15:00:33 <gordc> #startmeeting ceilometer
15:00:34 <openstack> Meeting started Thu Apr 16 15:00:33 2015 UTC and is due to finish in 60 minutes.  The chair is gordc. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:37 <sileht> o/
15:00:38 <openstack> The meeting name has been set to 'ceilometer'
15:00:44 <cdent> o/
15:00:56 <jasonamyers> o/
15:00:56 <eglynn> o/
15:01:46 <jd__> o/
15:01:50 <llu-laptop> o/
15:02:14 <_nadya_> o/
15:02:18 <ildikov> o/
15:02:29 <gordc> quorum.
15:02:40 <gordc> let's start :)
15:02:53 <gentux> o/
15:02:57 <gordc> #topic kilo-rc requests/bugs
15:03:20 <gordc> i think we currently have one patch proposed for rc2...
15:03:38 <gordc> i don't know if anyone else has a bug they've found that they'd like to have in kilo?
15:03:49 <gordc> (haven't seen any in launchpad)
15:04:16 * cdent nopes
15:04:21 <eglynn> #link https://bugs.launchpad.net/ceilometer/+bugs?field.tag=kilo-rc-potential
15:04:24 <eglynn> nada
15:04:30 <idegtiarov> o/
15:04:55 <eglynn> tagging with kilo-rc-potential is the signal if you've got something you want to raise up as a candidate for rc2
15:05:11 <gordc> ^ that will help a lot
15:05:46 <gordc> April 30 is the release date. obviously, bringing blockers up early is better.
15:06:20 <gordc> next?
15:06:22 <gordc> next
15:06:29 <gordc> #topic kilo release notes
15:06:37 <gordc> https://wiki.openstack.org/wiki/ReleaseNotes/Kilo
15:06:52 <gordc> i've added a few items that i can remember we added this cycle.
15:07:06 <gordc> please have a look and see if there's something important you think we should highlight
15:07:28 <eglynn> mention tech-preview gnocchi integration?
15:07:46 <ityaptin> o/
15:07:52 <gordc> eglynn: good point. was going to action item that for jd__
15:07:56 <eglynn> cool
15:08:02 <ildikov> gordc: I wonder if I should mention there the Event description in OS Manuals that we have finally and maybe in the same time we could clean up the dev docs part if that is needed?
15:08:03 <gordc> jd__: deflected :)
15:08:39 <jd__> gordc: ack
15:08:46 <gordc> ildikov: i guess we can add a link to event part?
15:08:57 <ildikov> gordc: ok, I will do that
15:09:04 <gordc> eglynn: is the page generally just pushed to ops list?
15:09:21 <llu-laptop> anyone can edit the release notes or it's just the PTL's privilege?
15:09:31 <gordc> llu-laptop: anyone can edit
15:09:31 <eglynn> llu-laptop: anyone, it's on the wiki
15:09:48 <gordc> llu-laptop: good point... please don't tell me to edit!
15:09:48 <eglynn> are we missing anything else from https://blueprints.launchpad.net/ceilometer/kilo ?
15:09:49 <ildikov> llu-laptop: you only need to sign in to the wiki
15:09:54 <eglynn> e.g. the hyper-v stuff
15:09:55 <llu-laptop> got that thx, might add something tommorrow
15:10:45 <jd__> gordc: I've just mentioned Gnocchi as a new dispatcher in those release notes
15:10:46 <_nadya_> ildikov: btw, do we have description of partition coordinator for agents in docs, for HA? I didn't find it and thinking of writing it
15:10:52 <gordc> eglynn: right. i'll add that. i guess a simple 'added more metrics for hyper-v' is suffice?
15:10:58 <eglynn> yeah, I think so
15:11:13 <gordc> eglynn: cool cool.
15:11:35 <gordc> _nadya_: i think we do... at least a blurb about it
15:11:43 <ildikov> _nadya_: there is in the Admin guide
15:12:10 <_nadya_> ildikov: aha, will find there, thanks a lot!
15:12:13 <ildikov> _nadya_: not sure about the dev docs at the moment
15:12:17 <gordc> _nadya_: i might try creating a diagram relating to that... but it might be more for dev docs
15:12:52 <ildikov> _nadya_: http://docs.openstack.org/admin-guide-cloud/content/section_telemetry-cetral-compute-agent-ha.html
15:13:34 <gordc> ildikov: any help needed on cleaning up existing docimpacts from kilo?
15:13:42 <_nadya_> gordc: don't worry, I was going to do it by myself :)
15:13:54 <ildikov> gordc: I agree, it seems a detail that should go to the dev docs on the first place and I also agree it would nice to have, so thanks in advance :)
15:13:55 <gordc> _nadya_: kk
15:14:04 <_nadya_> ildikov: great!
15:14:19 <ildikov> gordc: I will double check, it mostly should be fine
15:14:27 <gordc> ildikov: cool cool
15:14:41 <ildikov> gordc: the configuration update is already done
15:15:00 <ildikov> gordc: do you have anything in mind for install guide that we haven't covered already?
15:15:23 <gordc> ildikov: not currently. i'll probably give another read through in next few days
15:15:39 <ildikov> gordc: the other docs are not versioned yet, these are the two first priority...
15:15:46 <ildikov> gordc: cool, sounds good, thanks!
15:16:02 <gordc> for those interested: http://docs.openstack.org/, search telemetry, see if it makes sense. :)
15:16:34 <gordc> #topic gnocchi updates
15:16:56 <gordc> i think i deleted the previous floating topic by accident... but jd__, any updates?
15:17:06 <sileht> history have been landed
15:17:27 <sileht> archive policy rules too
15:17:28 <gordc> \o/
15:17:34 <cdent> \o/ indeed
15:17:37 <prad> yep ap rules landed
15:18:16 <eglynn> prad is working on gnocchi packaging for Fedora, should make it more consumable :)
15:18:19 <sileht> we have etag for resources :)
15:18:35 <gordc> are there docs we can point to from release notes (in case people want to test)?
15:18:47 * gordc does not understand packaging.
15:18:59 <sileht> I have retested the integration with ceilometer and devstack
15:19:10 <sileht> everything seems work out of the box
15:19:11 <eglynn> is this up-to-date ... http://gnocchi.readthedocs.org/en/latest/
15:19:16 <prad> yea packaging coming along, should  have something soon installable soon (even if its from copr) untile fedora reviews
15:19:27 <jd__> I'm trying to realease 1.0.0rc1
15:19:30 <eglynn> gordc: ^^^ re. docs question
15:19:39 <jd__> I'm having some permission issue on Gerrit I've sent a patch on infra for that
15:19:57 <jd__> eglynn: no it's on 18592421
15:20:02 <jd__> err
15:20:03 <jd__> http://docs.openstack.org/developer/gnocchi/
15:20:05 <gordc> eglynn: yeah. i think we'll probably need a 'using it with devstack' part later
15:20:08 <eglynn> a-ga, cool
15:20:18 <eglynn> a-*ha :)
15:20:33 <gordc> nm i see it
15:20:53 <gordc> jd__: cool stuff, we should add that to release notes
15:20:57 <jd__> I've deleted the readthedoc project
15:21:06 <jd__> well at least I've clicked the DELETE button
15:21:34 <jd__> no stale doc online!
15:21:42 <gordc> request google to stop showing it
15:21:59 <gordc> anything else on gnocchi?
15:22:06 <jd__> I think that's all
15:22:25 * gordc changing order
15:22:29 <gordc> #topic ceilometerclient release
15:22:46 <gordc> i made a 1.1.0 release for Liberty. it will include everything that's merge as of a few hours ago
15:23:17 <gordc> anything between 1.0.13 and 1.1.0 needs to be backported to stable/kilo so if you have anything important, start backporting
15:23:38 <prad> gordc, i did a few from your list
15:23:54 <prad> i think we have a couple more left, i should have those soon i think
15:24:00 <gordc> i believe the rule going forward is if a requirements patch is merged, we should probably bump the minimum version number and make a release
15:24:21 <gordc> prad: thanks!
15:24:41 <gordc> i don't think there's much to talk about this so onto summit talk!
15:24:53 <gordc> #topic summit session proposals
15:25:11 <gordc> i created a link for proposal: https://docs.google.com/spreadsheets/d/1vlEe75uZc5faytpvIMPrQdViOOsRLnfeThk_Rp_lIvI/edit#gid=0
15:25:36 <gordc> let me know if you can't edit it... honour system in place so don't randomly delete stuff.
15:26:09 <ildikov> gordc: I added a placeholder session there, so it is editable
15:27:03 <gordc> eglynn: you have the list of session spaces handy?
15:27:09 <eglynn> gordc: how about we change "Topic for Friday?" column to "Session type (fishbowl/working/meetup)"
15:27:09 <ildikov> cdent: now you can edit the document, you will not be the first one ;)
15:27:18 <cdent> :)
15:27:29 <gordc> eglynn: sounds good
15:27:32 <eglynn> gordc: 2 fishbowl, 8 working sessions, 1/2 day meetup on Friday
15:27:39 * gordc copy/pasted blindly
15:28:37 <gordc> so i guess we have quick a bit of time reserved so if you want face to face time, the summit will be a good time.
15:29:05 <ildikov> gordc: when do we plan to close the proposals?
15:29:38 <eglynn> ildikov: usually the near-final list is needed for sched.org about a week in advance
15:29:57 <gordc> ildikov: good question... i hadn't thought that far.lol let's follow the cutoff from last cycle
15:30:05 <eglynn> ildikov: so we'll need to be voting/deciding at least 2 week out from summit
15:30:21 <ildikov> eglynn: cool, thanks for the update
15:30:55 <ildikov> I wonder where it could be placed on the google docs to inform those, who plan to propose a session
15:31:00 <gordc> so i guess aim for your proposals to be summitted within next two week so we have time to vote.
15:31:40 <gordc> ildikov: let's tentatively set May 4 as cut off
15:31:58 <ildikov> gordc: cool, sounds good to me
15:32:10 <gordc> but again earlier the better.
15:32:19 <gordc> in regards to submissions
15:32:44 <gordc> keeping on topic of summit
15:32:50 <gordc> #topic ops-related ceilometer summit session
15:32:54 <eglynn> gordc: yeah maybe do a first round of voting/discussion at the meeting on April 30th?
15:33:16 <gordc> eglynn: sounds good
15:33:27 <ildikov> a bit close but ok
15:33:51 <gordc> ildikov: should be ok... lights a fire.
15:33:52 <ildikov> so I added a placeholder session to discuss the outcome of the ops session
15:34:03 <gordc> http://lists.openstack.org/pipermail/openstack-operators/2015-April/006730.html
15:34:11 <gordc> https://etherpad.openstack.org/p/YVR-ops-meetup
15:34:12 <ildikov> gordc: LOL :), from that PoV you're right
15:34:38 <gordc> there's an op session set for discussing ceilometer (and gnocchi)
15:34:52 <gordc> on Tuesday 12:05 - 12:45
15:35:26 <gordc> it's for developers to ask ops questions and vice versa
15:35:35 <eglynn> we need someone to moderate the ops feedback session
15:35:48 <ildikov> yeap, just wanted to point it out
15:36:25 <gordc> don't all fight over opportunity to moderate.
15:36:26 <ildikov> also somehow we need to balance in oder to get as many feedback as we can, but also answer questions or address issues we can
15:36:56 <ildikov> ... like we have now Gnocchi, etc. ...
15:37:00 <cdent> moderate _at_ the ops session or in our own meeting to think about it?
15:37:17 <gordc> cdent: _at_ the ops session
15:37:26 <eglynn> cdent: wot gordc said
15:37:44 <cdent> I can't: speaking right before
15:38:02 <gordc> cool cool
15:38:12 <cdent> but would otherwise like to be sure to be there as I think it is likely one of the more important things happening all week
15:38:31 <gordc> well aside from moderating, if you have questions to ask ops (ie. how they use ceilometer, what they find useful) please think about it before hand
15:38:54 <gordc> jasonamyers: i don't know if you have any items to add?
15:39:04 <jasonamyers> not at the moment
15:39:13 <eglynn> proposal: all cores for both ceilo & gnocchi who are travelling to Vancouver try to be present at the ops session
15:39:25 <eglynn> what say ye?
15:39:44 <jasonamyers> I've been tasked a get more involved in code for ceilometer starting at the summit so I'm looking forward to that
15:39:53 <ildikov> eglynn: +infinite
15:40:12 <gordc> jasonamyers: cool cool.
15:40:24 <llu-laptop> eglynn: +1
15:40:26 <ildikov> eglynn: this is the most important session on the summit
15:40:57 <ildikov> feedback is priceless and we also need some marketing, from which PoV the support on the ops session is importatnt too IMHO
15:41:59 <gordc> 'need some marketing' /me puts on suit from linked profile
15:42:07 <eglynn> LOL :)
15:42:28 <ildikov> gordc: if you do, then you can moderate the session too ;)
15:42:50 <cdent> Yeah, as much as marketing terminology makes me want to vomit, there does seem to be a need to do a bit brand management.
15:43:22 <gordc> ildikov: i'll need to shine my shoes... after i buy some shoes.
15:43:39 <ildikov> gordc: LOL :)
15:43:50 <ildikov> cdent: sorry for the word, couldn't find any better
15:44:11 <gordc> i don't have anything else on this... so...
15:44:12 <cdent> ildikov: you are right
15:44:14 <ildikov> cdent: but I hope you still got my point
15:44:25 <cdent> I was agreeing with you
15:45:02 <ildikov> cdent: okok, got it now, too many meeings in a row... :S
15:45:20 <gordc> so basically come with questions and answers... the session is meant to go both ways.
15:45:37 <gordc> #Open discussion
15:45:38 <eglynn> gordc: excellent point!
15:46:05 <gordc> ...well that's not how you use it apparently
15:46:07 <gordc> #topic Open discussion
15:46:09 <eglynn> anyone done any testing with tokumx?
15:46:30 <eglynn> (as a higher perf drop-in replacement for mongodb)
15:46:39 <prad> eglynn, i played around with it after you mentioned
15:46:48 <prad> eglynn, obviously dint test it at scale
15:47:13 * eglynn would love to see some hard data on how ceilo performance against it
15:47:19 <gordc> eglynn: is this equivalent to mariadb to mysql?
15:47:35 <eglynn> ... my kingdom for a realistic pre-prod env ;)
15:47:36 <ildikov> eglynn: unfortunately not, as I had to check in Google, what this is :)
15:47:50 <eglynn> gordc: I think it's a close equivalent, yes
15:48:04 <ildikov> gordc: at least sounds like
15:48:24 <gordc> 90% reduction in size?!?!
15:48:29 <sileht> eglynn, according the homepage, that's rocks :p
15:48:39 <eglynn> ".... as of October last year, they were in parity with Mongo 2.4, with the exception of geo-indices and full-text search, and 2.6 is still an open milestone."
15:48:45 <eglynn> https://news.ycombinator.com/item?id=7990766
15:48:49 <sileht> gordc, I love this kind of number :)
15:48:51 <eglynn> meh, that's old new
15:48:54 <eglynn> s
15:48:54 <gordc> these are some bold statements.lol
15:49:20 <prad> see gordc, thats marketing :)
15:49:51 <ildikov> prad: exactly :)
15:50:02 <gordc> prad: :) i'll take note.
15:50:11 <ildikov> they are listing some companies as users, maybe if anyone has connections there it worth a question
15:50:32 <eglynn> does anyone have facilities to profile, e.g. run the Mirantis benchmarks
15:50:40 <eglynn> ?
15:51:07 <eglynn> might be worth having some data in our back-pocket for that ops session
15:51:17 <_nadya_> eglynn: what benchmark :)?
15:51:20 <ildikov> eglynn: my notebook only looks large ;)
15:51:31 <gordc> ildikov: hahah!
15:52:02 <eglynn> _nadya_: the one that Ilya(?) ran and Dina crunched the results, IIRC
15:52:20 <cdent> I think perhaps it would be better if by the time we got to summit we had already convinced one of the larger installations to have tried out tokumx and report benchmarks. We need to make partners not customers.
15:52:37 <sileht> it seems this is a perconna product since two days
15:52:50 <ildikov> although the idea sounds good, I wanted to ask this too
15:52:58 <_nadya_> eglynn: aha, it was just scripts. not a 'tool' . Maybe we may consider Rally
15:53:04 <sileht> https://www.percona.com/percona-acquires-tokutek
15:53:10 <eglynn> cdent: one to raise with linuxhermit perhaps?
15:53:34 <cdent> here today in his jasonamyers guise?
15:53:47 <jasonamyers> I am here
15:53:51 <_nadya_> eglynn: but tbh, I'm not sure if we may apply it to test Ceilo. Will discuss that
15:54:42 <jasonamyers> cdent: numbers are very useful to operators
15:54:43 <ildikov> cdent: if you mean those as partners who use what we produce then I agree, but I wouldn't say this in general
15:55:01 <jasonamyers> especially if you are comparing to options on the same hardware/network setup
15:55:29 <cdent> The point I'm trying to make is that we maintain too much of a division between "us" and the "operators" and I think that is bad for the health of the "product".
15:55:45 <eglynn> _nadya_: cool, thanks
15:56:00 <cdent> If "we" can get "them" to help "us" with something "we" can't do easily (because of lack of data and tooling) but "they" can, then win
15:56:21 <cdent> """"quotes""""
15:56:49 <gordc> cdent: you are very proficient with them.
15:56:52 <ildikov> cdent: sure, i got that part
15:57:27 <jasonamyers> cdent: I can't agree more on that point
15:57:36 <jasonamyers> operators are us
15:57:48 <jasonamyers> they are they reason we need ceilometer etc right
15:57:59 <jasonamyers> if operators move on to other clouds that's terrible for openstack
15:58:30 <gordc> we throw balls at each other in our office... it is us vs them for at least some portion of day
15:59:09 <gordc> i'm just going to let this meeting run out... 1 min
15:59:31 <ildikov> jasonamyers: +1
15:59:39 <rjaiswal_> _nadya_, idegtiarov: Is multiple pipelines still an important use case for you  - https://review.openstack.org/#/c/171826/?
16:00:07 <gordc> outta time sorry
16:00:21 <gordc> please carry on in #ceilometer
16:00:27 <gordc> #endmeeting