15:00:01 #startmeeting Ceilometer 15:00:01 #meetingtopic Ceilometer 15:00:01 #chair nijaba 15:00:01 #link http://wiki.openstack.org/Meetings/MeteringAgenda 15:00:02 Meeting started Thu Oct 4 15:00:01 2012 UTC. The chair is nijaba. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:04 The meeting name has been set to 'ceilometer' 15:00:06 Current chairs: nijaba 15:00:12 Hello everyone! Show of hands, who is around for the ceilometer meeting? 15:00:12 o/ 15:00:14 o/ 15:00:17 o/ 15:00:32 o/ 15:00:38 o/ 15:00:51 nice! 15:00:56 #topic actions from previous meeting 15:00:56 #topic nijaba to share a first version of the slides with dhellmann 15:01:07 We've been working well together and I think we almost have a final version 15:01:21 about 20 slides for 40 minutes 15:01:29 cool 15:01:31 o/ 15:01:32 volunteers to review? 15:02:02 o/ 15:02:12 please pm me your google enabled address if you want to give us your feedback 15:02:26 #topic dhellmann to provide some bullet points on DH use case 15:02:39 that was done too as part of the slides 15:02:49 o/ 15:02:54 #topic dhellmann update status on http://wiki.openstack.org/EfficientMetering/RoadMap 15:03:02 dhellmann: ? 15:03:12 I did update it, but it is not entirely up to date 15:03:22 lo 15:03:28 I completed a review of all of the meters we're collecting earlier today, so I have enough info to finish that now 15:04:04 nice! 15:04:18 #topic gmb to email the list with a summary of how things stand for the release 15:04:36 gmb is busy, but he will be sending the email before the eod 15:05:15 #topic Release status 15:05:33 so, while gmb is out, I think we need to drive this forward 15:05:44 I think we are past our feature freeze date 15:05:50 but are still accepting features 15:05:59 so when should we be cutting off? 15:06:01 * dhellmann hangs head in shame 15:06:21 * jd__ votes $RANDOM 15:06:43 what feature are on the verge of delivery that we would like to have? 15:07:00 there are 2 things I'd like to include 15:07:12 1. option to disable meters through configuration 15:07:23 I think it would be nice to have release on the 14th latest, so we can say it is done at the summit 15:07:25 2. review and standardize the meter names 15:07:41 #1 isn't a requirement for a 0.1 release, but would be nice 15:07:45 dhellmann: do you think this can be completed before the end of this week? 15:07:55 #2 will be more difficult to change later, I think, after people start depending on the current names 15:08:04 so that next week can be dedicated to test and release? 15:08:26 I think so. jd__ what do you think? 15:08:36 hum 15:09:02 * dhellmann just noticed a bunch of related code reviews from jd__ 15:09:02 i think both requirement makes sense 15:09:12 I've started #2 a bit actually as you just noticed 15:09:23 I'll take #1 15:09:31 I know how I want that done, and Justin started the code already 15:09:32 we have a start for #1 I think, right? 15:09:38 ok :) 15:09:41 that's what I though 15:09:48 +t 15:10:03 I think it'll be enough for a first release 15:10:16 so, feature free on saturday? 15:10:21 we can call this "technical preview" or something like that 15:10:30 i'm sure there's a good marketing name :) 15:10:32 let's say sunday since I may not actually get to work on it until then 15:10:33 0.1 carrues that very well, I think :) 15:10:44 nijaba: indeed :) 15:10:48 0.1 should be marked an alpha, right? 15:10:58 #agreed feature freeze push to sunday 15:11:00 one day I'll write a book on software version numbers 15:11:22 ;) 15:11:27 :-) 15:11:29 dhellmann: call it what you want, I know crazy guys will be usingit in prod regardless ;) 15:11:44 heh 15:12:03 nijaba: yeah i was wondering is anyone running this ina prod env? 15:12:17 ok, then what do we need to do then to have a release? 15:12:20 cp16net: we will be soon at DreamHost 15:12:20 or just dev or local env 15:12:28 but so far only dev 15:12:32 ok cool 15:13:53 we need to open a new dev branch and gate the existing one to bug fix only. Who can do that on monday 15:14:04 I have no idea how to do this with git 15:14:20 volunteers? 15:14:35 I can try 15:14:55 do we need to have special access on github to do that? 15:14:58 OTOH I think we need to adapt to something like other projects, we continue to commit in master but create stable branch 15:15:12 +1 to stable branch 15:15:15 jd__: ok that works too 15:15:20 spn: not sure, I think i'll poke the -infra guys 15:15:27 also a milestone-proposed branch? 15:15:27 the guys in openstack-infra should be able to explain how 15:15:39 #action jd__ to create a stable branch on monday 15:15:39 cool 15:15:42 eglynn: we would need to define some milestones first, no? :-) 15:16:04 I think just a folsom/stable branch would be enough for now 15:16:16 0/ 15:16:18 it's not like we are expected to release real working stuff 15:16:20 :-> 15:16:21 dhellmann: the milestone-proposed branch is usually used for pre-release fixes 15:16:23 o/ 15:16:27 nijaba: we should, after (or at?) the summit put together a set of milestones for grizzly 15:16:33 dhellmann: +1 15:16:40 roooaaarr (grizzly) 15:16:41 and tag master as well 15:16:42 also shouldn't the stable/folsom only be cut once the release is done? 15:16:56 eglynn: cutting after the release makes sense 15:17:05 eglynn: we release on sunday and makes the branch on monday? 15:17:26 next week is QA week, we release the weekend after 15:17:35 nijaba: that makes sense 15:17:46 jd__ we feature freeze on Sunday, but release later in the week? 15:17:55 so we should work on bug fixes on the stabel branch, new features land on master 15:17:56 so fixes next week go on master first, then cherry picked onto miletsone-proposed 15:18:00 eglynn: the milestone-proposed branch may make more sense after we have our first release then? I'm not sure what it buys us now 15:18:09 ah 15:18:31 milestone-proposed for pre-relase fixes, stable/folsom for post-relase fixes 15:18:42 s/relase/release/ 15:18:42 I see 15:18:46 eglynn: ah, that makes sense 15:19:10 * nijaba wonder if ttx is lurking 15:19:20 FYI some more detail on the process: http://wiki.openstack.org/GerritJenkinsGithub#Authoring_Changes_for_milestone-proposed 15:19:42 thanks eglynn 15:19:46 http://wiki.openstack.org/BranchModel worth reading too 15:19:51 nijaba: no 15:20:24 need help ? 15:20:42 ttx: if you have best practices for handling our release 15:20:45 we're discussing how to copy branching model for stable and milestone from other project 15:20:48 right now we are on: 15:20:51 ttx: just clarifying the use fo miletsone-proposed versus stable branch 15:20:54 milestone-proposed for pre-relase fixes on monday 15:21:09 agreed 15:21:19 stable/folsom for post-relase fixes next weekend when we release 15:21:32 we tag master 15:21:37 each time 15:21:56 ttx: makes sense to you? 15:21:57 hmm, how is the version generated right now ? 15:22:07 what tarballs are you producing currently ? 15:22:23 * jd__ whispers 15:22:25 none I think? 15:22:27 I don't think we produce any 15:22:29 yet 15:22:31 I think we just have the version # hard-coded in setup.py 15:22:46 ok 15:22:53 let me think 15:23:13 if you don't have automatic tarball generation, that simplifies 15:23:54 you can probably create milestone-proposed on top of master, then stable/folsom on top of milestone-proposed (after release) 15:24:08 if you want to mimick how iot's done for core 15:24:11 I think this is what eglynnwas proposing 15:24:18 ttx: that's the point, thanks 15:24:21 and yes, we want 15:24:23 we'll do that :) 15:24:29 Been working on rewriting http://wiki.openstack.org/ReleaseTeam/HowToRelease 15:24:30 how do we get the branches pushed to github? 15:24:30 thanks a lot ttx 15:24:54 which is the complete process I use, but you need nothing of that complexity 15:25:50 since you have nothing like milestone codes or pre-generated tarballs 15:25:53 ok, so the next question is weather we are missing anything for a release? 15:25:57 release notes maybe? 15:26:35 also, do we need to produce a tarball? 15:26:36 we could add those to our docs 15:26:54 dhllemann: git remote add github git@github.com:user_name/my_app.git and git push -u github master 15:27:32 spn: I don't know if any of us have permission to push directly to github because of the way the repo was created. maybe nijaba has admin rights? 15:27:49 agreed. 15:27:52 #action nijaba to prime release notes in the docs 15:28:24 dhellmann: can I grant those rights to someone else? 15:28:34 may be adding other users as collaborators 15:28:38 nijaba: if you have admin rights, I think you can 15:28:48 nijaba: do you have admin? 15:29:05 * jd__ does not have admin 15:29:12 dhellmann: I have no idea. Dachary did the initial setup of the repo IIRC 15:29:12 * dhellmann does not have admin 15:29:23 how do I check? 15:29:31 admin page ( https://github.com/user/repo/admin ) and in the Collaborators tab 15:29:37 go to https://github.com/stackforge/ceilometer and look for the "admin" tab 15:29:41 see if you can add users 15:30:04 the admin tab should be on the right end after "graphs" etc. 15:30:25 anyway someone from openstack-infra has admin so don't bother I'll deal with them 15:30:32 dhellmann: no admin tab for me. I'll speak with dachary 15:30:37 ok 15:30:47 we should probably add you, me, and jd__ 15:30:50 at least 15:31:07 #action nick to find a way to give admin right to jd__ for github 15:31:09 https://github.com/stackforge?tab=members -> Monty :) 15:31:24 #action nick to find a way to give admin right to jd__ and dhellmann for github 15:31:50 #action nijaba to find a way to give admin right to jd__ and dhellmann for github 15:31:58 I think I have the action right this time! 15:32:05 we hope so :) 15:32:16 so, do we need to produce a tarball? 15:32:36 nijaba: that may be something we can have the infra team turn on for the project 15:32:41 like they do for nova, etc. 15:32:59 ok, I'll check into it at the same time then 15:33:21 #action nijaba to check if infrateam can generate a tarball for us 15:33:37 are you sure? that may overlap with my #action about putting the branches in place 15:33:49 not that it's a big problem :) 15:33:53 jd__: you want to cover it then? 15:34:03 as you wish, but I can handle all this at once I imagine 15:34:10 ok, sounds good 15:34:17 #action jd__ to check if infrateam can generate a tarball for us 15:34:39 anything else we might be missing for a release? 15:34:39 once I have kidnapped an -infra guy, I can do whatever is needed 15:35:10 i think it'll be a good start :) 15:35:10 :-) 15:35:18 * nijaba wonders if jd__ is really jack dauer 15:35:46 ok, next topic then 15:35:59 #topic Motion to review incubation request at the next TC meeting 15:36:07 We need to trigger the new TC to decide if we are worth being incubated. I think it would be nice if we ensured this would happen at their next meeting, which will be face to face in san diego. Are you ok with me sending them an email on that subject. If they give us a green light, according to the by-laws, the board of director will then have to confirm. 15:36:55 +1 15:37:05 +1, face to face is always easier to kidnap 15:37:09 dumb question: what's the timeline for incubation? 15:37:11 (what is wrong with me!) 15:37:16 eglynn: one cycle 15:37:20 (i.e. how long of a prior track-record is required?) 15:37:28 a-ha ok, cool 15:37:41 eglynn: prior track-record is unspecified, afaik 15:37:57 one cycle in incubation, at least, before being core 15:38:12 cool, thanks for the info 15:38:15 * dhellmann thought eglynn meant prior to incubation 15:38:25 dhellmann: I did :) 15:38:39 dumb q?: being in incubation and core are same? 15:38:51 spn: in terms of duties, yes 15:38:54 one preceeds the other 15:39:25 do we need to vote on my proposal or should I just mark that as agreed? 15:39:34 I sense agreement 15:39:58 a yea by acclamation I think 15:40:05 is anyone opposed? 15:40:12 no 15:40:29 #agreed nijaba to send an email to the tc to ask for a motion to put ceilometer in incubation 15:40:51 this bring us to the next topic 15:40:54 #topic ODS Grizzly 15:41:06 roooaaarr 15:41:06 show of hands, who is going to be there in person? 15:41:11 how many slots at teh ODS does ceilometer have? 15:41:23 just so that I can plan for cash to buy beers ;) 15:41:24 o/ 15:41:25 o/ 15:41:29 * eglynn just heard he'll probably be able to attend :) 15:41:36 eglynn: \o/ 15:41:37 wait, beer? o/// 15:41:54 :) 15:42:00 eglynn: :) 15:42:07 I think we have webex option this time 15:42:19 spn: yes, that was mentioned on the mailing list 15:42:20 and long nights for you then... 15:42:37 hihi.. 15:43:23 not sure whether we are allowed to interact on webex 15:43:49 spn: I'm not sure whether it's supported, but if not we could have someone in the room on irc relay questions and comments 15:44:02 no one else is going to be there? 15:44:08 o/ 15:44:16 dhellmann: that would be great 15:44:17 ah! 15:44:25 I am sure jaypipes will be too 15:44:31 what about jtran? 15:44:39 about design summit: I'll need your session descriptions for inclusion in the final schedule Monday 15:44:59 ttx: http://wiki.openstack.org/EfficientMetering/GrizzlySummit 15:45:05 so here you are 15:45:05 * jaypipes will be there Monday around noon 15:45:12 nijaba: i'll try to remember that 15:45:21 ttx: do you want an email? 15:45:27 nijaba: nah 15:45:30 k 15:45:40 jaypipes: will jtran be there? 15:46:15 ttx, do you have other questions about the metering integration sessions? 15:46:22 ttx: you wanted to talk about "the multiplication of metring integration sessions" 15:46:49 nijaba: yes 15:47:04 cool. and I know enovance is sending a few people too 15:47:24 nijaba: yes, Would like to reduce the duplication asap 15:47:35 nijaba: so that you don't end up with all of them refused 15:47:42 I think we all agreed to merge 3 of them into one last night by email 15:47:50 independently by the separate topic leads making independent decisions 15:47:52 115,116 and 118 IIRC 15:49:15 115,117 and 118 actually 15:49:26 yes, the latter 15:49:29 ttx: would this be good enough for you? 15:50:48 let me check which topic they were proposed under 15:50:50 I just talked to Mark about the quantum session (110) and he wants to keep that focused on quantum because it will cover more than ceilometer 15:51:09 i will be there. 15:51:40 so, I hope you have all noted that ceilometer 3 sessions will happen on moday morning 15:51:47 OK, let's merge 117-118 15:51:59 you can keep 115 since Horizon is missing stuff anyway 15:52:20 and it wil give you a late session (Thursday afternoon) on the subject to wrap it up 15:52:51 ttx: sounds good 15:52:54 dhellmann: can you make one out of two ? 117-118 ? 15:53:01 dhellmann: and let me know which one to remove 15:53:12 will be simpler than communicating the merge to markmc, isn't around 15:53:17 ttx: yes, I'll do that after lunch and email you the results 15:53:32 (in a couple of hours) 15:53:37 ok, just let me know which to remove. 15:54:29 ok, let's move to our final topic then 15:54:37 #topic open discussion 15:54:49 ttx, delete 118 15:54:56 on it 15:55:04 5 minutes, any topics... 15:55:08 http://wiki.openstack.org/Governance/Proposed/Ceilometer 15:55:19 This page has incubator details 15:55:22 at the bottom 15:55:25 spn: yes, that was our proposal 3 mo ago 15:55:29 we have integration status 15:55:43 I guess it is 3mon old status then 15:55:51 dhellmann: nuked. 15:55:52 I'll need to refresh that before sending my email 15:55:59 ttx: thanks! 15:56:59 #action nijaba to resfresh http://wiki.openstack.org/Governance/Proposed/Ceilometer before sending email to the tc 15:57:17 anything else? 15:57:29 * dhellmann has nothing to add 15:57:35 nothing from me 15:57:54 that was a full meeting! thanks a lot everyone! 15:58:04 thanks! 15:58:07 next week, same place, same time 15:58:09 thanks 15:58:16 #endmeeting