16:02:04 #startmeeting kolla 16:02:05 Meeting started Wed Jul 20 16:02:04 2016 UTC and is due to finish in 60 minutes. The chair is sdake. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:02:06 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:02:09 The meeting name has been set to 'kolla' 16:02:11 #topic rollcall 16:02:22 o/ 16:02:28 \o/ :) 16:02:43 o/ 16:03:15 \o/ 16:04:10 o/ 16:04:17 #topic midcycle discussion 16:04:24 this will probablybe a short meeting 16:04:32 i just want to focus on one thing the midcycle produced 16:05:00 https://etherpad.openstack.org/p/kolla-N-midcycle-priority 16:05:11 please open :) 16:05:55 so we flat ranked every priority 16:06:09 based upon our 1-whatever 16:06:14 that is what the numbers represent 16:06:16 hey 16:06:22 lower = more imporant to people 16:06:26 hey rhallisey 16:06:30 just going through priorities now 16:06:42 lets go through em one at a time 16:06:44 documentation 16:06:48 rhallisey is leading that up 16:06:54 we produced a new structure for the documentatoin 16:07:05 the big thing we are doing here is breaking out the conditional logic from the nonconditional logic 16:07:09 I put up some patches to re wrok it 16:07:13 e.g. "i centos do this if ubuntu do this" 16:07:30 documentation is one of the most important things a project can produce 16:07:37 pbourke, pointed out a better way to do it 16:07:39 https://raw.githubusercontent.com/openstack/openstack-manuals/master/doc/install-guide/source/environment-memcached.rst 16:07:42 it lowers the core team's time output on evals 16:08:04 o/ 16:08:14 hey dave-mccowan 16:08:27 so rhallisey are you going to use this mechanis mand redo all the quickstart with it? 16:08:54 yes 16:08:58 o/ sorry i'm late 16:09:01 how does only work 16:09:04 does it detect your host os 16:09:07 or do you select? 16:09:29 let me grab how it's rendered 16:09:33 i think it works on select 16:09:55 tripleo has similar doc structure 16:09:56 well anyway however its done 16:10:03 its something that needs to land prior to newton ;) 16:10:06 http://docs.openstack.org/mitaka/install-guide-ubuntu/environment-memcached.html 16:10:13 that's how it's rendered 16:11:03 how ubuntu/centos is selected? 16:11:10 bifrost and host playbooks 16:11:22 we also decided we want to make this ahappen befoe newton 16:11:33 if your interested in getting involved sean needs help 16:11:45 he is going to place items in teh work list in launchpad 16:11:59 rather work items in the work list ;) 16:12:05 customization 16:12:11 inc0 is leading nthis effort 16:12:21 we definately need this functionality 16:12:30 one interesting thing customizations may offerus in the future is pluggability 16:12:36 i tihnk we will have to think more about that next cycle 16:12:48 monitoring we pushed to occata 16:12:54 voting gates was hot topic 16:13:03 we are going to make the build gates voting as soon as feasible 16:13:12 this depends on me finishign the job on the mirroring 16:13:16 there are 3 mirrors to go 16:13:22 for ubuntu 16:13:25 then we will have an ubuntu voting gate 16:13:30 there are 4-5 mirrors for centos 16:13:35 then we will have a centos voting gate 16:13:37 this is for build only 16:14:22 any questions? 16:16:58 repo split 16:17:01 we moved to passt branch 16:17:09 still open for discussion 16:17:19 depends on core review team to handle backportss properly 16:17:30 sdake, I have a draft email for that 16:17:42 non-ini config 16:17:44 we need that 16:17:45 it will start the vote 16:17:47 inc0 is going to tackle that 16:17:59 send it out after the meeting 16:18:37 osic cluster 16:18:52 we have about 8-10 test cases 16:18:56 after which we will pubilsh a whitepaper on the effort 16:19:00 1-2 pages 16:20:00 this is good when are we starting it? 16:20:10 huh? 16:20:26 dates i mean 16:20:32 tentative 16:20:38 2-3 weeks out 16:20:42 i'd have to check the mailing list 16:21:06 i'll ask inc0 to send out a formal announcement 16:21:17 it starts in couple of days 16:21:33 3 weeks between 7/22 and 8/14 16:22:15 thanks 16:22:21 i guess we better get cracking on that then 16:22:39 that is what is officially on our plate for newton 16:22:43 the above 16:22:47 if you work on something else - sweet ;) 16:22:56 but these are the things we think are best for project progression 16:23:08 i idn't mention k8s 16:23:22 kolla-kubernetes is early - i expect folks there to work on whatever needs to be worked on 16:23:39 ya it's all spelled out in launchpad 16:23:52 once we have a working release - which could be 3-12 months out, we will probbly end up wit this same priority planning 16:23:59 ok well I know that was alot to absorb ;) 16:24:03 #topic open discussion 16:24:55 if no open discussion, we can end the meeting 16:25:02 no use of updating the agenda on wiki :/ 16:25:06 about COPY_ONCE 16:25:10 i think Jeffrey4l added something to meeting 16:25:16 here we go 16:25:16 there's some stuff in the agenda 16:25:24 my bad, i didn't think the agenda got updated 16:25:25 since we had the midcycle 16:25:51 should we remove COPY_ONCE now or future? 16:25:54 not all people attended midcyle :) 16:25:59 #topic remove copy_once 16:26:42 it cause many confuse and there are lots of codes to handle it. 16:27:05 it was the idempotent option 16:27:14 or however we talked about it 16:27:17 But i think few/none people use that copy mode. ( we have copy_always in default now) 16:27:34 ya I don't know who uses copy_once either 16:27:35 iirc, this is talked in the mitaka cycle. 16:28:01 my take is we keep it for now 16:28:03 after that, we move from copy_once to copy_always as default. 16:28:08 i have heard alot of people want to use it for products 16:28:15 copy_alwyas is already the default 16:28:42 yes. but remove copy_once save lots of code and logical. 16:28:45 :) 16:28:54 Jeffrey4l, how can we make it less confusing 16:29:09 well that's maybe more of a general question 16:29:18 remove it... rhallisey :( 16:29:24 i've been telling folks for 3 years we offer immutable containers 16:29:31 ok. 16:29:33 if we remove copy_once we remove immutable containers from our stack 16:29:40 i'm not opposed to making that announcement 16:29:51 but think it involves a more thorough discussion then an irc meeting 16:30:14 got. 16:30:38 immutable that's the word I was looking for 16:31:37 another question: how about heka? should we go back to logstash in the future? 16:31:56 #topic heka EOL 16:32:00 logstash is a nonstarter 16:32:05 logstash is EOL as well :( 16:32:21 so what isn't EOL 16:32:52 so what is the best tool? 16:33:04 ya annoying 16:33:09 that post mentions hindsight 16:33:16 there is another thing from the elastic people 16:33:47 It has currently no mailing list, the docs are minimal, and there are 16:33:47 no configuration examples (except the benchmarks dir). Any plan to 16:33:47 improve this? Heka's doc is one of its "selling" points. 16:33:54 it was unclear to me how long heka would live 16:34:02 ^from mail list about hindsight 16:34:05 Jeffrey4l where is that from 16:34:06 Yeah. I'm kinda anti-heka. 16:34:15 https://mail.mozilla.org/pipermail/heka/2016-May/001061.html sdake 16:35:05 https://github.com/elastic/beats is the designated 'elasticsearch' solution. 16:35:32 I rather much like, as a kubernetes snob, the notion of having logging go to stdout. 16:35:48 there's telegraf if another 'go' based option is desired 16:36:07 And then have all of the collection and tagging and organization of logs be done outside of the running process containers. That might be a tall order for Ansible. 16:36:12 Beats is also GO. 16:36:14 wirehead_, forcing all the app to print the log to stdout is hard 16:36:34 ok well - 3 months is not sufficient time to absorb this change 16:36:40 with all the other work we have on our plate 16:36:54 if someone wants to start prototyping a solution -w fm :) 16:37:36 anyone have a better solution? 16:37:37 Jeffrey4l: Indeed. Then again, that's how all of my non-OpenStack projects in the past 5 years have worked and it's worked out quite well. 16:38:39 sdake, no :( 16:38:54 ok so lets fix things up to beats 16:39:07 unless someone has a different plan 16:39:14 but lets do it at the start of the cycle 16:39:16 wirehead_, ya. use stdout is cool. But there are some app can not print log to stdout. 16:40:02 pity beats doesnt seem to have a ceph module yet 16:41:06 so move on? 16:41:13 Ceph bug in Liberty & Mitaka - https://bugs.launchpad.net/kolla/+bug/1604129 16:41:13 Launchpad bug 1604129 in kolla "storage_interface problem" [Critical,Triaged] - Assigned to Bartłomiej Daca (bartek-daca) 16:41:29 there is a bug on the agenda. need more eyes. 16:41:41 ya this is an intersting one 16:42:24 bug with ansible >2.0.0 16:42:40 other way around 16:42:44 < 2.0.0 16:43:15 right 16:43:17 in ansible >2.0.0 it is problematic too 16:43:36 bdaca, I though >2.0.0 fixes the issue? 16:43:49 yes, but open another one 16:44:14 problem is: xtra vars are always on top 16:44:48 so anything specified in inventory will fail 16:46:05 bdaca, there's a fix for this in master though 16:46:10 anyone have time to fix this problem? 16:46:14 thought I saw some patches for it 16:46:28 didn't see any 16:46:36 ok nvm 16:46:47 i can handle this, just tell me if proposed fix is ok 16:47:01 (in the last comment) 16:47:26 bdaca, ok will reivew 16:47:26 cool so lets just work this via the normal review process 16:47:32 rhallisey can you put a star on it plz 16:47:36 roger 16:47:45 ok anything else? 16:48:05 there's Project Mascot on the schedule 16:48:08 idk what that is 16:48:39 ya we already voted on this 16:48:40 we need to submit it before july 27 16:48:56 oh that's literally the mascot 16:49:07 I took that as something entirely different 16:50:41 its been submitted 16:51:32 #topic open discussion 16:51:48 guess took longer then i thought - apologies for not readin gthe agenda :( 16:52:21 sdake: np 16:53:24 ok thanks for coming 16:53:34 lets get cranking on getting our top dev priorities finished 16:53:56 if your looking for stuff not on the list, there are aother areas we can focus on such as the pain points 16:54:10 or just work on what you like 16:54:16 and we will sort it out at release time ;) 16:54:24 #endmeeting