18:00:34 <SergeyLukjanov> #startmeeting sahara 18:00:34 <openstack> Meeting started Thu Apr 16 18:00:34 2015 UTC and is due to finish in 60 minutes. The chair is SergeyLukjanov. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:38 <openstack> The meeting name has been set to 'sahara' 18:00:46 <huichun> hello 18:00:47 <SergeyLukjanov> #link https://wiki.openstack.org/wiki/Meetings/SaharaAgenda 18:00:50 <elmiko> yo/ 18:00:59 <SergeyLukjanov> #topic sahara@horizon status (crobertsrh, NikitaKonovalov) 18:01:12 <NikitaKonovalov> ok, so now the liberty is open 18:01:24 <SergeyLukjanov> #link https://etherpad.openstack.org/p/sahara-reviews-in-horizon 18:01:29 <NikitaKonovalov> our changes have got -2s removed 18:01:55 <NikitaKonovalov> not much reviews there yet, but I hope they'll come soon 18:02:18 <SergeyLukjanov> crobertsrh, any comment on sahara@horizon things? 18:02:19 <SergeyLukjanov> crobertsrh, hi 18:02:37 <NikitaKonovalov> I've update the event_log change to work with new Django 18:02:45 <crobertsrh> Not a ton at the moment. I need to tweak my template editing patches a bit and I'm going back fixing a few bugs that have been around awhile. 18:03:05 <crobertsrh> Have your UI ideas ready for summit! 18:03:39 <elmiko> javascript, javascript everywhere.... ;) 18:04:11 <SergeyLukjanov> :) 18:04:19 <SergeyLukjanov> #topic News / updates 18:04:22 <SergeyLukjanov> folks, please 18:04:49 <sreshetnyak> o/ 18:04:51 <vgridnev_> i'm working with deprecation of direct engine in sahara 18:04:53 * SergeyLukjanov now working on adding gates for images of all plugins and adding gate job to run fake-plugin based tests 18:05:02 <elmiko> i've been working on a possible bug in the domain proxy stuff, also looking into mapr config hints, and rewriting the external secret storage spec 18:05:02 <sreshetnyak> no updates from me 18:05:10 <elmiko> also writing more topics for summit =) 18:05:22 <tosky> SergeyLukjanov: fake-plugin tests in addition to full-cluster tests? 18:05:49 <SergeyLukjanov> tosky, fake plugin based in upstream gerrit (user Jenkins) 18:05:59 <SergeyLukjanov> tosky, sahara-ci will still running all plugins 18:06:02 <vgridnev_> it would be nice have reviews here 18:06:07 <vgridnev_> #link https://review.openstack.org/#/q/status:open+branch:master+topic:bp/deprecate-direct-engine,n,z 18:06:16 <tosky> SergeyLukjanov: ah, I see 18:06:16 <SergeyLukjanov> tosky, we're going to rework list of tests that sahara-ci is now running 18:06:27 <SergeyLukjanov> tosky, but it'll test all plugins for sure 18:07:07 <SergeyLukjanov> tosky, in a few words - we'd like to move testing of different topologies (direct vs. heat, nove-net vs. neutron) to upstream gate based on fake plugin 18:07:15 <elmiko> vgridnev_: ack 18:07:34 <SergeyLukjanov> tosky, and test heat-neutron-distributed in sahara-ci for latest versions of all plugins 18:07:52 <SergeyLukjanov> tosky, I will write down ideas and changes today and share it 18:08:19 <tmckay> hi folks, sorry I'm late 18:08:25 <tosky> SergeyLukjanov: thanks! 18:08:59 <tosky> oh, I have few questions about sahara-image-elements, I guess for later (no kilo branch yet? Should we backport the extjs fix to juno branch?) 18:09:04 <tosky> (Pino can do it) 18:09:29 <elmiko> tosky: nice, signing pino up for more work ;P 18:09:40 <SergeyLukjanov> tosky, it's a good question, probably yes, we need it 18:10:16 <SergeyLukjanov> I think there are no critical fixes for sahara itself 18:10:28 <tosky> SergeyLukjanov: and the kilo branch? Right now it's missing 18:10:43 <SergeyLukjanov> tosky, stable/kilo branch will be created at April 30 18:10:52 <SergeyLukjanov> with a release 18:11:09 <SergeyLukjanov> I will create proposed/kilo branch if we'll need to backport something 18:11:15 <SergeyLukjanov> and sounds like we need to do it 18:11:17 <tosky> SergeyLukjanov: ah, ok, so different than the other repositories where the branch is already there 18:11:30 <tosky> weren't proposed/ branches killed? 18:11:32 <tmckay> update for me -- currently running through a list of small (medium?) bugs and blueprint ideas for Liberty. Trying to register them all 18:11:38 <SergeyLukjanov> tosky, hm, do you have an example? 18:11:44 <tosky> SergeyLukjanov: I read an email about the process being changed a lot, let me find it 18:12:11 <SergeyLukjanov> tosky, oh, I got it 18:12:17 <SergeyLukjanov> tosky, yeah, I'm wrong 18:12:17 <tosky> SergeyLukjanov: for example https://github.com/openstack/cinder/branches 18:12:27 <SergeyLukjanov> tosky, yeah, yeah, I understant 18:12:29 <huichun> tmckay: we can help on these bugs and bps ;) 18:12:32 <tosky> oki :) 18:13:10 <tmckay> huichun, sure! :) Some of them are things that I just haven't had a chance to log. But, I figured now was the time to list them all as we plan ahead 18:13:34 <tmckay> huichun, I would love to see EDP (and Sahara in general) come out of Liberty very polished 18:14:06 <tmckay> get rid of the little things around the edges, in addition to adding "cool new features" 18:14:08 <SergeyLukjanov> tosky, I'll create stable branches after the meeting 18:15:59 <SergeyLukjanov> okay, let's move on 18:16:18 <SergeyLukjanov> #topic Design summit 18:16:49 <SergeyLukjanov> so, we have 2-5-2 slots on summit 18:17:00 <SergeyLukjanov> 2 big sessions, 5 small and the whole day on Friday 18:17:09 <huichun_> tmckay: we can talk offline :) 18:17:17 <tmckay> huichun_, sure 18:17:31 <SergeyLukjanov> folks, please share your thoughts - is it enough or too much in your opinion? 18:17:55 <elmiko> i thought it seems kinda small, but given the state of the etherpad, maybe not 18:18:06 <tmckay> SergeyLukjanov, re sessions, we have had an EDP planning session for the last few summits but I think given the other topics on the table, we can just carve out some time on Friday to prioritize EDP goals for Liberty. 18:18:11 <elmiko> #link https://etherpad.openstack.org/p/sahara-liberty-proposed-sessions 18:18:38 <SergeyLukjanov> tmckay, ++ 18:19:09 <elmiko> i'm guessing that ui and security are our fishbowls? 18:19:11 <tmckay> SergeyLukjanov, we also might have some big topics from the intel guys -- scheduling, coordinated job manager. But, no proposals submitted yet (bps) 18:19:21 <SergeyLukjanov> elmiko, if we'll not have more proposals 18:19:24 <tmckay> but I think Friday is good for that stuff 18:19:37 <SergeyLukjanov> tmckay, yup 18:19:53 <tmckay> elmiko, I think UI and security are good topics. Ease of use, and secure secure secure 18:20:06 <tmckay> this imho is what people want ;-) 18:20:09 <SergeyLukjanov> currently we should understand that it's not too much for us and if it isn't - than just feel it with our topics 18:20:10 <elmiko> i'm just going based on whats in the pad currently 18:20:41 <weiting> tmckay, we can add the proposals to liberty session. 18:20:49 <elmiko> SergeyLukjanov: given what we have proposed, it seems appropriate (2-5-2) 18:21:33 <tmckay> weiting, okay. SergeyLukjanov, weiting might have proposals for other working sessions? .... 18:21:45 <jvrbanac> exit 18:22:17 <SergeyLukjanov> weiting, it'll be great if you add proposals to the etherpad 18:22:26 <elmiko> +1 18:22:36 * SergeyLukjanov should dump thoughts to etherpad as well :) I have ~ 5-7 items 18:23:04 <elmiko> oh nice, maybe we'll need more sprint sessions then? 18:23:19 <elmiko> it sounds like space is at a premium for this summit 18:23:46 <tmckay> we could adjourn to a local restaurant :) 18:24:23 <SergeyLukjanov> :) 18:24:46 <elmiko> would be fun to do a meetup outside the summit 18:24:51 <SergeyLukjanov> I think that we all should dump all ideas to etherpad and on next meeting try to align it to 2-5-2 slots 18:25:12 <elmiko> sounds good to me 18:25:42 <SergeyLukjanov> because at least some topics are not for the full slot and some could be or should be moved to the meetup 18:26:00 <elmiko> would also be nice if people could create separate pads for the larger topics to help refine the focus 18:26:01 <SergeyLukjanov> okay, sounds like everything is ok with summit 18:26:27 <SergeyLukjanov> who will go to summit? 18:26:47 <SergeyLukjanov> from Mirantis - /me, alazarev and aignatov will be their this time 18:26:47 <elmiko> o/ 18:27:27 <tmckay> me 18:27:33 * tosky not this time 18:27:36 <tmckay> crobertsrh 18:27:43 <elmiko> not sure if egafford is going 18:27:47 <tmckay> mattf I believe 18:27:51 <SergeyLukjanov> will mattf join us this time? 18:27:53 <elmiko> yea 18:27:55 <SergeyLukjanov> tmckay, ok ;) 18:28:10 * crobertsrh will be there 18:28:46 <weiting> We got another topic in liberty session called "sahara-versionedobject". 18:29:01 <SergeyLukjanov> #topic Open discussion 18:29:06 <weiting> Ken just added a sample code, please help to review it. https://review.openstack.org/#/c/173654/ 18:30:28 <tmckay> I was looking at that this morning, looks like a lot of projects are going this way 18:30:42 <tmckay> I intend to study it more 18:32:54 <tmckay> I am unclear though how using versioned objects help migration. I'll have to read up on it. The alembic migrations work pretty well as it is. But I suppose it can help with backwards compatibility 18:34:12 <tmckay> oh, if we intend to ever implement distributed conductor (local vs remote), maybe we should do it in Liberty :) 18:34:29 <elmiko> tmckay: maybe the bp has a little more info, #link https://blueprints.launchpad.net/sahara/+spec/versioned-objects 18:34:32 <tmckay> just a thought. The Local/remote API classes have been sitting there for quite a while 18:35:00 <tmckay> elmiko, yeah, I took a look at that, I want to dig through the oslo lib stuff too 18:35:05 <tmckay> sounds interesting 18:35:15 <elmiko> tmckay: i'd say the distributed conductor is an excellent topic for summit ;) 18:35:35 <tmckay> ack. Implement it, or state why it should be abandoned 18:35:45 <tmckay> I'll add it to the pad 18:35:50 <elmiko> at least we can have more discussion about it 18:36:04 <elmiko> i have to admit, i could learn more about the reasoning for using it 18:36:29 <tmckay> "By using objects, the code will be insulated from the actual database schema, making it easier rolling upgrades." See, my contention is that with sqlalchemy and alembic, we' 18:36:37 <SergeyLukjanov> IMO we don't need distributed conductor 18:36:38 <tmckay> we're already insulated enough 18:36:47 <SergeyLukjanov> we're not going to run sahara-engine on computes 18:36:48 <elmiko> tmckay: oh, i meant the conductor 18:37:12 <elmiko> as for versioned objects, i'm not so sure on the wider implications 18:37:14 <tmckay> elmiko, I know, mixing thoughts :) 18:37:18 <elmiko> hehe 18:37:23 <elmiko> stop confusing me! ;) 18:37:24 <tmckay> I was just explaining my stance for wanting to dig more 18:38:13 <tmckay> how much more insulation do we need? is my question. Another layer could be nice, but it could be just another layer 18:38:22 <tmckay> to dig through 18:38:25 <tmckay> and change 18:38:49 * tmckay has late adopter hat on today ;-) 18:38:55 <elmiko> lol 18:39:48 <tmckay> hmm, rolling upgrades, maybe as in "hot upgrades". That might be the difference 18:40:03 <tmckay> don't take the db down, leave Sahara running. That could be cool. 18:40:16 <tmckay> like I said, time to read more 18:40:33 <tmckay> and stop armchair quarterbacking 18:40:39 <tosky> live upgrade to major version when the rest of openstack falls apart in that scenario? :D 18:40:39 <SergeyLukjanov> tmckay, ++ 18:41:11 <tmckay> tosky, as long as Sahara works, we've done our part ;-) 18:41:15 <tosky> (nothing to do with HA for sahara-api and -engine, right? I guess I will have to read a bit about this) 18:41:38 <SergeyLukjanov> IMO one of the biggest discussions for this summit will be the HA 18:41:46 <SergeyLukjanov> (after API v2 sure) 18:42:21 <tmckay> We need API v2 T-shirts "V2, just do it!" hmm, copyright? 18:42:26 <elmiko> lol 18:42:34 <tosky> ... 18:42:47 <elmiko> i'm curious to hear more about HA 18:43:14 <tmckay> HA and security are big, I agree. True enterprise stuff. 18:43:24 <tosky> I'm not so curious to hear more, but I will take what will be implemented :) 18:43:37 <elmiko> nice tosky 18:43:38 <SergeyLukjanov> :) 18:44:26 <SergeyLukjanov> BTW alazarev is working on HDP 2.2 support and sreshetnyak working on Hadoop HA for CDH and HDP 18:44:37 <elmiko> nice 18:44:58 <tosky> oh 18:45:00 <elmiko> speaking of HDP, what's up with the hwx folks? 18:45:04 <tmckay> SergeyLukjanov, you mean alazarev is adding HDP 2.2 version to the HDP plugin? 18:45:04 <tosky> of course this is Liberty 18:45:16 <SergeyLukjanov> tmckay, yup 18:45:27 <SergeyLukjanov> elmiko, they aren't intended to support hdp plugin 18:45:36 <tosky> but given the timeframe, doesn't it make sense to look at the upcoming HDP 2.3 later this year? (not sure about the date) 18:45:40 <elmiko> ahh, too bad =( 18:45:54 <tmckay> SergeyLukjanov, I should talk to him. I looked at the resource files a bit -- I believe they were sniffed from the REST api on a runnig/installing cluster 18:45:58 <SergeyLukjanov> tosky, we already have a working version with base functionality 18:46:01 <tmckay> not maintainable 18:46:26 <tmckay> maybe we should try to use blueprints, instead of the resource-style files from 2.0.6 18:46:34 <SergeyLukjanov> tmckay, I think alazarev think the same way :) 18:46:35 <huichun_> SergeyLukjanov: hi Serygey, is there any bps related to the HA for CDH and HDP? 18:46:57 <SergeyLukjanov> huichun_, not yet I think, sreshetnyak should create some soon 18:47:21 <SergeyLukjanov> tmckay, it'll mean to write plugin from scratch 18:47:47 <huichun_> SergeyLukjanov: ok ,cause we just raise the same bp today, so sreshetnyak is working on that now ? 18:48:19 <tmckay> SergeyLukjanov, yes, I could see that. But I think going forward it would be much easier to maintain 18:49:23 <SergeyLukjanov> huichun_, yeah, he's working on HA for CDH already, so, I think you should discuss it with him 18:49:33 <SergeyLukjanov> huichun_, have you already started working on it too? 18:49:56 <elmiko> SergeyLukjanov: are there specs up for any of this work? 18:50:13 <huichun_> SergeyLukjanov: we are currently not, just raised the bp today, but not start working 18:50:39 <SergeyLukjanov> tmckay, in theory - yes, but alazarev said that exported blueprint from working cluster failing to create the same cluster :) 18:51:08 <tmckay> bah. Okay, but that's a bug in Ambari 18:51:29 <sreshetnyak> huichun_: i'm not start working on implementation HA for CDH 18:51:34 <SergeyLukjanov> huichun_, elmiko, sreshetnyak is now investigating the right way to do it and experimenting with CDH to write a better spec 18:51:35 <sreshetnyak> only research 18:51:43 <SergeyLukjanov> sreshetnyak, hey :) 18:52:08 <elmiko> ok, ack. just thinking about better communication for work 18:52:27 <huichun_> sreshetnyak: ok, we can talk this topic detail offline 18:52:41 <SergeyLukjanov> huichun_, sreshetnyak, yeah, please discuss it 18:52:47 <sreshetnyak> huichun_: ok 18:52:57 <SergeyLukjanov> elmiko, probably we all should try to share plans in a single place 18:53:06 <SergeyLukjanov> like etherpad for internal plans / roadmaps 18:53:18 <elmiko> SergeyLukjanov: +1 great idea 18:53:24 <tmckay> +! 18:53:26 <SergeyLukjanov> w/o many details, just to now that we're not doing the same :) 18:53:28 <tmckay> heh 18:53:49 <elmiko> SergeyLukjanov: agreed, that's what i was thinking 18:53:55 <SergeyLukjanov> okay, I'll create an etherpad today and share it with overview of mirantis team plans for liberty 18:54:01 <elmiko> maybe something on the roadmap page for the wiki? 18:54:08 <elmiko> ack 18:54:12 <SergeyLukjanov> elmiko, it's too official 18:54:17 <elmiko> ok 18:54:29 <SergeyLukjanov> team's plans aren't so official, could be changed and not accepted :) 18:54:34 <elmiko> yea, for sure 18:54:46 <SergeyLukjanov> but probably we should resurrect the roadmap page 18:54:55 <elmiko> +1 18:55:03 <SergeyLukjanov> because sometimes people ask me about it and I'm answering oops ;) 18:55:08 <elmiko> hehe 18:55:18 <SergeyLukjanov> the good topic for summit's friday 18:55:23 <elmiko> i'm all for increased transparency with regards to our plans, even if they are unofficial 18:55:34 <SergeyLukjanov> or probably it could be highlights from team's roadmaps 18:55:43 <SergeyLukjanov> elmiko, ++ 18:56:16 <SergeyLukjanov> elmiko, tmckay, crobertsrh, do you have some kind of internal plan / roadmap ? 18:56:17 <tmckay> SergeyLukjanov, can alazarev push HDP 2.2 base patch as WIP? I'm interested in it 18:56:28 <elmiko> SergeyLukjanov: we'll be working on it in the next week or so 18:56:30 <crobertsrh> we're working on it :) 18:56:59 <SergeyLukjanov> cool, that'll be great to sync our plans and understand how to achieve the max results 18:57:09 <elmiko> agreed 18:57:27 <tmckay> SergeyLukjanov, yes, next week or two we'll have something. 18:57:31 <SergeyLukjanov> tmckay, it's currently done for Juno :) so, need to be ported to master 18:57:40 <SergeyLukjanov> tmckay, that's really great 18:57:40 <tmckay> gotcha 18:57:59 <SergeyLukjanov> for my team we already have it and I'll dump it to etherpad 18:58:06 <tmckay> oh, elmiko already said that, sorry :) 18:58:35 <SergeyLukjanov> few mins left 18:59:20 <tmckay> nothing left for me, have to run 18:59:27 <SergeyLukjanov> thank you folks, have a good day / night! 18:59:32 <SergeyLukjanov> #endmeeting