07:00:26 #startmeeting tripleo 07:00:27 Meeting started Wed May 28 07:00:26 2014 UTC and is due to finish in 60 minutes. The chair is tchaypo. Information about MeetBot at http://wiki.debian.org/MeetBot. 07:00:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 07:00:30 The meeting name has been set to 'tripleo' 07:00:39 o/ 07:00:41 hello 07:01:00 Ooh, we have people here already, that's exciting 07:01:06 :-) 07:01:14 O/ 07:01:22 not many it seems :-) 07:02:17 yeah. not really crowded 07:02:31 true. But it's the first one at this time, and it's my first time driving meetbot, so it's very exciting anyway 07:02:33 for me at last 07:03:07 #topic agenda 07:03:16 bugs 07:03:19 reviews 07:03:21 Projects needing releases 07:03:23 CD Cloud status 07:03:25 CI 07:03:27 Insert one-off agenda items here 07:03:29 open discussion 07:03:38 Anything else? 07:03:57 We had a few action items last week - http://eavesdrop.openstack.org/meetings/tripleo/2014/tripleo.2014-05-20-19.01.txt 07:04:13 but I'm guessing bnemec and rpodolyaka aren't here, so we can carry those over to next week. 07:04:21 tchaypo: I'm here 07:04:25 The third item was to have a meeting at this time and I seem to have managed that 07:04:33 rpodolyaka: excellent :) 07:04:45 No other items? 07:05:08 #topic bugs 07:05:12 https://bugs.launchpad.net/tripleo/ 07:05:15 https://bugs.launchpad.net/diskimage-builder/ 07:05:17 https://bugs.launchpad.net/os-refresh-config 07:05:19 https://bugs.launchpad.net/os-apply-config 07:05:21 https://bugs.launchpad.net/os-cloud-config 07:05:23 https://bugs.launchpad.net/os-collect-config 07:05:25 https://bugs.launchpad.net/tuskar 07:05:27 https://bugs.launchpad.net/python-tuskarclient 07:05:37 oh wow, serious red for tripleo bugs :( 07:06:10 o/ 07:06:14 o/ 07:06:15 baby in bed 07:06:17 it's been a complicated week :) 07:06:27 Yeah, 5 new unassigned undecided 07:07:05 one in os-cloud-config - StevenK could you jump on that one? 07:08:15 it looks like tripleo is the one project that needs some serious attention though. 07:08:24 closed https://bugs.launchpad.net/tripleo/+bug/1319978 07:09:22 StevenK and I are planning to spend friday in the office together; it would be good to clear them up before then, but we can plan to bash on them together then if they're still open 07:10:42 #info tripleo project has 12 new/incomplete bugs, many of them unassigned. This needs some attention 07:11:05 Anything else to add before we move on? 07:11:48 #topic reviews 07:12:44 http://russellbryant.net/openstack-stats/tripleo-openreviews.html still seems to be blank 07:12:51 I think this had a vote last meeting? 07:12:52 bnemec has a url 07:13:07 www.nemebean.com/reviewstats/tripleo-open.html 07:13:23 12 days \o/ 07:13:36 I think meetbot grabs anything with a url and I don't need to mention those with #link 07:13:37 we had a goal last week to bring it down 07:13:40 so I'm just leaving that be 07:13:47 Notes are at http://eavesdrop.openstack.org/meetings/tripleo/2014/tripleo.2014-05-20-19.01.txt 07:14:09 The group goal was to hit 12 days 07:14:22 which we have \o/ 07:14:54 I must be dense, I'm not seeing where to read that 07:15:11 o/ 07:15:24 tchaypo: 07:15:24 tchaypo: since the last revision without -1 or -2, 3rd quartile 07:15:25 Stats since the last revision without -1 or -2 : 07:15:25 Average wait time: 8 days, 17 hours, 38 minutes 07:15:25 1rd quartile wait time: 2 days, 3 hours, 11 minutes 07:15:25 Median wait time: 6 days, 15 hours, 39 minutes 07:15:27 3rd quartile wait time: 12 days, 2 hours, 48 minutes 07:15:29 lo pblaho 07:15:45 oh, of course, that's "since-negative-review" 07:16:01 I was looking for the words "since negative review" #tooliteral 07:16:42 I'm not going to propose a new vote - I think we have too few people, and my understanding is that the meetings aren't the best place for a decision anyway 07:16:44 but 07:17:09 #info Target from last meeting - 3rd quartile wait time since last negative review - has been achieved 07:17:21 #info Let's aim for 11 days next week! 07:17:35 o/ woops. you'd think i'd remember the first meeting at an actually convenient time 07:18:31 I planned to be online 12 hours ago just in case there was confusion, but I managed to sleep in for a change. it was lovely :) 07:18:43 tchaypo: lol :) 07:18:43 Okay, unless we have anything further to say about reviews... 07:19:14 #topic Projects needing releases 07:19:30 not a problem, I can volunteer this week too 07:19:40 tchaypo: 12 hours ago, there was just me and jrist :-D 07:19:50 #info rpodolyaka to save our bacon once again 07:20:00 Actually 07:20:06 rpodolyaka: you got a link to those notes you made for the release process? seems like if you or slagle are away would be good to have backup 07:20:16 #action see above 07:20:17 rpodolyaka: does it take a lot of time in your week? 07:20:26 marios: 1-2 hours 07:20:51 marios: mostly reading the changes, deciding what part of the version to bump 07:21:05 marios: https://wiki.openstack.org/wiki/TripleO/ReleaseManagement 07:21:05 rpodolyaka: i think you documented it somewhere no? 07:21:09 rpodolyaka: great thanks mate 07:21:13 marios: np 07:22:07 So... 07:22:10 coing up next... 07:22:19 #topic CD Cloud status 07:22:28 lifeless: I hear tell this hasn't been swell 07:25:05 its been rocky 07:25:12 so there is an etherpad 07:25:20 https://etherpad.openstack.org/p/tripleo-ci-r1-trusty 07:25:32 which is about getting hp1 entirely onto trusty 07:25:39 without that we should expect more issues still 07:26:55 we've also hit the stuck session in nodepool again 07:27:04 lifeless: so R1 isn't currently being used by nodepool, 07:27:05 filed a bug for that on novaclient and made a (little) fuss about it 07:27:11 derekh_: rh1 ? 07:27:20 lifeless: should we just leave it that way until we get to trust ? 07:27:45 derekh_: IMO yes 07:27:51 lifeless: R2 is running ok, but struggle to keep up with the full load in the absense of R1 07:28:16 lifeless: # worker nodes was finially increased from 10 to 18 last night 07:28:25 derekh_: nodepool has ~ 60 active slaves in hp1 07:28:35 derekh_: why do you say its not being used ? 07:28:47 ah, they are mostly missing public ips 07:28:51 lets see if that helps and if it doesn't maybe we should get rid of an overcloud job for a bit 07:29:00 lifeless: I don' think they're reaching READY 07:29:12 lifeless: when I boot an instance, it never fully establishes an SSH connection 07:29:26 SpamapS: I tested instances with ssh just fine 07:29:29 lifeless: nodepool most be spinning thme up without deciding their usable 07:29:30 interesting 07:29:37 perhaps it was just one bad compute node? 07:29:42 lifeless: none are getting used goodsquishy.com/downloads/s_tripleo-jobs.html 07:29:53 lifeless: did you actually SSH in, or just telnet to port 22? 07:30:11 SpamapS: ssh in 07:30:13 SpamapS: sudo'd 07:30:13 lifeless: I got fails after connection, traffic would stop flowing 07:30:16 SpamapS: apt-get updated 07:31:04 lifeless: perhapes those 19 jobs in the ? region were on R1, ? means they failed befor the log entry allowing me to decipher the region 07:31:08 lets -> #tripleo for details 07:31:21 its clearly bong, since only one node has a public IP 07:31:28 * derekh_ nods 07:31:30 we'll likely need infra logs to see whats up 07:31:40 public IP is pretty much the first thing that nodepool should be doing 07:32:14 Do we have a specific action item to record here? 07:32:56 no 07:32:58 tchaypo: action have a side meeting about R1 in #tripleo after main meeting 07:33:03 or yues :) 07:33:41 #action derekh_ SpamapS lifeless to discuss CD in R1 further 07:33:54 #info it's been a rocky week 07:34:06 #topic CI 07:34:22 I feel like this isn't going to be much different 07:35:05 derekh_: there was an action item from last week for bnemec to follow up with you about ci improvements and spec - did that happen? 07:35:17 tchaypo: spec for ci improvements is up, finding link 07:35:35 https://review.openstack.org/#/c/95026/ 07:35:56 reviews welcome :-) 07:36:06 #info As per action item from last week, https://review.openstack.org/#/c/95026/ is a spec for ci improvements 07:36:43 #info derekh_ welcomes reviews :) 07:36:46 anything else? 07:37:59 #topic Insert one-off agenda items here 07:38:34 #info all the action items from last week have already been covered #yay 07:38:37 Oh I had something I wanted to mention 07:38:49 The floor is yours :) 07:38:52 gerrit no longer does auto abandon of patches, 07:38:59 bahahahaha 07:39:02 that's quite a change 07:39:16 could people look through old patchs and abandon things that are irrelevant 07:39:24 Could you say that again with an hashtag-info ? 07:39:27 derekh_: didn't know that. it's both a good and a bad thing though. 07:39:37 #info could people look through old patchs and abandon things that are irrelevant 07:39:49 #info as Gerrit doesn't auto-abandon any more 07:40:08 #topic open discussion 07:40:09 tchaypo: are we going to try and match (upcoming) specs to juno release schedule? do we want to think (or already have) a date for specs to be in by? 07:40:33 marios: I feel like that's a question we should ask elsewhere 07:40:39 marios: ya its a little good but I think its mostly bad as stale patches will end up staying around and taking up peoples time 07:40:54 tchaypo: yeah, mostly rhetorical. don't expect full discussion/decisions here 07:40:55 so core can mark patches as WIP 07:41:00 either on the mailing list, or simply by proposing a spec that says we match the specs to the release schedule and see how that goes 07:41:01 but I don't think thats sufficient 07:41:10 but we can review with WIP hidden 07:41:16 as an interim measure 07:41:43 lifeless: just bringing it up as it's something i saw in neutron 07:41:54 lifeless: agreed WIP isn't sufficient, if we start reviewing with WIP's hidden then whats the point of people using WIP at all for get feedback 07:41:57 lifeless: juno-1 has been made up with all expected bugs/reviews etc 07:42:06 We seem to be near the end of my first meeting, and lifeless hasn't felt the need to use the second set of pedals yet. Iguess that means I can graduate to p-plates next time? 07:42:09 derekh_: WIP is for 'don't review yet' :) 07:42:11 lifeless: any specs from now on will be taken per case etc 07:42:12 derekh_: explicitly so 07:42:27 marios: sorry, I don't quite follow the question 07:42:43 lifeless: are we going to try and match (upcoming) specs to juno release schedule? 07:42:51 do we want to think (or already have) a date for specs to be in by? 07:42:53 marios: what does that mean ? 07:42:58 haha 07:43:03 lifeless: we have specs 07:43:11 lifeless: (either approved or under discussion) 07:43:17 lifeless: we have a juno release schedule 07:43:51 lifeless: do we want to make a list of when we expect certain specs to have been implemented, bugs closed etc, for juno release schdule 07:44:11 lifeless: I though WIP was more of a "here is the direction I'm going in if ya wanna take a look but don't review in detail yet cause I not finished" button 07:44:23 So we have a juno folder in the specs repo; I had been assuming that anything in there was targetted at Juno release, but we can move things if they slip 07:45:16 tchaypo: sure, i meant more like 'juno-1' we try and concentrate on these things, if they land ok juno-2 will be this and this etc etc 07:45:56 marios: so, I generally think that organising every little thing like that is a boiling-the-ocean problem 07:45:59 Do we know when the mid cycle sprint is ? 07:46:01 e.g. like this https://launchpad.net/neutron/+milestone/juno-1 07:46:05 Mayhap it would be useful to make juno/1, juno/2 and so on? 07:46:07 derekh_: I'm waiting on jcoufal for that 07:46:19 To me it sounds like it would realy be useful to have a spec describin how the spec repo operates 07:46:24 derekh_: though if he doesn't post to the list soon I'm going to have to find alternate venues since folk won't have time to book 07:46:25 lifeless: ok 07:46:25 lifeless: just a question/suggestion as i hadn't come across it before and thought it was possibly a useful resource to refer to during the cycle 07:46:30 lifeless: derekh_ I will send some updates today or tomorrow the latest 07:46:30 jcoufal: hint, hint 07:46:46 jcoufal: cool, thanks 07:47:01 marios: I' 07:48:10 marios: do you know if nova have done that kind of targeting of specs to milestone releases? 07:48:17 marios: that said, part of approving a blueprint will be identifying if its current work or future work 07:48:26 marios: so we can certainly target the next bit of work 07:48:31 If they have maybe we can copy them; if they haven't, maybe there's a reason they haven't. 07:48:35 lifeless: sure 07:48:40 I'm very much against having large numbers of idle unimplemented specs 07:48:49 neutron has quite a different situation to us 07:49:05 lots of vendors with very specific axes to grind 07:49:10 right 07:49:22 and a relative paucity of work-in-the-core feeling 07:49:27 they're trying to change that 07:49:56 (indeed, the axes are becoming much less specific from what i can see) 07:51:30 #action jcoufal to send an update on mid-cycle dates today or tomorrow 07:53:15 lifeless: so if I can summarise, it sounds like you don't want us to have idle specs; that's going to need scheduling decisions when specs are approved, but will probably need review after that as well. Do you want to own deciding on how we do that review and how specific we want to be with targetting work (to milestones, or just to releases, or somethign else)? 07:53:41 tchaypo: its on me for now I think 07:53:45 We still have 6 minutes left, is there anything else anyone wants to bring up? 07:54:20 #action lifeless to follow up on process for review and (re-)scheduling/targetting of specs 07:55:41 tchaypo: actually I was proposing to do nothing 07:55:49 lol 07:55:50 tchaypo: #undo will remove the last #command :) 07:55:55 #undo 07:55:56 Removing item from minutes: 07:56:00 haha 07:56:11 $action lifeless to sit on his hands 07:56:16 ;) 07:56:33 If there's nothing else I'm going to call time 07:56:53 thanks to those that pushed for the alternate time 07:57:05 its great to be making coffee vs going to bed 07:57:17 Heh 07:57:29 #agreed that those of us present love the alternate meeting time, those not present probably don't love it as much 07:57:41 hehe 07:57:46 heh 07:57:47 #endmeeting