16:00:07 #startmeeting Cinder 16:00:08 Meeting started Wed Jul 6 16:00:07 2016 UTC and is due to finish in 60 minutes. The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:11 The meeting name has been set to 'cinder' 16:00:12 ping dulek duncant eharney geguileo winston-d e0ne jungleboyj jgriffith thingee smcginnis hemna xyang tbarron scottda erlon rhedlind jbernard _alastor_ vincent_hou kmartin patrickeast sheel dongwenjuan JaniceLee cFouts Thelo vivekd adrianofr mtanino yuriy_n17 karlamrhein diablo_rojo jay.xu jgregor baumann rajinir wilson-l reduxio wanghao thrawn01 chris_morrell 16:00:13 Hello 16:00:15 Hello. 16:00:17 hi 16:00:17 <_alastor_> o/ 16:00:20 hi 16:00:21 hi 16:00:22 hi/ 16:00:22 Hi! o/ 16:00:23 hey 16:00:27 Just in case... 16:00:27 o/ 16:00:29 hi 16:00:31 hi 16:00:35 #chair jungleboyj 16:00:36 Current chairs: jungleboyj smcginnis 16:00:41 #chair xyang1 16:00:43 קךךם 16:00:45 hello 16:00:46 Current chairs: jungleboyj smcginnis xyang1 16:00:57 hi 16:01:07 smcginnis: What does that mean? 16:01:07 Hello 16:01:11 smcginnis: why so many chairs?:) 16:01:13 hi 16:01:27 o/ 16:01:35 Too many cooks. 16:01:37 jungleboyj, xyang1: In case my internet connection goes out, just wanted to add someone else to be able to at least stop the meeting. :) 16:01:46 ok:) 16:01:59 hi 16:02:04 #topic Announcements 16:02:15 smcginnis: Gotcha. 16:02:26 Same stuff. Please take a look at the review focus etherpad. 16:02:45 I've checked on the new drivers and divided them between the ones with CI reporting and the ones without. 16:02:57 Really kind of a surprising number of them in total. 16:03:20 #link https://etherpad.openstack.org/p/cinder-spec-review-tracking Review tracking 16:03:40 #link https://etherpad.openstack.org/p/cinder-spec-review-tracking Midcycle planning 16:03:52 Please add topics to the midcycle planning etherpad. 16:04:06 #undo 16:04:07 Removing item from minutes: 16:04:15 #link https://etherpad.openstack.org/p/newton-cinder-midcycle Midcycle planning 16:04:23 correct one :) ^^ 16:04:38 OK, that's all I have for announcements. 16:04:45 smcginnis: Hi, I don't see my company (reduxio) in your list - though we've commited a patch (driver) and are in the process of setting up the CI 16:05:08 rdxnoam: OK, I must have missed that. Please add a link to the driver list. 16:05:29 #topic Config categorization 16:05:37 #link http://paste.openstack.org/show/520980 Proposal 16:05:40 thrawn01: Hi 16:05:46 o/ 16:06:09 thrawn01: So this is something nova has done? 16:06:36 Or rather, is doing. 16:06:54 yes, part of this is in progress 16:06:55 link to previous review or something? 16:07:12 #link https://review.openstack.org/#/c/295543/9/specs/categorized-configuration-options.rst Nova spec 16:07:19 some of it is still waiting on oslo patches 16:07:47 i suspect this is problematic until we also sort out the handling of [DEFAULT] vs backend sections issues we've been discussing recently 16:07:51 smcginnis, thrawn01 is there actually code progress review link within nova? 16:07:55 actual* 16:08:02 besides the spec 16:08:20 eharney: +1 16:08:22 there are several moving parts to this spec 16:08:31 eharney: +1 ... 16:08:32 thrawn01: So it would be changes in oslo_config that would then require changes in each project? 16:08:34 eharney: +1 16:08:46 smcginnis, thrawn01 is there actually code progress review link within nova? 16:08:49 Can we figure out what we want to do there with this in mind and then tackle this? 16:08:59 eharney: seems overlapping with patrickeast's work 16:09:01 The reorder of the sample file is a feature that should be added to oslo 16:09:10 I think our config options need a lot of care and feeding. Going to need to do it little by little though. 16:09:24 jungleboyj: +1 16:09:24 geguileo: +1 16:09:38 jungleboyj: +1 16:09:44 thingee: I'm looking for the centralize config patch, which I think is already merged, but I'mn not great at finding this stuff 16:09:46 geguileo: +1 16:09:57 thrawn01: Is that one of the patches centralizing configs: https://review.openstack.org/#/c/258984/ ? 16:10:12 thrawn01: ok, if you're familiar with the code, maybe git blame would help 16:10:20 just a sha or something to look at previous work would be great 16:10:47 lol, I don't have the nova code checked out =( 16:11:07 dulek: looks like a piece of it. thanks 16:11:33 So IMO that's cool refactoring, but not a number one priority. 16:11:58 someone mentioned patrickeast is doing something similar? 16:12:06 thrawn01: I think some of this looks interesting. I would want to wait until the oslo changes. 16:12:09 Definitely we can benefit in looking through configuration options to see if we haven't any that don't make sense anymore. 16:12:28 xyang1: ^? 16:12:29 Option descriptions also need some care. 16:12:37 thingee: patrickeast is doing something with default group 16:12:44 dulek: All good points. 16:12:45 dulek: +1 16:12:45 patrickeast: around? 16:12:51 thingee: https://review.openstack.org/#/c/330767/ 16:12:58 smcginnis: ++ 16:12:59 geguileo: thank you thank you 16:13:06 Good for the Oslo work to get in first. 16:13:14 "Adopt Option Groups IE: 'common', 'driver', 'volume', 'solidfire'" - that's probably related to eharney's deprecation of not using enabled_backends. 16:13:30 smcginnis: Added the reduxio driver to the list as you requested. I have some questions on the process of final approval. Please ping me when you're ready as i don't want to interrupt the meeting flow 16:13:49 dulek: I think it's very similar, I don't really see option groups as a priority, just wanted some dissucsion around it 16:13:51 rdxnoam: Sure, will follow up later. 16:14:03 so glancing at the two different proposals it seems like patrickeast's is more focused on shared configs between backends 16:14:20 thrawn01: I found this particular interesting and seems useful from deployer point of view: https://blueprints.launchpad.net/oslo.config/+spec/option-interdependencies 16:14:21 thrawn01: Thanks for bringing it up. It helps that we are aware of the effort. 16:14:40 winston-d: Oh, good one. I hadn't seen that. 16:14:53 xyang1: not sure if they actually conflict. 16:15:05 thrawn01: Thanks for making us aware. :-) Good to be able to plan ahead. 16:15:10 xyang1: thrawn01's proposal seems to be share configs with respect to projects 16:15:25 am I understanding that right thrawn01 ? 16:15:58 not really *sharing* config options, mostly the organization of said options. 16:16:09 thingee: sure, I am just saying they are related 16:16:16 smcginnis: unfortunately, that blueprint is marked as obsolete 'Obsolete: We just use "related options:" in the normal help text' 16:16:23 xyang1: got it 16:16:26 the goal being, better orginzation and descriptions of options, and presentation of options to end users 16:16:36 winston-d: Yeah, noticed that after reading through. 16:16:42 yeah I'm board. agree with the dependencies on the oslo side maybe cuttin' it close 16:16:47 thrawn01: I like that as the end goal. 16:16:48 I'm on board* 16:17:03 thingee: And bored? :P 16:17:06 lol 16:17:27 especially if there's no code yet on the cinder side produced to depend on the oslo stuff 16:17:32 thrawn01: Cool, anything else on this for now? 16:18:01 I'm on board as well, but 'better orginzation and descriptions' is very high level goal and I'm wondering how we can technically achieve that 16:18:02 I think tackling the backend vs default issues is the higher priority right now. 16:18:07 maybe a priority of what I should focus on, out of the changes I was proposing. 16:18:13 jungleboyj: Agree 16:18:26 instead of depending on developers to be good with docs. 16:18:45 winston-d: We know that's never going to happen. :) 16:18:58 I'm refering to these proposals http://paste.openstack.org/show/520980/ 16:19:19 someone had mentioned _class options where planned for removal? 16:20:17 thrawn01: Not aware of that. Could be. 16:20:25 thrawn01: Is the oslo deprecated thing new? 16:20:29 yes 16:21:25 thrawn01: That one may be a good target to start with if it's available. 16:21:49 thrawn01: 'Prosoed change' session of the spec here: https://review.openstack.org/#/c/295543/9/specs/categorized-configuration-options.rst 16:21:52 k, will do, I'll wait on the other items. 16:22:16 thrawn01: i see only 1st bullet item is a technically achieveable goal. 16:22:30 the rest looks like still relies on people 16:23:00 thrawn01: Anything else to discuss here for now? 16:23:15 no, I'm good thanks all for the feedback 16:23:17 thrawn01: we can discuss offline, of course. 16:23:35 thrawn01: OK, thanks for raising it. 16:23:42 #topic Migrate volume between backends in an async way 16:23:48 wilson-l: Hi, you're up. 16:23:49 winston-d: thanks, I would like to pick your brain on "Improving the sample file" but thats about it 16:24:01 hi :-) 16:24:07 #link https://review.openstack.org/#/c/312853/ Spec 16:24:09 I have commit a patch for proof-of-concept: https://review.openstack.org/#/c/334923/ 16:24:21 thrawn01: I have done some work with improving the sample file as well if you wanted to chat about it. 16:24:22 #link https://review.openstack.org/#/c/334923 Proof of concept 16:24:25 not a full-blown patch, just a proof-of-concept 16:24:41 diablo_rojo: awesome 16:25:51 wilson-l: I need to read through the latest rev of the spec. 16:26:04 wilson-l: Are there any points you wanted to specifically discuss here? 16:26:28 not really much:-) 16:26:54 maybe we can revisit it again 16:27:19 wilson-l: OK, just raising awareness then. Since this has an impact on backends - driver maintainers, it's probably worth your time to review and see what this will mean to you. 16:27:41 wilson-l: OK, thank you. 16:27:48 wilson-l, do we have a spec ? 16:28:09 hemna: There is one out there. 16:28:10 heman: https://review.openstack.org/#/c/312853/ 16:28:10 we have discussed the spec ago 16:28:15 sweet 16:28:20 thnx 16:28:21 hemna: ^^ oops 16:28:50 #topic Open Discussion 16:28:55 Any other topics? 16:29:00 this sounds like something like the old c-data service discussion could solve 16:29:22 hemna: That topic does keep coming up. 16:29:27 yup 16:29:35 we really could use it for lots of stuffs 16:29:39 hemna: I think if a backend can handle it effeciently then that should be the first place. 16:29:42 * dulek think that he actually never been part of that. Just heard stories… 16:29:51 hemna: But having a data service to handle things like this could be good. 16:29:58 yah I think so 16:30:02 smcginnis: is that on the topic for midcycle? 16:30:07 man, if only I had a job, I would love to take that on 16:30:08 dulek: I think it's probably been a year or so since it was last discussed. 16:30:09 Does manila actual have something like c-data? 16:30:10 hemna: Yeah, that discussion as been dead for a while. :-( 16:30:17 thingee: Not sure, but it probably should be. 16:30:31 winston-d: yes, bur that is a generic solution 16:30:43 Right before we discuss taskflow again. :D 16:30:56 the proposal here is driver specific, not generic 16:31:11 * jungleboyj runs away from smcginnis screaming .... 16:31:19 winston-d: manila uses c-data for migration, I think since, Mitaka 16:31:20 :) 16:31:34 * dulek starts to feel good about joining mid-cycle only remotely… ;) 16:31:43 xyang1: right, wilson's proposal is driver specific but since hemna brought up c-data, that was supposed to be generic. 16:31:44 :) 16:31:50 dulek: Lol :) 16:32:15 winston-d: sure, that will be different from wilson's 16:32:36 xyang1: yeah. 16:32:37 Added to the etherpad for the midcycle. 16:32:42 dulek: ;) 16:32:56 👍 16:33:06 dulek: No fair muting us to avoid the fun. 16:33:48 Anything else? 16:34:11 OK, thanks eveyone. 16:34:19 Thanks! 16:34:23 Thanks :) 16:34:25 thanks 16:34:26 thx! 16:34:30 thanks 16:34:30 #endmeeting