16:02:42 <xarses> #startmeeting fuel 16:02:43 <openstack> Meeting started Thu Jun 18 16:02:42 2015 UTC and is due to finish in 60 minutes. The chair is xarses. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:02:44 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:02:47 <openstack> The meeting name has been set to 'fuel' 16:02:53 <xarses> #chair xarses 16:02:53 <openstack> Current chairs: xarses 16:03:07 <xarses> hi all 16:03:12 <kozhukalov_> hi 16:03:24 <ogelbukh> hi 16:03:27 <akasatkin_> hi 16:03:40 <mattymo> hi 16:03:41 <xarses> do you want me to lead kozhukalov_ ? 16:04:06 <kozhukalov_> xarses: yes 16:04:23 <kozhukalov_> xarses: if you have time of course 16:04:26 <xarses> #chair xarses 16:04:26 <openstack> Current chairs: xarses 16:04:41 <xarses> ok we have a short list schedule here, 16:04:46 <xarses> #topic flexible(advanced) networking feature status (akasatkin) 16:04:56 <akasatkin_> spec:https://review.openstack.org/115340 All tasks which do not fit into 7.0 are moved to other tickets. 16:05:05 <akasatkin_> As scope is still big, there is a certain risk that full functionality will not be implemented and tested in time. We have a backup plan and we are going to run it in parallel with the basic one. Design of a backup plan is in progress. 16:05:13 <akasatkin_> Even if we do not land everything into master for the basic plan, we will have changes in library part (apparently, Neutron ML2(both vlan/tun in one env) and at least several separated network roles) and corresponding changes in backend and UI. 16:05:30 <akasatkin_> Basic plan will be implemented incrementally and should allow to not break current functionality in the middle of the developmen cycle. Development is started. We have done a number of tasks in nailgun already which allow library team to implement and test their part. 16:05:54 <xarses> So what can we do to break it into phases so that we can make progress in 7.0? 16:06:12 <xarses> Has the spec been updated to reflect this? 16:06:59 <akasatkin_> yes. now spec reflects nailgun tasks only 16:07:47 <akasatkin_> we can make progress in 7.0 anyway 16:08:32 <xarses> yes, but do we have seperate sub-topics that we have identified and a action plan of what we think should make 7.0? 16:08:33 <akasatkin_> and it will allow to continue in 7.1 if it will not be finished in 7.0 16:09:14 <akasatkin_> yes. we have a backup plan as well 16:09:35 <akasatkin_> it will costs additional resources though 16:09:59 <xarses> ok, I ask because it still looks very big and at risk to over run the schedule 16:10:48 <xarses> OK, thanks moving on then? 16:10:56 <kozhukalov_> are we going to move network stuff into a separate service somewhen? 16:11:49 <xarses> kozhukalov_: I think its a future goal that we might start to address in 8.0 16:12:01 <akasatkin_> we have no strict plan on that 16:12:08 <ogelbukh> should we think about rather adopting Neutron for this? ;) 16:12:16 <kozhukalov_> i mean something like, when nailgun will call kinda API to allocate networks 16:12:58 <ogelbukh> from what I read in spec, it looks pretty much like we should 16:13:03 <akasatkin_> it looks more as a plugin for me now 16:13:10 <kozhukalov_> ogelbukh: yes, it is exactly what i have in my mind when talking about this 16:13:44 <xarses> ogelbukh: its about setting up underlay network though 16:14:01 <ogelbukh> xarses: right, but still 16:14:18 <ogelbukh> xarses: we can use it with custom plugin if we need 16:14:39 <xarses> it's an interesting point, I'm not sure it would work reliably enough. It needs more thought 16:14:39 <kozhukalov_> the same is about volume manager 16:14:45 <akasatkin_> there are a number of entities there: networks, network roles, nics, 16:14:48 <ogelbukh> kozhukalov_: + 16:15:16 <kozhukalov_> it would be great to deal with volumes via a separate service or at lease separate python module 16:15:28 <ogelbukh> akasatkin_: neutron has it all as well 16:15:38 <akasatkin_> yep 16:15:55 <ogelbukh> and a community to support it 16:16:02 <ogelbukh> which we don't 16:16:27 <ogelbukh> well, that's just a thought 16:16:52 <ogelbukh> but I'd like to plant it :) 16:16:59 <kozhukalov_> for volumes we probably could adopt cinder, but i am not sure about this 16:17:04 <akasatkin_> let's discuss it separately then 16:17:11 <ogelbukh> sure thing 16:17:34 <xarses> ogelbukh: it would be good to get something written down on the ML about how it might work 16:17:37 <ogelbukh> kozhukalov_: I'm not sure cinder is good for managing what we have 16:18:04 <ogelbukh> xarses: I'll see what I can do here 16:18:27 <xarses> thanks, It's and interesting topic and would need more details 16:18:29 <ogelbukh> I think we should move on though 16:18:41 <xarses> #topic open-discuss 16:19:05 <xarses> thats the end of our topics scheduled. 16:19:16 <xarses> does anyone have any other specs or bugs to discuss 16:19:17 <ogelbukh> looks like we already had it :) 16:19:34 <ogelbukh> well, I've got 2 upgrade-related specs 16:19:46 <ogelbukh> targeted to 7.0 16:21:05 <xarses> please continue 16:21:14 <ogelbukh> #link https://review.openstack.org/#/c/185136/ 16:21:56 <ogelbukh> #link https://review.openstack.org/#/c/192551/ 16:22:17 <ogelbukh> first one is an external logic that does upgrade openstack env 16:22:20 <mattymo> looks like the bot doesn't want to behave on links 16:22:34 <ogelbukh> that's because I'm not a chair, I guess 16:22:51 <xarses> shouln't matter =( 16:22:58 <xarses> shouldn't even 16:23:11 <ogelbukh> second one is an extension to nailgun API to facilitate upgrade of cluster settings 16:23:41 <ogelbukh> it's rather modest at the moment 16:24:16 <ogelbukh> as it should talk more about proper versioning of cluster object and stuff 16:25:05 <ogelbukh> but we are rather limited on dev resources 16:25:33 <ogelbukh> so I don't want to bite more than we can chew 16:25:46 <kozhukalov_> ogelbukh: thanks for links, interesting 16:26:12 <ogelbukh> kozhukalov_: I would much appreciate any feedback on them 16:26:34 <kozhukalov_> ogelbukh: yes, will have a look at them 16:26:37 <xarses> So like with advanced networking, lets identify what parts we can Implement ad increments and ensure that we can get partial progress 16:26:41 <ogelbukh> thanks! 16:26:48 <ogelbukh> xarses: that 16:26:57 <ogelbukh> is exactly what I tried to do 16:27:15 <xarses> thanks! 16:27:24 <xarses> any other topics to discuss? 16:27:31 <ogelbukh> thus the nailgun api proposal is very limited in scope 16:27:45 <ogelbukh> not from my side, thank you 16:27:59 <xarses> anyone else? 16:28:08 <kozhukalov_> not from me as well 16:28:21 <xarses> ok, I will close the meeting thanks! 16:28:28 <xarses> #endmeeting