:18 ok. let's not spend the whole meeting discussing that 21:09:19 soren: *nod* sorry for stating obvious :) 21:09:22 there is stuff that has been churning back and forth and i would hate to postpone it for gate reasons 21:09:32 going forward, perhaps that's something that should be run only on a regular basis on the tree? i.e., nightly, we run that set of tests and log bugs for any errors that crop up? 21:09:33 let's discuss this in #openstack-dev after this meeting 21:09:39 in all cases we add a day ? 21:09:42 There is a tempest hourly build. If folks acted like that breaking was the same as a gate failure for urgency, that could be a compromise. 21:09:44 we could just be more open to FFEs of course 21:09:46 Shit *will* happen on the last day, whenever that last day is. 21:09:55 davidkranz: we turned off the hourly 21:09:59 to save CI resources 21:10:06 sdague: Heh. 21:10:06 because we had the gate 21:10:18 sdague: We could reverse that temporarily. 21:10:27 please. Let's discuss that in 50min in #openstack-dev. We have lots of ground to cover in meeting 21:10:40 in all cases we add a day ? 21:11:07 yes 21:11:08 +1 for adding a day imo, fwiw 21:11:19 ok 21:11:20 +1 21:11:24 #info Feature Freeze pushed back to EOD Wednesday, Feb 20 21:11:31 #info Starting Thursday, only features granted a FeatureFreeze exception should be approved. 21:11:37 In this meeting we'll review the remaining open blueprints and see if they should be granted an exception if they fail to make it tomorrow 21:11:46 So it's time to be strong and start saying "no". 21:11:55 noooooo 21:12:00 markmc, mordred, annegentle, sdague/davidkranz/jaypipes: quick updates from Stable/CI/QA/Docs teams ? 21:12:24 Book Sprint Next Week! 21:12:35 Discussion on opportunity to turn off some gate tests at 22:00 UTC in #openstack-dev 21:12:37 nothing to report from me on stable, 2012.2.4 scheduled for 2013-04-11 21:12:45 great, thx 21:13:17 #topic Oslo status 21:13:23 #link https://launchpad.net/oslo/+milestone/grizzly-3 21:13:36 2 open blueprints: 21:13:46 cfg-move-opts-between-groups 21:13:46 advanced-matchmaking 21:13:49 the first will definitely make it, it's basically got 2 +2s 21:13:57 the second I'm behind on 21:14:07 there is a patch, some concerns about testing 21:14:08 should the other just be deferred to H if it fails to make it in time ? 21:14:19 it would be a fine FFE, unlikely to impact anything else 21:14:26 and has been in the works for quite a while 21:14:35 slow reviewers was mainly the problem (including me) 21:14:36 I'm reluctant with FFEs in oslo libraries 21:14:50 that one will only affect zmq rpc driver 21:14:57 so i think it's ok for FFE 21:15:08 now, there is the question of ... 21:15:18 markmc: how much time do you need ? 21:15:19 code getting into oslo-incubator doesn't mean the code is synced to projects yet 21:15:28 true story. 21:15:36 ttx, I can probably make a call on it tomorrow 21:15:46 markmc: ^ that's what I mean by being reluctant with oslo FFEs 21:15:49 the other oslo thing 21:15:55 you still need to sync the code in 21:15:55 is making projects use oslo-config 21:15:59 https://review.openstack.org/#/q/I4815aeb8a9341a31a250e920157f15ee15cfc5bc,n,z 21:16:16 nova, glance and cinder patches all still need to be merged 21:16:27 being bitten by gate issues and lots of conflicts 21:16:34 should make it tomorrow 21:16:38 markmc: ok, would be great to get them in gate tomorrow :) 21:17:12 Looking at g3-targeted bugs, 1 bug left. Is it G3-critical, or should it be moved to the RC1 buglist ? 21:17:27 yeah, I must ping mikal about that 21:17:27 (or deferred ?) 21:17:32 it's probably not all that critical 21:17:45 BTW: making projects use oslo-config for the venv setup is a real PITA 21:17:56 Anything else on the oslo topic ? 21:17:58 merged that in this morning, and it promptly bit us hard 21:18:02 heckj, how do you mean? 21:18:18 You need oslo-config to run the process which installs the project depedencies. 21:18:24 chicken and egg kind of issue 21:18:29 that ^^ 21:18:34 ah 21:18:40 yeah, that caught me by surprise 21:18:48 * markmc isn't even sure why it's using cfg 21:18:51 will investigate 21:19:05 markmc: I'll find the specifics and ping you separately 21:19:06 it's not used by jenkins 21:19:09 since everything is tox 21:19:13 that's what I tested with too 21:19:30 ok 21:19:41 heckj, thanks 21:19:41 #topic Keystone status 21:19:46 heckj: o/ 21:19:51 #link https://launchpad.net/keystone/+milestone/grizzly-3 21:20:12 A lot of stuff still in progress here... 21:20:14 heyoo 21:20:16 Given recent progress I'm not really convinced giving it a few more days will get them all merged... 21:20:28 What do we absolutely need to have in grizzly release to make it coherent ? 21:20:53 actually, I'm pretty confident about all but the trusts piece 21:21:11 confident as in... merging today ? 21:21:12 The final reviews have been chugging through nicely, and all the devs are in sync 21:21:16 yeah, as in merging today 21:21:22 good news 21:21:25 trusts… well, here's the hat coming out 21:21:35 I'd like to ask for a featurefreeze extension for trusts 21:21:42 how much time do you need ? 21:22:10 ayoung was asking for end of the week earlier 21:22:21 (friday) 21:22:35 I'm ready to grant anything... just need to be convinced that it will make it. Been promised this for "the end of the week" quite a few times alareday 21:22:46 already 21:22:49 ayoung - you here to speak on your coding behalf? 21:22:56 (he was very head's down earlier today) 21:23:12 also that's only if the rest really makes it, obviously :) 21:23:32 yep 21:23:34 #infor considering FFE for trusts 21:23:43 You also have two in the grizzly series without milestone set: 21:23:52 domain-name-spaces and domain-scoping (completed?) 21:23:56 what about them ? 21:24:13 ttx: s/#infor/#info/ 21:24:29 #info considering FFE for trusts 21:24:34 Vek: thanks 21:24:40 welcome. 21:24:48 both are going in with the final bits of authn for V3 21:24:55 which we're expecting merged this afternoon 21:25:05 heckj: ok, I see. 21:25:26 most of it was already there, just pending the auth side of it 21:25:26 Looking at targeted bugs, 5 bugs on list. Should any of those be fixed in G3, or can I move them all to the RC1 buglist ? 21:25:39 all can be safely moved to the RC1 buglist 21:25:43 ok thx 21:25:47 Anything more about Keystone ? 21:26:16 that's it from me 21:26:17 #topic Swift status 21:26:21 notmyname: o/ 21:26:23 #link https://launchpad.net/swift/+milestone/1.8.0 21:26:31 Let's make it quick ;) Is 1.8.0 starting to crystallize ? 21:26:47 ya, we're fine, AFAIK 21:27:02 sometimes in first half of March ? 21:27:20 what's the final date for inclusion on grizzly? 21:27:39 we'll cut the release when it's ready. I'm still anticipating mid-march 21:27:44 The first RCs are expected in first half of March 21:27:48 ok 21:27:51 notmyname: Anything you wanted to mention ? 21:28:10 nothing omes to mind 21:28:14 #topic Glance status 21:28:17 notmyname: thx 21:28:20 bcwaldon: o/ 21:28:23 #link https://launchpad.net/glance/+milestone/grizzly-3 21:28:24 ttx: hey! 21:28:28 lalalalala 21:28:45 rainya: you're on thin ice 21:28:49 looks like this was trimmed down 21:28:58 ttx: yes, actively working on it 21:29:05 4 blueprints left 21:29:07 ttx: I want an FFE for image locations and image sharing 21:29:17 the domain logic bp has one patch left that markwash won't let me see 21:29:23 and common image properties is in review 21:29:38 bcwaldon: how much time do you need for those two ? 21:29:45 ttx: well, that's hard to answer 21:30:01 ttx: image sharing has several reviews up, so I expect that to be early next week at the latest 21:30:10 ttx: image locations is dependent on me not being lazy tonight 21:30:23 but I would give that the same target 21:30:38 #info FFEs for glance-api-v2-image-sharing and multiple-image-locations 21:30:52 yes 21:30:54 all the rest will be deferred if it doesn't pass the deadline tomorrow 21:31:13 already did so for a couple of bps 21:31:20 those bugs can bump to rc1 too 21:31:21 saw that. Thanks 21:31:31 all bugs can bump ? ok. 21:31:47 including bug 1102476 ? 21:31:48 Launchpad bug 1102476 in glance "Latest json-patch draft (10) is incompatibile with Images API v2 implementation" [Critical,Triaged] https://launchpad.net/bugs/1102476 21:31:52 ttx: to rc1 21:32:00 or am I thinking of things incorrectly 21:32:00 Anything more on Glance ? 21:32:10 not from me 21:32:24 #topic Quantum status 21:32:42 danwent: hi! 21:32:44 #link https://launchpad.net/quantum/+milestone/grizzly-3 21:32:48 hello 21:32:54 do i need to remind you? 21:32:58 no I read 21:33:05 :) 21:33:22 3 High stuff 21:33:28 lbaas-namespace-agent 21:33:28 lbaas-haproxy-driver 21:33:28 multiple l3 and dhcp agents 21:33:37 which ones do you need FFEs for ? 21:33:54 I do not expect to need an FFE for the l3/dhcp agents one 21:34:02 we will need one for the lbaas blueprints. 21:34:06 danwent: you mean, should make it in time ? 21:34:18 ttx: yes, should make it by tonight 21:34:37 I'll yell if it doesn't 21:34:42 i'm still planning on closing gate to non-high blueprints tonight 21:34:49 we've already had too much last minute stuff thrown in 21:34:52 #info FFe for lbaas-namespace-agent / lbaas-haproxy-driver 21:34:56 and I don't want to encourage it more :) 21:35:05 yes. 21:35:14 The remaining ones are all Medium/Low, should be deferred to H if they don't make it tomorrow. 21:35:20 we also have an important nova blueprint that I want to highlight 21:35:30 https://blueprints.launchpad.net/nova/+spec/nova-quantum-security-group-proxy 21:35:45 vishy and markmc have been giving it love though, so hopefully its in good shape 21:35:49 21 bugs in your g3-targeted buglist. Should any of those be fixed in G3, or can I move them all to the RC1 buglist ? 21:36:11 We'll have a short G3 gestation period due to pushing back FF one day 21:36:12 that's what i'm figuring out today. will work with markmclain, our bug master on that 21:36:30 well, quantum will have a bonus bug day then :P 21:36:49 danwent: could you keep me posted on that ? 21:36:58 ttx: on moving bugs out of g-3? 21:37:12 yeah. or move them all to RC1 21:37:20 (the milestone was created) 21:37:28 yes sir. will do it today. some bugs have already been moved over. 21:37:29 I need to know if I should block on it 21:37:38 Anything else on Quantum ? 21:37:38 i'm not aware of any blockers 21:37:44 but will recheck 21:38:03 nope 21:38:08 #topic Cinder status 21:38:18 danwent: thx! 21:38:21 jgriffith: hi! 21:38:24 #link https://launchpad.net/cinder/+milestone/grizzly-3 21:38:30 hello 21:38:44 Should be good 21:38:44 let's see, 7 left 21:38:49 especially with another day 21:38:58 thingee will likely have the two docs ones tonight 21:39:06 I'll have snap meta finished tonight as well 21:39:19 The others just need to get through the gates and rebasing 21:39:20 which ones should get an FFe if by pure lack of luck they don't make it ? 21:39:37 Maybe just snapshot-meta AFAIC 21:39:41 * ttx 's best friend is called Murphy 21:39:53 And the Fibre Channel 21:40:05 Oh... backups 21:40:06 crap 21:40:09 best friend. Worst enemy. same difference... 21:40:24 Ok... those three, but that's it! 21:40:43 jgriffith: if you end up needing more than one I'll shoot the others randomly 21:40:44 But I expect them all to make it in on their own 21:40:53 K 21:41:23 #info preemptive FFe for volume-backups update-snap-metadata fibre-channel-block-storage, if only one is needed 21:41:38 ttx: thx 21:41:53 #info preemptive FFe for volume-backups update-snap-metadata fibre-channel-block-storage, if only one is neededit 21:42:00 arh 21:42:04 heh. 21:42:19 1 bug in the targeted bug list 21:42:32 yeah, thingee is on it 21:42:35 Should be done tonight 21:42:42 ok 21:42:48 Anything more in Cinder ? 21:42:53 Nope 21:42:55 #topic Nova status 21:42:59 vishy: o/ 21:43:02 #link https://launchpad.net/nova/+milestone/grizzly-3 21:43:07 hi 21:44:08 OK, so which one(s) of the remaining stuff would you ask an FFe for if they fail to make it tomorrow ? 21:44:24 ok so going down the list 21:44:47 real unique keys I don't know if we can mark complete if the next patch merges since there are more unique keys we need 21:45:12 the api-samples additions we should just target to rc-1 21:45:13 no, just defer it to havana for more unique keys work 21:45:29 vishy: seems like something we could just re-scope and call "complete", and open a new bp for what we want to get done in havana 21:45:31 they are testing additions so they don't need feature freeze 21:45:32 vishy: yeh, I was talking with boris, it's one of 10 21:45:37 russellb: that is fine with me 21:45:39 #2 and #3 are testing 21:45:40 russellb: ++ 21:45:53 vishy: +1 21:45:58 sdague: i notice that his tests actually failed 21:46:09 migration testing sdague thoughts on that one? 21:46:20 vishy: I added a new test in boris's code, we need to resolve something 21:46:34 if we can't get https://blueprints.launchpad.net/nova/+spec/nova-quantum-security-group-proxy in by tomorrow i think an ffe is reasonable 21:46:35 vishy: I'd say we've got about 50% of the interesting cases covered 21:46:49 powervm-compute-resize-migration is marked completed but has https://review.openstack.org/#/c/22243 open 21:46:50 I figured the rest we could get after g-3 as it's test additions 21:46:58 nova-quantum-security-group-proxy 21:47:04 sdague: ok so migration testing is rc1 also 21:47:04 #info ffe for nova-quantum-security-group-proxy 21:47:08 i think the quantum one is FFE worthy 21:47:27 the quantum one is big, so it should have eyes if it's getting an FFE 21:47:30 i think this one is also worthy of an ffe if it doesn't make it: https://blueprints.launchpad.net/nova/+spec/default-rules-for-default-security-group 21:47:41 although that one probably will make it 21:47:51 vishy: +1 to the last one, but I think it's just about in 21:48:02 #info FFE for default-rules-for-default-security-group (if needed) 21:48:05 utilization i think we decided to push 21:48:11 vishy: anything else ? 21:48:15 the fibre channel one will make it 21:48:25 it just got owned by gating twice 21:48:32 the hotplug one is close 21:48:41 rxtx should be in 21:48:42 owned by gating. 21:49:17 vishy: 9 bugs in your g3-targeted buglist. Should any of those be fixed in G3, or can I move them all to the RC1 buglist ? 21:49:20 I'm here....and yes, I was heads down 21:49:21 so i think we're good 21:49:25 looking at bugs 21:49:32 bug 1130146 is critical 21:49:34 Launchpad bug 1130146 in nova "Unit tests broke with recent commit" [Critical,In progress] https://launchpad.net/bugs/1130146 21:49:44 the critical one has approved patches 21:49:46 in line to merge ... 21:49:51 ayoung: how much time do you need for trusts 21:50:06 the ultimate bug is in testr, though, I think. 21:50:24 Vek: ah, well workaround in line to merge at least 21:50:25 although what's merging covers legitimate bugs. 21:50:29 vishy: ok, so review the bugs and defer to RC1 anything not g3-critical 21:50:35 k 21:50:50 vishy: powervm-compute-resize-migration is marked completed but has https://review.openstack.org/#/c/22243 open 21:51:12 looks like it is completed 21:51:13 ttx, We are close....It depends on how brutal dolph is in the code reviews 21:51:15 and that is just a bug fix 21:51:35 russellb: ok 21:51:40 * russellb approves 21:51:42 Any question on Nova ? 21:52:10 #info FFE for keystone/trusts, but need to be completed this week 21:52:18 ttx, right now, it is rebased behing the v3 api fix. I'd like to say a day. 21:52:44 #topic Horizon status 21:52:52 gabrielhurley: hey 21:52:56 #link https://launchpad.net/horizon/+milestone/grizzly-3 21:52:58 hi 21:53:15 Only quantum-lbaas left 21:53:26 yep 21:53:28 :-D 21:53:44 would you require an FFe for that if it misses the deadline tomorrow ? 21:53:51 probably not 21:53:56 the code there will likely be ready in time for the deadline 21:53:57 ok 21:54:05 but it depends on Quantum finishing their haproxy plugin code 21:54:27 well, they requested an FFe for that 21:54:37 so it's a bit unlikely to make it 21:54:46 ttx: hey... 21:55:17 yeah, I'm not really sure what to do about it 21:55:34 whether it's better to make the Horizon BP an FFE contingent on the Quantum FFE, or what... 21:55:37 gabrielhurley: if its not ready, we can live with cli only for lbaas 21:55:49 danwent: +1 21:56:18 gabrielhurley: Still have 8 bugs targeted to grizzly-3. Can I move to RC1 any that doesn't get merged by tomorrow ? Or is there anything in particular you'd rather see fixed in grizzly-3 ? 21:56:25 I'm not gonna merge it until I get the OK from danwent so let's discuss it again when they get their FFE request in 21:56:37 ttx: rc1 is fine for bugs 21:56:45 gabrielhurley: ok, i will keep you in the loop. thursday is our next touch point on lbaas. 21:57:05 gabrielhurley: means you need an ffe for quantum-lbaas 21:57:14 apparently so 21:57:14 (it won't land in g3) 21:57:20 sorry, I'm multi-tasking horribly right now 21:57:23 #info FFe for quantum-lbaas 21:57:35 Anything more on Horizon ? 21:57:53 nope. got a ton of great stuff merged lately, so that's awesome. 21:58:20 #topic Incubated projects 21:58:34 Anyone from heat/ceilometer ? 21:58:37 hi 21:58:40 #link https://launchpad.net/heat/+milestone/grizzly-3 21:58:48 all green 21:58:51 cool 21:59:12 ~10 bugs in rc1 21:59:17 need to figure out if the tarball generation is up to par 21:59:28 last time I checked it was... different 21:59:30 i merged your patch this morning 21:59:33 ok 21:59:36 and merged a version patch 21:59:40 so hopefully that fixes it 21:59:44 yeah, might have fixed it 21:59:51 anyone from ceilo ? 21:59:55 #link https://launchpad.net/ceilometer/+milestone/grizzly-3 22:00:18 circa 4 bugs that will have to retarget'd at RC1 I think 22:00:31 eglynn_: What should I do with anything that is not completed by tomorrow ? 22:00:31 one BP also in danger of not making it 22:01:02