15:00:15 <gordc> #startmeeting ceilometer
15:00:16 <openstack> Meeting started Thu Apr 30 15:00:15 2015 UTC and is due to finish in 60 minutes.  The chair is gordc. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:20 <openstack> The meeting name has been set to 'ceilometer'
15:00:26 <sileht> \o
15:00:27 <jasonamyers> o/
15:00:36 <llu-laptop> o/
15:00:37 <cdent> o/
15:00:40 <prad> o/
15:00:41 <fabiog> hi
15:00:43 <eglynn> o/
15:01:22 <gordc> just a heads up before we start. similar to last year. i'll give those who proposed talks a quick elevator pitch to explain their sessions
15:01:45 <gordc> so you have x minutes while we run through other topics to type up a blurb (if you feel it necessary)
15:01:57 <gordc> that said...
15:02:02 <gordc> #topic kilo release
15:02:09 <idegtiarov> o/
15:02:28 <gordc> i'm currently not tracking any critical bugs
15:02:37 <ityaptin> o/
15:02:39 <gordc> i believe we are still on track to release today?
15:02:45 <gordc> eglynn is that true?
15:02:56 <eglynn> yes kilo-rc2 will be final
15:03:02 <cdent> ttx has announced
15:03:04 <prad> i see tweets that its out
15:03:35 <eglynn> just say "non!" to more RCs ;)
15:03:37 * gordc is going to resist twitter
15:03:47 <gordc> eglynn: agreed.
15:03:49 <prad> :)
15:04:04 <gordc> so yeah, i guess kilo's done. good job everyone.
15:04:08 <ildikov> o/
15:04:14 <eglynn> here here! :)
15:04:15 <cdent> go us
15:04:29 <gordc> if you do see bugs, tag them for kilo-backport-potential.
15:04:58 <gordc> #topic summit ops session
15:05:21 <gordc> as mentioned in previous meetings, we have our slot set
15:05:32 <gordc> now we have an etherpad: https://etherpad.openstack.org/p/YVR-ops-ceilometer
15:05:42 <gordc> if you have topics you want to discuss with ops put it there ^
15:05:54 <llu-laptop> what's YVR?
15:06:04 <cdent> vancouver airport code
15:06:46 <llu-laptop> gee, why I didn't remember see this, did I book the wrong flight, checking... :)
15:06:47 <eglynn> airport code shothand really only works when it's obvious
15:07:08 <gordc> yyz in toronto... i don't know why. yul in montreal.
15:07:10 <fabiog> I guess next time will be NRT
15:08:02 <prad> s/YVR/Liberty
15:08:08 <jd__> o/
15:08:29 <gordc> basically... i don't know if anyone has questions regarding the session. i guess we've discussed it enough
15:08:45 <gordc> #topic recurring: ceilometerclient release?
15:09:20 <gordc> i made a ceilometerclient 1.0.14 release for kilo on monday... i'm not track anything else. so yeah... yell now for another release of something
15:09:49 <gordc> no? no. done.
15:09:56 <gordc> #topic recurring: Gnocchi status
15:10:09 <gordc> jd__: sileht: cdent: anything?
15:10:30 <jd__> I'm working on adding Alembic support for 1.1
15:10:41 <jd__> and I'm polishing stable/1.0 branch to release 1.0.0rc2 today I hope
15:10:47 <jd__> and a final 1.0.0 next week
15:10:56 <jd__> I think sileht is doing some testing or something
15:11:08 <gordc> lol
15:11:10 <jd__> he does not talk much but he comes with a lot of bug reports and fixes from time to time :p
15:11:33 <prad> i'm looking into moving dispatcher code to ceilo from gnocchi https://review.openstack.org/#/c/178209/
15:12:13 <sileht> jd__, gordc I try to make the swift driver working well in devstack
15:12:40 <gordc> prad: do you need to update to pull in jd__'s patch to filter swift metrics?
15:13:08 <sileht> I think we should move the dispatcher just after the 1.0 have been released, WDYT ?
15:13:15 <gordc> sileht: cool cool. that'd be good. i'm updating ceilometermiddlware right now
15:13:17 <prad> gordc, probably i'll check, i'll pulled in sileht's patch on http one recently
15:13:43 <gordc> sileht: yeah. that might be safer? i assume you guys are still making changes while testing integration
15:13:48 <sileht> prad, thanks your change looks up2date, I will hold of touching the dispatcher, now :)
15:13:50 <jd__> sileht: yep but I think 1.1 is going to be released while Kilo is out so we can't remov the dispatcher yet
15:13:59 <jd__> sileht: I expect people to use Kilo + Gnocchi 1.1 if they want
15:14:00 <prad> :)
15:14:12 <jd__> sileht: though with the requirements mess it might not work lol
15:14:52 <sileht> jd__, I have seen a mail that say they rethink the thing (again) :p
15:15:00 <jd__> sileht: -_-
15:15:10 <jd__> sileht: anyway it's too late for Kilo
15:15:17 <sileht> yep
15:15:23 <jd__> ceilometer is going to have crap such as oslo.db<1.8
15:15:31 <jd__> and Gnocchi 1.1 oslo.db>=1.8 :D
15:15:57 <gordc> jd__: ... that's going to be interesting.
15:16:08 <sileht> jd__, distro will hate us
15:16:14 <gordc> jd__: there's imaginary numbers in there that overlap right?
15:16:56 <prad> packaging update: i submitted gnocchi rpms for fedora review
15:17:15 <gordc> prad: cool stuff.
15:17:34 <gordc> did we ever find someone to do deb?
15:17:46 <eglynn> zigo?
15:18:05 <jd__> gordc: that's crap because Ceilometer Kilo is working with oslo.db 1.8
15:18:14 <jd__> gordc: so the requirement is wrong
15:18:31 <jd__> I talked to zigo about Gnocchi for Debian but not sure what is the status
15:19:11 <jd__> http://anonscm.debian.org/cgit/openstack/gnocchi.git/
15:19:20 <gordc> jd__: :( i follow that to see how we can get that working. i'll probably be looking at release stuff more going forward.
15:19:34 <jd__> gordc: good luck with that ;)
15:19:40 <jd__> as sileht said, it might change
15:19:49 <eglynn> jd__: if the oslo.db<1.8 requirement is wrong, that's not terrible as the distros don't need to slaveishly follow the upstream requirements.txt (i.e. it's more of a "hint")
15:20:01 <gordc> jd__: hopefully it involves 0 effort.
15:20:03 <eglynn> not ideal either though, obviously
15:20:31 <jd__> eglynn: … except that I *think* the file is copied by pbr into dist-package and read at startup or something in some cases such as loading drivers via stevedore
15:21:10 <sileht> jd__, I confirm
15:21:12 * gordc is still hoping for 0 effort/work
15:21:16 <jd__> eglynn: because of the fact pbr is taking the install_requires from requirement.txt which is actually a wrong behaviour we need to change (discussed on openstak-dev recently)
15:22:44 <gordc> jd__: i guess we will watch that thread. you have link handy?
15:22:44 <eglynn> jd__: hmmm, maybe I'm missing some aspect of the packaging workflow ... wouldn't the first time that happened
15:23:37 <jd__> gordc: eglynn http://lists.openstack.org/pipermail/openstack-dev/2015-April/061282.html
15:23:53 <gordc> jd__: cool cool
15:24:14 <gordc> so anything else gnocchi related?
15:24:56 <gordc> #topic summit session proposals - VOTING
15:25:12 <gordc> here are topics: https://docs.google.com/spreadsheets/d/1vlEe75uZc5faytpvIMPrQdViOOsRLnfeThk_Rp_lIvI/edit#gid=0
15:25:35 <gordc> there's a column for you to put your name if you're interested
15:25:58 <gordc> you can still add stuff but yeah, i guess someone might not see it if you add it later
15:26:15 <gordc> we'll close voting sometime next week close to meeting and we'll schedule things accordingly.
15:26:50 <cdent> what's the difference between interest and a vote?
15:26:52 <gordc> so open elevator pitch time... if you have something additional to add right now, raise your hand.
15:27:26 <gordc> cdent: ...i didn't even see the vote part.lol i'm goig to assume votes is just a tally of interested
15:27:44 * cdent gets gordc a wider screen
15:27:53 <eglynn> on the oslo.versionedobjects one ... there are corresponding sessions on the olso track and for many individual projects also
15:27:55 <fabiog> gordc: this is my elevator pitch
15:28:08 <fabiog> Ceilometer should be comprised of a set of components (micro-services) as git repositories. These components should be as independent as possible and have the database technolgy that is best suited for the task and not have to support a matrix of alternatives. Morevoer pipelines for Metric, Events should be completely sepearate as well as their API. So I can deploy a Event only solution (Ceilometer-Collection + Ceilometer
15:28:14 <cdent> fabiog is your session a lot the same as mine? can we merge, overlap or define the distinction?
15:28:32 <gordc> fabiog: ^ was going to say the same.
15:28:38 <jd__> I just wrote that cdent, fabiog lol
15:28:47 <sileht> cdent, fabiog yeah that looks the topic
15:28:48 <cdent> fabiog: if you look at the linked etherpad from mine you'll see
15:28:53 <gordc> we might merge the two and you might want ot copresent
15:29:03 <fabiog> cdent: sure
15:29:12 <fabiog> gordc: I am ok with co-present
15:29:23 <cdent> FISTICUFFS
15:29:28 <gordc> fabiog: cool cool. if we have time we might just split it across two sessions
15:29:38 * cdent is not sure what we'd argue about though. violent agreement
15:29:53 <fabiog> gordc: I just want to decouple the role ceilometer plays, like monitoring vs. metering from the fact that the code should be organized differently
15:29:59 <ildikov> the alarm/event session proposals seem to be overlapping too
15:30:16 <gordc> fabiog: kk. i think that is a shared idea by others
15:30:31 <gordc> ildikov: yeah, we'll probably be merging quite a bit.
15:30:39 <fabiog> cdent: are you going to be at the summit?
15:30:40 <llu-laptop> yeah I see our event alarm proposal are very much alike with Ryota's getting more out of events,
15:30:54 <llu-laptop> but he beats ours in proposing time
15:30:57 <llu-laptop> we could merge
15:31:05 <gordc> llu-laptop: yep. Ryota, will be at summit so you can sync up.
15:31:13 <ildikov> gordc: so I think we are quite good with time and proposals to allocate enough for discussions like the componentization/identity one
15:31:15 <llu-laptop> good to hear that
15:31:26 <gordc> ildikov: i think so too
15:31:56 <gordc> eglynn: is the oslo.versioned proposal dependent on the other sessions?
15:32:30 <eglynn> gordc: my main interest is just figuring out what olsoversionedobjects can do for ceilo (if anything)
15:33:06 <gordc> eglynn: kk. disclaimer: i didn't read many of the oslo.versioned threads
15:33:31 <gordc> eglynn: but will be good to look into that
15:33:47 <gordc> anyone else have a elevator pitch?
15:33:49 <ildikov> eglynn: gordc: we have a specs for that, right? or at least I think I saw something related
15:34:03 <cdent> specs and in progress changes
15:34:04 <gordc> ildikov: i think it got proposed to all the projects
15:34:24 <ildikov> gordc: a-ha, ok, I haven't checked that part
15:35:09 <gordc> ildikov: yeah... i don't know how they did it. everyone must really like dansmith. :)
15:35:12 <ildikov> as for elevator pitch, I think it would be nice to agree on the meter deprecation process that we would like to follow and then cover it with docs and also implementation to be more user friendly
15:35:45 <ildikov> gordc: lol :)
15:36:16 <gordc> ildikov: just for reference i added the deprecation topic to ops meeting too
15:36:46 <ildikov> gordc: ok, cool, maybe we can merge it then with the ops session outcome slot
15:37:17 <ildikov> gordc: let's see when we get there to schedule the session proposals
15:37:21 <gordc> ildikov: kk. someting to consider when we assign slots
15:37:37 <gordc> cool cool... anything else?
15:38:13 <cdent> bowling
15:38:20 <gordc> sshhh!
15:38:27 <cdent> ha ha only serious
15:38:34 <cdent> (not really)
15:38:47 <gordc> we'll close voting next thursday?... let's say when i wake up. EST
15:39:11 <gordc> so just guess when i wake up... and vote based on that.
15:39:27 <gordc> #topic open discussion
15:39:39 <llu-laptop> just want to make an announcement
15:39:45 <gordc> cdent: i think puppet team wanted to a bowl off against us.
15:39:52 <llu-laptop> we brought out 3rd party CI for ceilometer
15:40:00 <llu-laptop> it's about testing the ipmi meters
15:40:15 <gordc> llu-laptop: cool! i saw that new check.
15:40:18 <cdent> has been seeing the mail, is it working out?
15:40:22 <llu-laptop> only one test case now, but we're planning to add more test cases, like snmp related too
15:41:08 <gordc> llu-laptop: so ping you if it starts failing?
15:41:31 <prad> llu-laptop, meant to ask about that, its a non-voting check i assume?
15:41:55 <llu-laptop> gordc: ok. but just for a notice, right now, most of the failure comes from network connection issues in our crappy intel's network environment
15:42:08 <llu-laptop> prad: it's not voting now
15:42:33 <prad> cool
15:42:45 <gordc> llu-laptop: lol appreciate the honesty
15:43:22 <llu-laptop> more than 90% of the failure comes from those network issues, we have 4 3rd party CIs, for different OS project
15:44:27 <jd__> yay more spam
15:44:53 * cdent is pretty sure jd__ knows about filters
15:45:15 <gordc> cdent: i use hotmail. it has crappy filters...
15:45:22 <cdent> heh
15:45:27 <gordc> i get everything or nothing... i chose nothing
15:45:36 <fabiog> I have to go. Talk to you all next week
15:45:54 <gordc> cool cool. anything else?
15:46:14 * cdent thinks about tim hortons
15:46:35 <jd__> cdent: :-)
15:47:19 <gordc> cdent: you can dream in os-ceilometer.
15:47:28 <gordc> shutting it down
15:47:36 <gordc> thanks everyone.
15:47:39 <gordc> #endmeeting