15:04:48 #startmeeting Ceilometer 15:04:48 #meetingtopic Ceilometer 15:04:48 #chair nijaba 15:04:48 #link http://wiki.openstack.org/Meetings/MeteringAgenda 15:04:49 Meeting started Thu Oct 11 15:04:48 2012 UTC. The chair is nijaba. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:04:50 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:04:51 The meeting name has been set to 'ceilometer' 15:04:53 thanks 15:04:54 Current chairs: nijaba 15:05:01 Hello everyone! Show of hands, who is around for the ceilometer meeting? 15:05:01 o/ 15:05:03 o/ 15:05:07 o/ 15:05:08 o/ 15:05:10 thanks jeblair 15:05:10 o/ 15:05:12 o/ 15:05:14 o/ 15:05:31 #topic actions from previous meeting 15:05:43 #topic jd to create a stable branch on monday 15:06:19 so I think this was consider too much of a pai by the infra team, right jd__? 15:06:26 yep 15:06:37 pai? 15:06:42 pain 15:06:42 pain 15:06:48 so we will use master as our stable branch? 15:06:53 and we can't ask all of that because we're not official nor incubated 15:07:08 * dhellmann wonders how hard it is to make branches 15:07:08 actually, miletsone-proposed was refused 15:07:09 well 15:07:14 ah 15:07:19 not doing a folsom branch 15:07:27 so i think we wanted to set up permissions so you can make your own branches.... 15:07:36 but we haven't gotten around to that yet 15:07:38 jeblair: that would be a good compromise 15:07:39 yes,that would be nice 15:07:42 we'll be able to create a folsom branch just by asking infra to create it with the commit id 15:07:58 so in the mean time, we should just create the branch for you instead of giving you a hard time. :) 15:07:59 (well unless jeblair give us permissions) 15:08:08 +s 15:08:31 jeblair: can I mark an action for you? 15:08:47 (sorry to have pulled you in a trap) 15:08:53 nijaba: sure. :) 15:09:00 do we want whatever HEAD is now, or do we need to think about which revision should be used? 15:09:15 #action jeblair to give us rights t create our own branches 15:09:37 dhellmann: let's see that under the release status topic 15:09:45 nijaba: ok 15:09:47 so in the end, it was decided not to go ahead with the milestone-proposed style branching for the run-up to the release? 15:10:02 eglynn: yep, we did not have much coice 15:10:11 nijaba: fair enough 15:10:41 #topic nijaba to prime release notes in the docs 15:10:52 his was done and is now visible at 15:10:52 #link http://ceilometer.readthedocs.org/en/latest/releasenotes/index.html 15:10:52 Please send patches if we need to add to this 15:11:20 doc/sources/releasenotes/ in the code 15:11:35 any comments or suggestions? 15:12:06 thanks for doing so much work on the docs this week, nijaba, they look really good 15:12:19 you are welcome dhellmann 15:12:45 I'll skip the next action as it was already discussed as part of the first acton 15:12:57 #topic jd to check if infrateam can generate a tarball for us 15:13:19 jd__: what did infra say on this? 15:13:44 would that be different than the tarballs github produces in some way? 15:13:53 didn't ask, but we can have some on github I think 15:14:20 so that's link to us being admin of our projec on github, I guess? 15:14:33 somehow yes, we need to be able to tag our project 15:14:55 I wonder if we can just push a tag to that repo through gerrit? 15:15:19 ok, we already have an action on this, but will need to chase this quickly to be able to release 15:15:37 i think globally that infra doesn't know more than us how to handle our project and our release, so once we want to release we'll have to go ask for things *now* and they'll have to find a way :) 15:16:03 jd__: maybe if we just focus on getting the permissions, we can experiment 15:16:12 dhellmann: +1 15:16:35 ah sure, but I didn't feel they wanted to give us any permission 15:16:40 but I might be wrong :) 15:16:53 jeblair: any comments on this? 15:16:58 now that we are asking for so many thing they might change their mind :-D 15:17:24 jd__: I think it was only a matter of not having time to help us 15:17:44 hi 15:18:01 * nijaba propose that we go in #infra just after the meeting to get to the bottom of this 15:18:19 #vote yes 15:18:23 so i think the way we want to handle this is for one or more of you guys brush up on how jenkins jobs are built 15:18:36 and then you can add it to the infra yourself 15:18:49 * dhellmann is confused 15:18:50 that's agood idea. 15:18:53 all the jobs, etc, are in a git repo anyone can submit changes to 15:19:00 what does jenkins have to do with branching our repo? 15:19:11 dhellmann: this was tarballs, yeah? 15:19:20 oh, I see 15:19:44 nijaba: is the tarball a release requirement? 15:19:59 dhellmann: I would think so... 15:20:09 dhellmann: what would distro use otherwise? brnahc? 15:20:15 branch even... 15:20:26 nijaba: could not just point them at the tarball github produces? 15:20:52 ^could not^could we not 15:20:52 dhellmann: Error: "could" is not a valid command. 15:20:55 dhellmann: if is clearly identifiable and durable, yes 15:20:58 dhellmann: no 15:21:12 the tarball github produces is not a good tarball 15:21:23 (you guys are python, right?) 15:21:29 mordred: yes 15:21:30 we are 15:21:50 sadly, the github tarballs are not python sdist tarballs 15:22:10 so they don't have metadata that's needed for the tarballs for fully-well interact with stuff like pip 15:22:13 ok, so we have 2 things we have to do for the release 15:22:19 1. make the stable branch 15:22:20 2 15:22:20 15:22:21 s 15:22:22  15:22:27 2. set up the job to make the sdist tarball 15:22:37 ++ 15:22:53 ok, let's park this for now. dhellmann, jd__ are you ok to followup on this in -infra right after the meeting? 15:22:55 3. be able to push to the stable branch 15:22:59 nijaba: yes 15:23:02 jd__: +1 15:23:02 nijaba: I'm in 15:23:27 #action dhellmann, jd__, nijaba to follow on infra fr our release reqs 15:23:41 #topic nijaba to resfresh http://wiki.openstack.org/Governance/Proposed/Ceilometer before sending email to the tc 15:23:50 That was done and the email to the tc was sent. Looking at the tc ml archive, it seems that the vote on incubation will not happen during the face to face meeting for lack of logging and supposed openess of the process. 15:24:36 do we know when they are holding their next online meeting? 15:24:58 archives: http://lists.openstack.org/pipermail/openstack-tc/2012-October/thread.html 15:25:18 http://wiki.openstack.org/Governance/TechnicalCommittee 15:25:35 it lists the in person meeting as the next meeting, next is tbd 15:25:47 can anyone subscribe to that list? 15:26:13 "First regular TC meeting will happen the week after the Design Summit, time tbd at the inaugural meeting. " 15:26:30 o/ 15:26:41 dhellmann: I haven't tried, I just look at the achives now and then 15:26:53 ok 15:27:07 I guess that's it for this week's actions... 15:27:32 #topic Release status 15:27:44 gmb, the floor is yours 15:29:09 gmb? 15:29:31 Sorry, was getting pinged in another channel 15:29:32 So 15:30:00 All the items on the RoadMap that are required for Folsom.0 / 0.1 are committed (thanks jd__ for taking care of that last bug). We're good to go. 15:30:25 I would have targeted all the bugs to a 0.1 milestone 15:30:32 As I said in my email last week 15:30:39 anything blatently missing (we'll get to bugs in a minute)? 15:30:43 But Launchpad is being uncooperatve. 15:30:49 nijaba: Not AFAICT. 15:31:00 anyone else? 15:31:24 nop 15:31:32 do we need to add a trove classifier to our setup.py indicating this is an alpha? 15:31:55 I don't know if any packaging tools other than pypi pay attention to that field 15:32:04 dhellmann: I do not think we need that 15:32:11 ok, then I think we're ready 15:32:39 it looks like I7837e80940e7ee9622bdcd3b1f253d582f383d1a (current HEAD on github) would be a good place to branch for the release 15:33:01 oops, that's the change id not the commit id, I meant : f2aee4b75d07a5f4d8ecf381934452e2a8dc9a3d 15:33:16 dhellmann: almost. as I am going to propose one last change before release 15:33:26 lol 15:33:28 but that in 2 topics from now 15:33:44 ok 15:33:49 suspense 15:33:59 * dhellmann moves to the edge of his seat 15:34:04 hehe 15:34:19 #topic Outstanding bugs 15:34:39 so, do we have any outstanding bugs that we need to fix before release? 15:34:55 I had a look and I can't find any... what do you guys think? 15:35:23 most of the open tickets are either wishlist, large, or small enough not to matter 15:35:38 I feel good about cutting things off where we are 15:35:42 +1 15:35:57 +1 15:36:18 anyone against (or shutup until next release) ;) 15:36:18 * eglynn is sad that https://review.openstack.org/14175 didn't make it ... 15:36:32 other than that, +1 15:36:50 +1 15:36:51 +1 15:36:58 eglynn: I'll approve that right after the release is branched 15:37:00 +1 15:37:07 eglynn: I think that landed a bit too late in our "feature freeze" 15:37:08 dhellmann: cool 15:37:16 nijaba: understood 15:37:39 ok, so we are good to go, but.... 15:37:42 #topic Naming convention for Meters 15:37:57 root_disk_size 15:37:57 ephemeral_disk_size 15:37:57 floating_ip 15:37:57 volume_size 15:37:57 seem like the remaining counters that we should rename 15:38:14 they just stic out when looking at http://ceilometer.readthedocs.org/en/latest/measurements.html 15:38:22 jd__, we talked about the disk names didn't we? 15:38:28 What I would propose is that we always use dot as separators and go from least to most discriminent. Ie: 15:38:28 disk.root.size 15:38:28 disk.ephemeral.size 15:38:28 etc... 15:38:38 yes 15:38:42 and I would propose to do that before release 15:38:47 I don't know about floating.ip though, maybe just floatingip? 15:38:47 there's also "instance:flavor" 15:39:00 yes, floatingip 15:39:09 instance flavors have dots, so maybe instance.flavor.m1.tiny etc? 15:39:10 since it is a pain to change for people having it prod 15:39:14 one issue with the dotted notation is how it gets embedded in a REST API 15:39:14 +1 to pre-release counter naming changes, much messier to do post-release 15:39:25 * nijaba offers to propose the pacth 15:39:29 dhellmann: yeah why not 15:39:29 +1 15:39:41 nijaba: thanks :) 15:39:48 DanD: think about a particular issie? 15:39:51 s/i/u/ 15:39:54 nijaba: can we keep the instance:flavor meter for now and add the new one? I already have code in production that depends on that 15:40:18 nijaba: we can leave it out of the documentation, and I will submit a patch to clean it up in a week or two 15:40:19 yes, I think this is a very particular case 15:40:52 I would be personally in favor to say that any variable name should be using a : before the var starts 15:41:17 well, that's an idea 15:41:20 and therefore keep it that way 15:41:32 * nijaba makes up grammar on the fly 15:41:40 ok, I meant to use both meters and keep the old version just for backwards compatibility 15:42:13 but using a special indicator that part of the name is a variable makes sense, too 15:42:21 yeah not a bad idea 15:42:50 so the proposed rule is "always use dot as separators and go from least to most discriminent. when a part of the name is a variable, it shold always be at the end and starts with a ':'" 15:42:57 woud you like to vote on this? 15:43:14 sure 15:43:42 #startvote agree on meter convention? yes, no 15:43:43 Begin voting on: agree on meter convention? Valid vote options are yes, no. 15:43:44 Vote using '#vote OPTION'. Only your last vote counts. 15:43:49 #vote yes 15:43:50 yes 15:43:50 #vote yes 15:43:51 #vote yes 15:44:01 #vote yes 15:44:04 #vote yes 15:44:10 #vote yes 15:44:15 #vote yes 15:44:34 #endvote 15:44:35 Voted on "agree on meter convention?" Results are 15:44:36 yes (7): jd__, nijaba, eglynn, dhellmann, gmb, spn__, DanD 15:45:01 ok, great. Patch will be done tonight to fix existing meters 15:45:21 #action nijaba to patch existing meter name to follow rule 15:45:38 #action nijaba to put the rule in the doc too 15:45:47 * dhellmann was just about to suggest that 15:46:08 ok, so I guess that once this has landed, we should be able to release tomorrow 15:46:19 \o/ 15:46:24 (with some illustrative examples in docco, always good ...) 15:46:35 should we have a release meeting tomorrow at 3M UTC to finalize? 15:46:42 in our regular chan 15:46:53 sounds good 15:47:00 we might want to merge https://review.openstack.org/#/c/14338/ too 15:47:06 * jd__ sending last-minute-patches 15:47:19 * nijaba discovers it 15:47:59 that's the tool for putting data in the db to test the api, right? 15:47:59 3PM UTC is bad for me tomorrow 15:48:03 dhellmann: yes 15:48:11 jd__: later? 15:48:46 or earlier? 15:48:47 jd__: looks pretty acceptable for me. 15:49:09 I don't think I can after 14:00 UTC, so do it without me 15:49:23 that tool wasn't really meant for end users, but we can commit it if the group feels it is important 15:49:51 dhellmann: i think it's good to have it working so people can try with fake data :) 15:49:59 dhellmann: for the same reasons I can't think why not to merge it 15:50:04 jd__: that makes sense 15:50:05 n 15:50:05 i 15:50:06 j 15:50:11 * dhellmann needs a new irc client 15:50:22 nijaba: just being strict :-) 15:50:26 hehe 15:50:44 dhellmann: ERC! :) 15:51:02 ok, let's review it and merge it then... 15:51:16 +2 15:51:47 did we settle on a time for that meeting tomorrow? 15:51:50 so, back on the release meeting then. I can do 14UTC is jd__ is ok with that, busy the hour before though 15:51:55 oh snap, i accidentally prematurely approved that patch 15:52:12 lol 15:52:14 wait, i guess i am 2nd reviewer. so gtg 15:52:30 np, thanks jtran 15:52:56 14UTC works for me 15:53:07 gmb? 15:53:09 cool for me 15:53:15 Works for me. 15:53:30 still waiting for jd__ 15:53:31 nijaba: 14:00 UTC is the time I leave, so… :) 15:53:43 :/ 15:54:01 ok, so we'll do it at 14UTC, but without jd__ :( 15:54:11 and french trains in the country don't have good network connectivities 15:54:31 jd__: you could email the list if you have anything you want us to address 15:54:35 #agreed release meeting tomorrow @14UTC in #openstack-metering (no logs!) 15:54:50 dhellmann: sure, no prob 15:55:10 ok, let's move to the most important topic: 15:55:19 #topic Open Discussion 15:55:21 I can meet as early as 11UTC if we want to move it earlier 15:55:53 just a quick reminder that sdake floated the idea of a pre-summit meetup on Sunday evening at the conference hotel 15:56:20 eglynn: i'd be up for that 15:56:26 eglynn: yes, me, too 15:56:33 not that i'm contributing anything :) 15:56:34 eglynn: great idea, but I am already booked for that evening :( 15:56:39 jtran, dellmann: cool 15:56:44 some time that evening, around or after 18:00? 15:56:48 but feel free to do it without me 15:57:04 my flight gets in at 18.15, so maybe a bit later? 15:57:17 eglynn: sure, maybe drinks instead of dinner then? 15:57:32 dhellmann: sounds good 15:57:42 I've started a text/html render quickly for fun for ceilometer-api http://naquadah.org/~jd/Screenshot%20from%202012-10-11%2017:51:49.png 15:57:42 eglynn or dhellmann: if you ca send me your cell #, I might be able to join you a bit later 15:57:57 nijaba: will do 15:58:04 eglynn: thanks 15:58:12 nijaba: great 15:59:01 Almost running out of time. should we clse the meeting and go in -infra? 15:59:07 +1 15:59:20 #endmeeting