15:02:36 #startmeeting scheduler 15:02:36 Meeting started Tue Oct 29 15:02:36 2013 UTC and is due to finish in 60 minutes. The chair is garyk. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:38 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:02:40 The meeting name has been set to 'scheduler' 15:03:01 the schedule is out for the summit 15:03:58 http://icehousedesignsummit.sched.org/ ? 15:04:24 Seems I don''t see the scheduler performance session. Boris didn't do that? 15:04:44 http://icehousedesignsummit.sched.org/overview/type/nova#.Um_OgiSInns 15:05:03 http://icehousedesignsummit.sched.org/event/cde73dadfd67eaae5bf98b90ba7de073 15:05:13 rethinking scheduler design - http://icehousedesignsummit.sched.org/event/cde73dadfd67eaae5bf98b90ba7de073 15:05:29 smartter placement for work loads - http://icehousedesignsummit.sched.org/event/57775e21d82cf716d435f02c9185db1e 15:05:43 metrics - http://icehousedesignsummit.sched.org/event/630d65c8879935f7bc085833f47bb537 15:05:56 instance groups api - http://icehousedesignsummit.sched.org/event/b23505a1c59fa2e9e5f94e9467b5a76a 15:06:17 russellb has been really generous and nearly the whole of wednesday is scheduling 15:07:07 he has also scheduled a number of unconference sessions 15:07:38 glad to see it 15:07:51 in a sec i will post the link - people can register for 10 minute slots 15:08:20 Already? Or when we get there? 15:08:22 are there specific unconference sessions for scheduler, 15:08:24 right. There are Nova specific unconferences so there should be a fair number of slots to bring up unscheduled topics 15:08:25 https://etherpad.openstack.org/p/NovaIcehouseSummitUnconference 15:08:33 or for Nova in general 15:08:39 feel free to sign up this week 15:08:47 Yathi: it is specifically for Nova. 15:09:10 and if others don't sign up, you can have more time 15:09:31 at the moment myself and a few colleagues have registered a topic on resource tracking 15:09:43 the one on tuesday might be better for wednesday now that scheduling sessions got moved to wednesday 15:10:34 russellb: correct - i'll move it to weds 15:10:37 ok 15:11:23 the sessions are going like hot cakes :) 15:12:49 are there any issues that we would like to discuss today? 15:14:12 Should we make the scheduler related unconference sessions on Wednesday as well ? I thought it will be overloaded for Wednesday hence submitted one for Tuesday 15:14:51 Yathi: that is what russellb suggested above 15:14:59 I think those of us advocating optimization approach would like it to be in the minds of those in the "rethinking" session 15:15:16 but "rethinking" comes before the Wed unconference. 15:15:18 regarding demo's there are also opnetsakc unconference sessions where you can register on the day and then get a slotted time 15:15:33 10 min also? 15:15:43 those sessions will also be to folks who may not be on the nova track 15:15:51 MikeSpreitzer: yes, i think so 15:16:29 MikeSpreitzer: in the past there was a board that people could write their topics. It is a good place to show demos etc. 15:17:12 Bird in the hand and all that 15:17:48 OK if I sign up for a Nova unconference until I get a general one? 15:18:36 sure, if you have a topic sign up. If you change later then someone else can use the slot 15:18:48 OK, that's my plan 15:19:24 i guess that if there is an overflow then the PTL may intervene and prioritize 15:19:52 of unconference sign-ups? yeah, if i have to 15:20:02 would rather just let it be open and first come first serve :) 15:20:19 i guess that if we are having scheduling discussion on wednesday then maybe we can all get together for drinks or lunch on tuesday - put faces to names 15:20:39 sounds good 15:20:46 russellb: cool. thanks for arranging everything. 15:20:48 have some initial discussions and plans before the session 15:21:00 Tuesday get together sounds good 15:22:16 so how about tuesday lunch? we can sync the meeting time in the morning 15:22:38 Huh? Lunch time is indicated in the schedule 15:23:35 i think 1 − 2 is free for lunch. so maybe we can meet next to the nova room at 1pm and then go and get a table together 15:23:58 http://icehousedesignsummit.sched.org/ says 12:45 — 1:20 15:24:53 so lets make it 12:50 at the nova room. 15:25:07 +1 15:25:14 that will give people a chance to attend lightening talks if they wish 15:26:12 do we have any other topics to talk about of should we all go to planning our etherpads? 15:26:47 you mean etherpads for the session 15:27:00 sorry it is my first summit.. 15:27:05 russellb: on question - in neutron there was talk of not having presentations but just using etherpads for the sessions. will this be the same for nova? 15:27:09 will need to work with you all 15:27:21 yes, please 15:27:52 Hmm, I was hoping to show templates, screen shots, etc 15:27:55 it should be primarily discussion, and only using etherpad seems to be working best 15:28:14 having some reference info available is ok ... but just know it's not supposed to be a presentation 15:28:31 it's not like materials are forbidden, just not typically used in the sessions that go the best 15:28:32 IMO 15:28:35 OK. But we're debating API, so concrete examples are really good 15:28:42 that's fine 15:28:44 How about always havign an Etherpad (its a great way to capture feedback, etc) - but allow a few slides to be shown if it helps illustrate a specific point 15:28:52 yes that makes sense 15:29:41 particularly if what you want to show has a URL that can go in the etherpad, I suppose 15:31:03 at times a picture/diagram is very important as it helps capture an idea. but we also need to have the dicussion and engage one another 15:31:26 right, that's the whole point 15:33:23 we have a few documents being written up with the ideas 15:33:35 they will be good reading material for the sessions 15:34:19 Yathi: i think that we should share the ideas before and then discuss things. the etherpads whould have the links to the documents. the etherpads will later be used as part of the BP's or tracking 15:34:23 Yes, I was confusing session with demo. For session, writeup beforehand is good. 15:34:58 reading beforehand is even better 15:35:04 garyk: But what I am worried is there might be a lot of suggestions overlapping views, etc. how to coordinate 15:35:10 and manage on a etherpad 15:35:48 Yathi: from past experience what has worked is a short introduction with diagrams and maybe flows and then discussion. 15:36:21 garyk: ok 15:37:05 garyk: for the Instance group API work - can I create a new blueprint - or use the old one ? we have newer model, newer apis, etc.. 15:37:59 Yathi: i think that we can use the existing BP. I think that we just need to make sure that it is updated to contain the latest details 15:38:29 Ok. I will add some links there.. with the latest doc, etc 15:38:36 Yathi: thanks! 15:39:14 i really think that it is important that we try and get things in in the first or maximum second milestone. that is the only way we may get this through in icehouse 15:39:35 timing is of the essence 15:41:26 the discussions I am seeing in the mailing list, everyone wants the whole thing to solved at once 15:41:32 i hope that we can convey the ideas in the session that we have. this will certainly help move it forwards 15:41:35 :) 15:41:37 but we should realize that we should start simple 15:41:49 I am thinking in a roadmap fashion 15:41:52 and that is the attempt to get it started within Nova 15:41:56 But I do not want to lose sight of the goal. 15:42:03 yup, i agree with the simple. we need it to be robust so that we can build on it 15:42:25 agreed. 15:42:26 MikeSpreitzer: it is a process and without stable foundations we will find it hard to get anything done. 15:42:37 garyk: agreed 15:43:12 i am hoping that we will get to speak about this all at the summit. it is the best place for people to voice their reservations and more importnatly they can express ideas which can help us achieve a common goal 15:43:38 at the summit, we should finalize what can be done in the icehouse time frame 15:43:46 agreed 15:43:53 some of the discussions may take a few more summits to implement 15:43:58 that being the primary purpose of a summit, right? 15:44:06 correct. 15:44:30 first we crawl, then walk and hopefully one day run 15:46:02 for the development cycle we also need to make sure that the patches can be reviewed - that is posting one huge patch is not a good idea. we need to do small and coherent iterations that can be viewed as pieces in the puzzle... 15:46:09 sorry for babbling on 15:46:35 garyk: please go on 15:46:37 it helps 15:46:45 for new comers 15:47:19 i hope that at the summit we can break down all of the pieces, assign them to interested parties and then try and sync our work along the way. 15:47:28 I think the existing Instance group work provides a model 15:47:36 it was broken down into several pieces 15:47:58 MikeSpreitzer: yup, that was the plan. hopefully we can continue with that model moving forwards 15:48:41 garyk: I am not able to update that blueprint. I would like to add details and change the owners 15:48:50 one thing i think is very importnat for the community is for us to try and let everyone know where we are in the cycle of development and what is current done, being done and what needs to be done 15:49:13 Yathi: i think russellb or a core reviewer may be able to change the owner 15:49:18 good point. Seeing the whole roadmap is good, not just what's been done so far 15:49:24 unless we get senhau to come and work with us again :) 15:49:48 :) 15:49:55 Senhua is not in our team anymore 15:50:24 yeah 15:50:27 A blueprint could include in its whiteboard a roadmap, right? 15:50:31 russelb: Can you please make me the drafter? and assignee, and let me edit details for this blueprint - https://blueprints.launchpad.net/nova/+spec/instance-group-api-extension 15:50:34 just tell me the blueprint and who to assign it to 15:50:47 anyone on the nova-drivers launchpad team has the necessary permissions 15:51:12 my full name - Yathiraj Udupi if you search for it 15:51:59 how to join nova-drivers launchpad team ? 15:52:13 Yathi: i think that you need to be nova core 15:52:30 that is kind of like running the comrades marathon and finishing in a top ten place :) 15:52:54 garyk: I am aiming high 15:52:58 :) 15:53:03 some day in the future I hope 15:53:48 anything else to be discussed? 15:54:38 I guess this week will not be very productive. preparing for the summit travel, etc. and wrapping up other things 15:56:00 ok. so looking forwards to having fruitful discussions. 15:56:02 I hope we can continue some discussion in the ML for a couple more days. 15:56:15 travel safe and see you guys on Tuesday 15:56:26 be seeing you 15:56:43 see you all at the summit .. 15:56:49 signing off until then 15:56:54 #endmeeting