Wednesday, 2017-11-08

*** jkilpatr has quit IRC02:21
*** dims has quit IRC06:05
*** dims has joined #openstack-cyborg06:12
*** mpaolino has joined #openstack-cyborg08:26
*** zhipengh[m] has quit IRC09:43
*** ttk2[m] has quit IRC09:47
*** ttk2[m] has joined #openstack-cyborg09:57
*** zhipengh[m] has joined #openstack-cyborg10:04
*** crushil_ has joined #openstack-cyborg10:37
*** crushil has quit IRC10:40
*** crushil__ has joined #openstack-cyborg10:57
*** crushil_ has quit IRC11:00
*** crushil_ has joined #openstack-cyborg11:04
*** crushil__ has quit IRC11:07
*** crushil__ has joined #openstack-cyborg11:15
*** crushil_ has quit IRC11:18
*** crushil has joined #openstack-cyborg11:25
*** crushil__ has quit IRC11:27
*** crushil_ has joined #openstack-cyborg11:31
*** crushil has quit IRC11:34
*** crushil has joined #openstack-cyborg12:19
*** crushil_ has quit IRC12:22
*** mikeH has joined #openstack-cyborg12:29
*** jkilpatr has joined #openstack-cyborg12:32
*** crushil has quit IRC12:37
*** crushil has joined #openstack-cyborg14:07
*** zhipeng has joined #openstack-cyborg14:17
*** crushil has quit IRC14:25
*** zhipeng has quit IRC14:28
*** zhipeng has joined #openstack-cyborg14:33
*** crushil has joined #openstack-cyborg14:39
zhipeng#startmeeting openstack-cyborg15:01
openstackMeeting started Wed Nov  8 15:01:50 2017 UTC and is due to finish in 60 minutes.  The chair is zhipeng. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:01
*** openstack changes topic to " (Meeting topic: openstack-cyborg)"15:01
openstackThe meeting name has been set to 'openstack_cyborg'15:01
zhipeng#topic Roll Call15:02
*** openstack changes topic to "Roll Call (Meeting topic: openstack-cyborg)"15:02
zhipeng#info Howard15:02
mpaolino#info Michele15:02
zhipenghey Michele15:04
mpaolinohi Howard!15:04
zhipenglet's wait to see if there are other folks coming15:06
zhipengcrushil ttk2[m] zhuli you guys there ?15:06
jkilpatrdaylight savings time tho15:07
jkilpatrwe're an hour early by my calender15:07
zhulihi folks, sorry for the late15:07
jkilpatror is my calender screwed up15:07
crushilIt's DST jkilpatr15:07
jkilpatrno biggie15:07
jkilpatrwe can just do a 2 hour meeting and then anyone who wants to show up at the other time can :P15:07
jkilpatr#info Justin15:08
jkilpatrafter all we don't want to make Howard have to stay up any later than he is now.15:09
zhipengwe want the bot to rest as well15:09
jkilpatrlol bots rights15:10
crushil#info RUshil15:10
jkilpatrwe've got another 5 years before we have to worry about the meeting bots unionizing zhipeng15:10
zhipengjkilpatr if they do we just close business15:10
zhipeng#topic Sydney Summit Initial Feedback Gathering15:10
*** openstack changes topic to "Sydney Summit Initial Feedback Gathering (Meeting topic: openstack-cyborg)"15:10
zhipengSo I think I'm very thankful for Richard to step in for us to do the presentation15:11
zhipenghe even got the joke right (about our meeting time)15:11
zhipengfor those of you haven't seen the recording15:11
jkilpatrhow was attendance?15:12
zhuliawesome guy, is there a video about the presentation, would love to see it15:12
zhipengand for presentation slide itself15:12
zhipengcrushil I was gonna to ask you, any feedback from Rich >15:13
crushilIt went really well15:13
zhipengabout the attendance, etc ?15:13
zhulithanks for the link, will watch later15:14
crushilIt was a lighting talk15:14
crushilSo, there were like 15-20 folks there15:14
zhipenganyone approached him after the talk ?15:14
crushilObviously there was no time for questions15:14
crushilThat we didn't get a chance to talk about15:14
crushilIt was around midnight when he finished15:15
zhipengah ok15:15
crushilI'll catch up with him my evening (his morning)15:15
zhipengsounds great :)15:15
crushilSydney is 15 hours ahead of EST15:15
crushilIt's like 3 hours ahead of you15:15
zhipengye we have 3 hour difference15:15
zhipenglet's try to get together in Dublin next year, for rocky ptg15:16
zhipenglate Feb15:16
jkilpatrI figure I'll get sent to the PTG, to be honest we get more work done there anyways15:17
zhipengmight freeze my balls off tho, considering the weather ...15:17
zhipengjkilpatr for sure15:17
zhipengany other comments re this topic ?15:17
jkilpatrspeaking of getting work done, can we try and wrap up the interaction patch? Sorry I haven't gotten my nova hack ready been on PTO15:18
zhipengso on that note, moving on to the next topic15:19
zhipeng#topic Placement-Cyborg Interaction Discussion15:19
*** openstack changes topic to "Placement-Cyborg Interaction Discussion (Meeting topic: openstack-cyborg)"15:19
zhipengso last week crushil made a really good proposal15:19
zhipengthat before Xmas we should have like weekly video sprints15:19
zhipengi asked Sean about the release sched15:19
zhipengFeb 5th 2018 is the hard cutoff date for coding15:20
zhipengthat will be when RC is released15:20
jkilpatrI can refocus on Cyborg last week of November and into December15:20
jkilpatrWrapping up a lot of other projects in the next few weeks.15:20
zhipengso I think let's just concentrate on Nov and Dec and try to hammer everything we need out in this period15:20
zhipengjkilpatr no problem15:20
zhipengthen when would be the dates works for everyone about the sprint ?15:21
zhipengcrushil jkilpatr zhuli15:22
jkilpatrNov 27 - Dec 4th short break for Ansible training then Dec 8 - EOY and probably beyond.15:23
jkilpatrI can attend meetings and maybe get some stuff done earlier in November but I'15:24
jkilpatrI'm getting a few racks of gear for release qualification so I need to focus on that when I have it.15:25
zhipenghow about you crushil ?15:25
crushilI should be ok to start sprint next week. Still wrapping up one project that has been pending for months15:26
zhipengthen I think zhuli and I could start with crushil first, starting next week15:28
zhipengon driver side and traits15:28
zhulisince the summit is done, next week is ok for me too15:28
zhipengand jkilpatr could jam in later on the agent and conductor parts15:28
zhipengthen maybe EOY we could have some E2E stuff, hopefully15:28
jkilpatrbeen looking at Nova don't see any reason why we can't add devices to the whitelist live15:28
jkilpatrthere's actually not too much going on with the whitelist at all, just gets loaded and checked15:29
zhipengjkilpatr I could also ping Matt if we encounter any problems at Nova side15:29
jkilpatreasier to write code and ask for a review than ask them to do it :P15:29
jkilpatrespecially if it's short and easy to review15:29
zhipengyep :P15:30
zhipengthen crushil which day of the week shall we pick ?15:30
zhipengI think time wise the meeting time would be ideal15:30
jkilpatrI agree15:31
zhipengwe could do like the same time but on Tuesday, for example15:32
jkilpatrthat's fine too but then what do we talk about during the Wednesday meetings?15:32
crushilya, that works15:33
zhipengjkilpatr to get you update :P15:34
zhipengor shall we just use the weekly meeting time ?15:34
jkilpatrI think we should just use the weekly meeting time and be sure to note things in IRC don't want to turn away newcomers in the channel15:35
zhipengcrushil that ok for you ?15:35
zhipengokey great, then I will send out ZOOM links for the video conf15:37
zhipengstarting next week15:37
zhipengjkilpatr regarding your current spec, I will have another set of comment, but I think we could wrap it up this week15:37
jkilpatrsounds good let me know15:38
jkilpatrzhuli, these times good for you?15:39
zhulino problem15:39
zhipengjkilpatr last week meeting, don't know if you have comb through the log yet, I have thought about adding another dimension of resource provider15:40
jkilpatrnope haven't read it, care to summarize?15:41
zhipengwe could implement it later, but I think I should submit a spec about it15:41
zhipengso in a gist15:41
jkilpatrlink me when you have it up15:41
zhipengcurrently Placement uses Resource Provider to model resources15:41
zhipengwhich consists of two dimensions: traits on the quantative side of things, and resource class for the qualitive side of things15:42
zhipenge.g. you could use trait to define what you have and resource class for how much/many you have15:42
zhipenghowever, i think for cyborg to deal with accelerator15:43
zhipengpurely enumeration would cause problems15:43
zhipengsince for example for FPGAs15:43
zhipengyou can't trait and resource class with every VF/PF combination/layout15:43
jkilpatrwhere you have possible traits that consume the same resource class?15:43
jkilpatrzhipeng, I would have a resource class fpga that was consumed by many traits which would be the VF/PF15:44
zhipengjkilpatr my point is that many traits would be really hard to manage15:44
zhipengwe need a more fuzzy way to describe accelerator resource, besides traits and resource class15:45
jkilpatrwhat do you propose15:45
zhipengso I think propably something like capability, that it could generally map to a set of resource that user would need15:45
zhipenginstead of specifically id the resource15:45
zhipengfor example for deep learning15:46
zhipengthere will be cases that operators just want to run their framworks on OpenStack Infra15:46
zhipengdoesn't really care about which GPU, FPGA, SoC solution it should use15:46
zhipengand then if say, the infra got some GPU good for temporalachitecture15:47
zhipengand some FPGA cards that good for spatial architecture15:47
jkilpatrbut we need to get more specific at some point?15:48
zhipengwe should just label these like CAPABILITY: DL/TEM15:48
zhipengwe will still have traits ands resource class15:48
jkilpatrthe way I see it we can just map a buncha traits to the same resource and it will be kinda ugly but work.15:48
zhipengye that is what I think15:48
zhipengso my thought was that if we could add another dimension, to use with trait and resource class together15:48
zhipengso that operators could do a specific request15:49
zhipengor a fuzzy request15:49
zhipengbecause unlike the general computing infras, which could almost be enumerated15:49
zhipengaccelerators just got so many types and combinations15:49
jkilpatr so we let operators aggregate things when their workload could run on either and schedule agains that?15:50
zhipengnot necessarily aggreagating, but just select a (super/sub) set of resource that satisfy the fuzzy request15:51
zhipengif it deems to be a super set, then it is aggregation15:51
zhipengif it just needs a subset, say just some of the FPGA card funcationalities15:51
zhipengwe could also do that mitigation15:51
jkilpatrstill not sure I understand the usecase15:52
zhipengI think I will just writeup a spec and we could review it :P15:52
jkilpatrsure, a better outline of the potential usecase would help a ton15:52
zhipengI need to elaborate more15:52
zhipeng#agreed have weekly sprint video conf using team meeting timeslot starting Nov 15th15:53
zhipeng#topic AoB15:53
*** openstack changes topic to "AoB (Meeting topic: openstack-cyborg)"15:53
zhipengI think we have covered the most important ones15:54
zhipengany other business ?15:54
zhipengokey let's conclude the meeting then, thanks everyone :)15:58
*** openstack changes topic to "#openstack-cyborg"15:58
openstackMeeting ended Wed Nov  8 15:58:59 2017 UTC.  Information about MeetBot at . (v 0.1.4)15:59
openstackMinutes (text):
*** zhipeng has quit IRC16:00
*** crushil has quit IRC16:02
*** crushil has joined #openstack-cyborg16:06
*** mpaolino has quit IRC16:09
*** jkilpatr has quit IRC16:41
*** jkilpatr has joined #openstack-cyborg16:55
*** jkilpatr has quit IRC17:05
*** jkilpatr has joined #openstack-cyborg17:19
*** jkilpatr has quit IRC18:05
*** mikeH has quit IRC18:17
*** jkilpatr has joined #openstack-cyborg18:18
*** amol has joined #openstack-cyborg18:52
*** amol has quit IRC18:53
*** jkilpatr has quit IRC19:00
*** jkilpatr has joined #openstack-cyborg19:13
*** crushil has quit IRC19:20
*** mikeH has joined #openstack-cyborg19:27
*** crushil has joined #openstack-cyborg19:39
*** crushil_ has joined #openstack-cyborg20:15
*** crushil has quit IRC20:16
*** crushil_ has quit IRC20:17
*** crushil_ has joined #openstack-cyborg20:17
*** crushil_ has quit IRC20:18
*** crushil_ has joined #openstack-cyborg20:19
*** crushil has joined #openstack-cyborg20:23
*** crushil_ has quit IRC20:25
*** jkilpatr_ has joined #openstack-cyborg20:32
*** jkilpatr has quit IRC20:35
*** mikeH has quit IRC20:38
*** jkilpatr_ has quit IRC21:32
*** crushil has quit IRC21:47
*** jkilpatr has joined #openstack-cyborg21:57

Generated by 2.15.3 by Marius Gedminas - find it at!