15:01:00 #startmeeting openstack-helm 15:01:01 Meeting started Tue Oct 24 15:01:00 2017 UTC and is due to finish in 60 minutes. The chair is mattmceuen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:04 The meeting name has been set to 'openstack_helm' 15:01:09 GM/GE everyone! 15:01:13 o/ 15:01:38 o/ 15:01:39 here's the agenda -- please take a moment to update w/ anything you'd like to discuss today https://etherpad.openstack.org/p/openstack-helm-meeting-2017-10-24 15:01:54 #topic rollcal 15:01:58 #topic rollcall 15:02:46 o/ 15:02:59 o/ 15:03:06 o/ 15:03:40 o/ 15:04:12 o/ 15:04:29 #topic update to official project status 15:04:42 The update to the status is that the project is official :) 15:04:56 great!!! 15:05:21 Congrats everyone - this is going to be a great opportunity to get more folks using and contributing to our project 15:05:26 \o/ 15:05:30 :) 15:06:09 on the coattails fo that: 15:06:15 #topic who is going to sydney 15:06:35 Sydney will be a good opportunity to make folks aware of OSH as well 15:07:01 * portdirect tries to set topic to "whos coming back from sydney" 15:07:15 https://www.irccloud.com/pastebin/jXdjNFjI/ 15:07:39 jayahn who is the +3? 15:08:14 seungkyu ahn, robert (our ci guy), and one network guy 15:08:27 awesome 15:08:45 staceyf is going as well 15:09:09 and bomman from juniper 15:09:45 o/ 15:09:46 this would be great at sydney 15:09:57 Madhukar from Juniper will be there too 15:09:59 * jayahn who wants to do sky diving with us? 15:10:12 jayahn: when? 15:10:24 we are targeting sunday 15:10:38 jayahn: before? 15:10:50 portdirect: good point 15:10:51 if so I'll need to check when my flight gets in :D 15:11:11 #topics summit talks of note 15:11:25 it will be good discussion with staceyf since we are trying to do basic HA test automated in ci pipeline 15:11:47 I won't paste them here -- if anyone has any talks they think would be relevant to OSH folks, please add them into the etherpad! https://etherpad.openstack.org/p/openstack-helm-meeting-2017-10-24 15:12:01 jayahn: yes definitely 15:12:31 at&t keynote in the morning will mention osh? 15:12:50 We'll see :) 15:13:12 pls put a work "skt" in your keynote.. :) 15:13:30 Just run up on stage jayahn 15:13:40 i can do that :) 15:13:48 * srwilkers imagines jayahn pulling a kanye west 15:14:49 #topic Log-based metrics 15:14:53 okay, community meeting. need to be a bit more serious 15:14:57 The floor is yours mateuszb! 15:15:19 log-based metrics is a serious topic 15:15:24 Ok thanks 15:15:26 https://review.openstack.org/#/c/514655/ 15:15:42 I prepared a PS with the demonstration of fluent-prometheus plugin for log-based metrics purposes. @srwilkers - may I ask you to please take look at it to see if it's heading toward a good direction? There's a detailed description in PS's comment. 15:16:11 mateuszb: definitely 15:16:34 i really would like to suggest to separate our logging effort into several steps 15:16:55 it seems like we are thining too much in one shot 15:17:27 o/ sorry late this morning. 15:17:32 @srwilers thank you. @jayahn for example? 15:17:40 IMO the scope is minimal - I may only separate a Docker image from the PS. Apart from that I don't even touch integration with nagios and alarms. It just presents the simplest use case of prometheus-fluent plugin 15:18:32 there will be more discussion and several ways of implementing alerts. we can totally do that after we first get "fluentbit-fluentd" gathering pipeline done 15:19:35 Sure, I don't want to push the change together with fluentbit-fluentd 15:20:13 My PS is done on top of fluentbit-fluentd and I assume it will be merged only after fluentbit work is done 15:20:30 i know. i was happy to see your ps. 15:21:30 Cool. I also wanted to make sure @jayahn you accept the usage of fluent-prometheus plugin. Last time you talked about writing our own plugin for fluent-bit but I don't know anyone willing to do that :) 15:22:48 i was not suggesting we must develop our own plugin, i was presenting that idea just to back-up my reasoning of doing alert on either on edge or on log store. 15:22:55 :) 15:23:22 ah gotcha. 15:23:40 Ok. btw @jayahn how about alarms definitions? 15:24:18 i am preparing that right now. i can probably upload the first ps at the beginning of next week 15:24:29 awesome 15:25:25 I have seen you asked @srwilers about a reference implementation for alerting system. Is anything established in this matter? 15:25:30 btw, we all agree on this >> https://review.openstack.org/#/c/507023 moving fluent-logging from openstack-helm addons to openstack-helm repo 15:25:48 great 15:25:48 Anything else we can hash through on that one, mateuszb? 15:26:23 mateuszb: not really. we might need to setup some separate meeting (probably web-conference) for that topic 15:26:40 @jayahn yes, once it's moved I'll also move my PS 15:26:44 ok 15:26:59 that's all from my side :) 15:27:06 (please disregard me, was having some connectivity issues and missed your messages :) ) 15:27:13 ok cool 15:27:20 thanks mateuszb 15:27:37 #topic Nova placement service 15:27:49 pls review it. 15:27:55 This PS was getting some discussion in the etherpad, thought I'd bring it here 15:28:28 my quesiton would be not limited to nova placement 15:28:47 as portdirect wrote on etherpad, i have listed things we need to do for ocata. 15:29:27 getting these done would be not difficult, however, then gating should have a separate ocata build 15:29:44 jayahn: +2 to that 15:29:49 agree 15:29:51 what is our plan on completely switching to ocata 15:29:57 can you have a look at the gate work ive been doing in osh-infra 15:30:30 portdirct: i can ask our ci guy to look at the gate work 15:30:55 Re: when is our plan to completely switch to ocata -- once we get Newton done :) 15:31:04 And we're getting close 15:31:14 But still more work to be done 15:31:33 I'd like to work to "catch up" to master over time 15:32:22 then, i would like to have "the list" stating what means done. (we probably had similar discussion on ptg, but i think we just need to have more concrete "task list" for this) 15:32:56 Good idea. 15:33:03 Would we not want to move to master? 15:33:14 #action Matt will work on a consolidated task list for OSH Newton 15:33:19 and have a gate for each of the releases leading from netwon to that? 15:33:55 i am sure we definitely want to move to master. 15:34:40 portdirect: that would be ideal 15:35:53 #topic Review Needed 15:36:11 Jay already brought up https://review.openstack.org/#/c/493422/ 15:36:59 Anyone patiently waiting for reviews that could use a nudge? 15:38:13 okay, then we need to have a stair to heaven - steps we would like to take from current state (newton) to ideal state (catching up master) 15:39:10 Agree. Along with an understanding of which releases our users are most interested in 15:39:19 (and when-ish) 15:40:35 And again - hopefully we'll be getting a lot more users, and hence a lot more input soon :) 15:41:22 #topic Open Discussion 15:41:26 since we just started seeing EOL on newton discussion on mailing-list, probably we need to move up to more recent release soon 15:41:58 I have no doubt we can fast follow a 1.0 newton release with the next 15:42:01 +1 15:42:29 in really short order given the latest handling of config templates 15:42:50 I think the only thing we are waiting for is an official cut of newton 15:43:15 and that mostly releys on moving things to the correct repos 15:43:21 *relies 15:45:33 portdirect: is your thought that we'd branch OSH repo per-release, but have OSH-Infra for example usable by multiple OSH versions? Or am I reading in too much? 15:46:08 * jayahn i think i am doing really good after four pints of beer and three Tequila shots 15:46:13 def +2 to the latter 15:46:41 I hope we could do the former via overrrides, which would not require a branch 15:46:47 jayahn the next OS Summit needs to be in South Korea, that's all I know :-D 15:47:04 so a version of osh, could work with multiple versions of OS 15:47:12 but I need to think about this more 15:47:15 ahh 15:47:53 hm.. multiple versions of openstack can be done only via overrides. that would be ideal. 15:48:06 but, as portdirect said, we probably need to think a bit more. 15:49:04 Any other topics you all would like to discuss? 15:50:00 Hey @renmak__ :) 15:50:10 Hey Matt 15:50:13 Just joined in 15:50:34 Just in time - unless you have any roundtable, I think we're ready to call it a meeting! 15:50:54 Yeah one thing. 15:51:03 go for it 15:51:12 Looking for more feedback on this PS 15:51:13 https://review.openstack.org/#/c/457754/ 15:51:19 This is for Ceph OSD per POD 15:51:35 It needs to be reviewed and merge. Few other US are depending on this change. 15:51:57 So it would be great if folks take some time and provide feedback on this PS 15:52:16 Agree - let's try to round this out this week if possible 15:53:55 That's all 15:54:13 Ok, thanks team - good meeting. Have a great day (and go lay down jayahn :) ) 15:54:16 #endmeeting