Wednesday, 2018-10-17

openstackgerritLi Liu proposed openstack/cyborg master: Added Glance Client for Image downloading  https://review.openstack.org/60990602:44
*** zhipeng has joined #openstack-cyborg07:07
*** helenafm has joined #openstack-cyborg07:20
*** dims has quit IRC13:40
*** munimeha1 has joined #openstack-cyborg13:48
*** Sundar has joined #openstack-cyborg13:57
*** wangzhh has joined #openstack-cyborg14:01
*** Li_Liu has joined #openstack-cyborg14:01
sum12meeting ?14:03
*** dims has joined #openstack-cyborg14:03
Li_Liuyup14:05
Li_Liulet's wait for few more min14:05
Li_Liu#startmeeting openstack-cyborg14:06
openstackMeeting started Wed Oct 17 14:06:18 2018 UTC and is due to finish in 60 minutes.  The chair is Li_Liu. Information about MeetBot at http://wiki.debian.org/MeetBot.14:06
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:06
*** openstack changes topic to " (Meeting topic: openstack-cyborg)"14:06
openstackThe meeting name has been set to 'openstack_cyborg'14:06
Li_Liu#topic Roll Call14:06
*** openstack changes topic to "Roll Call (Meeting topic: openstack-cyborg)"14:06
Sundar#info Sundar14:06
Li_Liu#info Li_Liu14:06
wangzhh#info wangzhh14:06
Li_LiuLet's wait for shaohe14:07
SundarWhat is the agenda?14:08
Li_LiuWe will go over all the pending patches and tasks. and discuss the demo plan for the summit14:08
SundarI'd like to bring up the need to get reviews for: https://review.openstack.org/#/q/status:open+project:openstack/cyborg-specs14:08
*** xinran has joined #openstack-cyborg14:08
SundarNext week is Stein-1 milestone14:09
Li_Liuyes14:09
Li_Liuwe should get the pending one merged as much as we could14:09
Li_Liuok, let's get started first14:11
Li_Liu#topic Pending patches14:11
*** openstack changes topic to "Pending patches (Meeting topic: openstack-cyborg)"14:11
Li_Liuhttps://review.openstack.org/#/q/status:open%20project:openstack/cyborg14:11
SundarMay be we can start with wangzhh's patch on drivers14:11
Sundarhttps://review.openstack.org/#/c/596691/ ?14:12
Li_LiuAdded Glance Client for Image downloading  --- Please help review this so that we can merge. The summit demo needs this one14:12
wangzhhOf course. My patch is based on ovo now.14:12
wangzhhAnd Sundar, I have sent an email about your questions. Did u get it?14:13
Sundarwangzhh: Sorry, don't remember seeing it. When was it sent?14:14
*** shaohe_feng has joined #openstack-cyborg14:14
*** Li_Liu_ has joined #openstack-cyborg14:14
Li_Liu_I am back14:15
shaohe_feng#info shaohe_feng_14:15
Li_Liu_did I miss anything?14:15
Li_Liu_Hi shaohe14:15
shaohe_fengmorning  Li_Liu_14:15
wangzhhSundarMaybe this morning for you.14:15
xinran#info xinran14:16
xinranHi all14:16
*** dims has quit IRC14:16
Sundarwangzhh: OK. I haven't got it yet. I see your new patch set and responses. Thanks.14:16
wangzhhHi, xinran shaohe.14:16
*** Li_Liu has quit IRC14:16
*** shaohe_feng has quit IRC14:17
*** Li_Liu has joined #openstack-cyborg14:18
Li_Liuback again..14:18
Li_Liuare we gonna merge Xinran's spec?14:18
Li_Liudoes it still depend on anything?14:18
SundarLi_Liu:  Re. Glance client, I gave some reviews. Wondering why can't we just call Glance client directly? Do we need a wrapper?14:18
SundarLi_Liu: which spec?14:19
xinranLi_Liu:14:19
Li_Liuhttps://review.openstack.org/#/c/597991/14:19
*** Li_Liu_ has quit IRC14:19
*** dims has joined #openstack-cyborg14:19
xinranLi_Liu:  I have discussed with Sundar offline, the current solution after PTG is different with this one.14:20
Li_LiuSundar, of course, but I just wanna do it similar to Nova --> created a wrapper API for it14:20
Li_Liuxinran: any plan on updating it?14:20
Li_Liudid I drop?14:22
Li_Liuah.. i am good :)14:22
xinranSundar has a new spec https://review.openstack.org/#/c/603955/ in nova community14:22
Li_Liuxinran, so that one will replace yours right?14:22
*** shaohe_feng has joined #openstack-cyborg14:22
shaohe_fenghi wangzhh14:23
xinranyes in the future14:23
wangzhhDidn't get the review email about this patch...14:23
shaohe_fengseems I miss something14:24
shaohe_fengsomething wrong with my net.14:24
shaohe_fengno any message on the screen for a long time.14:24
*** Li_Liu_ has joined #openstack-cyborg14:24
SundarThanks, Xinran. All, please review https://review.openstack.org/608624 . That is the proposed implementation (high-level design) of Cyborg APIs for Nova14:25
xinranbut you guys can use https://review.openstack.org/#/c/596187/ this API if  it's urgent. It can works with nova, but of course need change code on nova side.14:25
*** Li_Liu__ has joined #openstack-cyborg14:26
Li_Liu__Sundar, could you help me run the meeting for now...  my connection is so bad14:26
Li_Liu__I am going through all the patches in https://review.openstack.org/#/q/status:open%20project:openstack/cyborg14:26
wangzhhxinran: So this patch should be abandoned? https://review.openstack.org/#/c/597991/14:26
*** shaohe_feng has quit IRC14:27
SundarLi_Liu: Sure. NP.14:27
*** Li_Liu has quit IRC14:27
SundarXinran: On https://review.openstack.org/#/c/601150/, this patch tries to create RP trees from the agent, IIUC. It is better to do it from the conductor14:29
Sundarfor the reasons we have discussed in the past14:29
*** Li_Liu_ has quit IRC14:29
SundarAlso, it is desirable to have all db access from the conductor (and not from the agent)14:29
*** shaohe_feng has joined #openstack-cyborg14:29
xinranHmmm,  Sundar do you wanna continue work on this spec https://review.openstack.org/#/c/597991/ or you prefer to create a new one?14:29
SundarXinran, I think this spec should be dropped -- that is a better use of your time and energy :)14:30
shaohe_fengwhy dropped?14:31
SundarShaohe: I already explained in the reviews. Basically, this is not in line with the Nova flow at all14:31
shaohe_fengA patch is not perfect, every reviewers should give his comments and let the auther to improve it.14:32
shaohe_fengthat's the rule14:32
*** Li_Liu_ has joined #openstack-cyborg14:32
SundarI gave my comments14:32
SundarThink about how this even relates to the Nova spec14:33
*** Li_Liu__ has quit IRC14:33
shaohe_fengso the summary of new cyborg design:14:34
shaohe_feng1. split the nova flavor to device profile14:34
shaohe_feng2. change the accelerator api to var api.14:34
xinranSundar: I can modify this spec according to your nova spec if you want once your spec is merged.14:34
shaohe_fengright?14:35
xinranSundar:  or you prefer to write a new one by yourself?14:35
SundarXinran, Shaohe: we need to start with the Cyborg API signatures proposed for Nova, and work downwards into the details. That is what we do in https://review.openstack.org/#/c/608624/14:36
*** openstackgerrit has quit IRC14:36
*** Li_Liu_ has quit IRC14:36
shaohe_fengwhat is signatures?14:36
*** shaohe_feng has quit IRC14:37
SundarShaohe: API parameters, responses, HTTP return codes14:37
SundarShahe: re. your question about 2 parts, the device profiles are an important part. So is the whole flow about creating VARs in an unbound state, then binding them, then attaching.14:38
*** Li_Liu_ has joined #openstack-cyborg14:40
*** shaohe_feng has joined #openstack-cyborg14:40
SundarAso, the device model where we map deployables to RPs, as we discussed last week14:40
shaohe_fengI do not care you name is var or accelerator.14:41
Sundar A VAR is not an accelerator.14:41
shaohe_fengbut as user var is puzzle, like vport.14:41
Sundar"var is puzzle" -- what does that mean?14:42
shaohe_fengvar is used for a choose a device to bind to a VM.14:42
SundarA VAR is a Virtual Accelerator Request. It abstracts the state of the request, starting from user requirements, moving on to the Nova's selection of an alloc cand with device RPs, and then Cyborg bindings14:43
shaohe_fengany project, you seen they expose a var object to user?14:43
SundarThis is conceptually different from what we had before the PTG14:43
shaohe_fengvport, vvolume?14:44
Li_Liu_shaohe, are you more concern with the naming or the concept?14:44
shaohe_fengthe initial design, is Request a Accelerator to VM.14:44
shaohe_fengbut we does not name it virtual  accelerator, we just named it accelerator14:45
shaohe_fengno need to use virtual14:45
SundarShaohe: is this about names, or concepts?14:45
shaohe_fengno need to emphasize virtual14:46
shaohe_fengyou can discuss with Li_Liu_ about the accelerator history14:46
shaohe_fengabout the name14:46
Li_Liu_I think "virtual" word indeed confuses a bit14:46
wangzhhShaohe: Accelerator is not same as VAR. Maybe u confusion about them?14:47
*** shaohe_feng has quit IRC14:47
Li_Liu_but I agree it's different from concept of accelerator14:47
SundarNeutron has something called a VIF -- virtual interface. The word 'virtual' is used in many contexts. I don't understand the issue.14:48
*** munimeha1 has quit IRC14:49
Li_Liu_I guess when you put all of these    virtual accelerator / accelerator request / virtual accelerator request together. that's where the confusion starts14:50
Li_Liu_Sundar, does nova needs to care if a accelerator is virtual or not?14:51
SundarLi_Liu: We have accelerators and we have VARs. Former is a resource represented in Placement. VAR, as you said, is different -- it is a request for an accelerator -- not known to Placement. There are only 2 things14:51
wangzhhI think it's better to have a clear definition of VAR. In the doc, for other developer or users.14:52
wangzhhSundar: Agree.14:52
SundarLi_Liu: Accelerators are resource class inventories -- they are not physical hardware, even in the previous Rocky proposal14:52
SundarLi_Liu: AFAIK, we never used the term 'virtual accelerator' to refer to an accelerator.14:53
Li_Liu_I know, I think Cyborg maybe should hide physical/virtual accelerator concept from nova.14:54
*** shaohe_feng has joined #openstack-cyborg14:54
Li_Liu_or can we rename it to Accelerator Virtual Request14:54
Li_Liu_as, according to you, the request is virtual but not the accelerator?14:55
shaohe_fengaccelerator is OK.14:55
shaohe_fengand Request is verb14:55
Li_Liu_shaohe_feng, we do have accelerator, but sunder needs something different to work with placement14:56
SundarLi_Liu: yes, the request is virtual. If we just call it an Accelerator Request, say ARQ or something, are you ok with that?14:56
shaohe_fengRequest  and be apply to a noun accelerator14:56
shaohe_fengRequest  can be apply to a noun accelerator14:56
shaohe_fengthis is not conflict with RESTFUL style.14:57
SundarLi_Liu: "hide physical/virtual accelerator" Please note that the term 'accelerator' never refers to any physical entity :)14:57
SundarIn Placement, an accelerator type is a resource class (e.g. CUSTOM_ACCELERATOR_FPGA). There are number of resources of one RC in a single RP14:58
Li_Liu_ARQ is kinda cool to me14:58
wangzhhSundar, could u explain the differences between Acc and VAR? I just know parts of them. I think shaohe is confused about them.14:58
SundarSo, it is just a number that Placement is counting as RC inventory14:59
shaohe_fenggenerally, for a REST, a resource belong to a collection, and the the resource can support different verb action14:59
Li_Liu_btw shaohe, Request cal also be a noun15:01
shaohe_fengso request is a resource15:01
Sundarwangzhh: OK. Placement has the notion of Resource providers (RPs) and Resource Classes (RC). We model each accelerator type as an RC. An accelerator is a unit of offload that can be assigned individually (to a VM, container, ...)15:01
shaohe_fengbind is verb to this request resource15:01
Li_Liu_shaohe_feng, I think that's Sundar's idea, is it? Sundar15:02
shaohe_fengOK, seldom see this rest style.15:03
SundarVARs are indeed resources in the REST API. We apply HTTP operations like GET, POST etc. to that resource.15:03
SundarShaohe: have you reviewed the Nova spec?15:03
shaohe_fenghave a look at  https://review.openstack.org/#/c/608624/2/specs/stein/approved/cyborg-api-wflows-for-instance-ops.rst15:04
wangzhhSundar, Yes.  So is there any differences on data structure?15:04
shaohe_fengany way,  virtual Accelerator Requests is puzzle me.15:04
wangzhhBetween VAR and Acc.15:05
shaohe_fengso VARs is a collection, and var is resource.15:05
shaohe_fengVARs  means multi requests.15:05
Sundarwangzhh: Getting back to your question :) Yes, a VAR is an OVO that is stored in Cyborg db. An accelerator is not a data structure per se: it is the inventory of a resource provider maintained by Placement n its db15:05
shaohe_fengvar  is one request15:05
shaohe_fengright?15:06
shaohe_fengand why we need var? just for consistency´╝č15:07
SundarYes, VAR is singular, VARs is plural. The Nova spec 603955 defines APIs on the collection VARs, so that we can do batch operations15:07
*** Li_Liu has joined #openstack-cyborg15:09
*** Li_Liu_ has quit IRC15:09
shaohe_fenganyway, here you define the VAR is same to the accelerator  concept in our  initial design15:10
SundarNo15:10
SundarAccelerator != VAR15:10
shaohe_fenganyway, here you define the VAR is same to the accelerator  concept in my initial design15:11
SundarThe concept of accelerator as a RC existed from Rocky cycle15:11
wangzhhSo, when cyborg-agent started, It collects which one? I think It is acc. And VAR will be created when user want accs to attach?15:11
SundarVAR is new from Stein PTG15:11
shaohe_fengand your Accelerator  is same to the allocation unit in  my initial design :)15:11
shaohe_fengother's no different15:13
shaohe_feng just name is different15:13
Sundarwangzhh: Cyborg agent collects device (physical hardware) and accelerator info from the driver. Please see the driver report structure in https://docs.google.com/presentation/d/1Anud3Qbcb0P3G245HpHduHhslx1MJljGD6wqPDy7o9E/edit#slide=id.g44d3e3519f_4_13415:13
shaohe_fengOK, let go ahead for others.15:13
SundarI am 15 minutes into my next meeting, and it is a bit difficult to flip back and forth. I need to drop out now.15:15
SundarShaohe: I am open to discussion at other times or meetings.15:15
shaohe_fengSundar: OK. we can also discuss it off line.15:15
shaohe_fengSundar: have a good day. bye.15:16
*** Li_Liu_ has joined #openstack-cyborg15:16
*** Li_Liu has quit IRC15:18
*** Li_Liu has joined #openstack-cyborg15:19
Li_Liuping15:20
Li_Liuyou guys still there?15:20
Li_Liuanyone?15:22
*** Li_Liu_ has quit IRC15:22
*** Li_Liu has quit IRC15:25
*** Li_Liu has joined #openstack-cyborg15:25
Li_Liuif no one is here, I will end the meeting for today15:26
Li_Liuwe can carry on the discussion offline15:26
Li_Liu#endmeeting15:26
*** openstack changes topic to "A zuul config error slipped through and caused a pile of job failures with retry_limit - a fix is being applied and should be back up in a few minutes"15:26
openstackMeeting ended Wed Oct 17 15:26:25 2018 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:26
openstackMinutes:        http://eavesdrop.openstack.org/meetings/openstack_cyborg/2018/openstack_cyborg.2018-10-17-14.06.html15:26
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/openstack_cyborg/2018/openstack_cyborg.2018-10-17-14.06.txt15:26
openstackLog:            http://eavesdrop.openstack.org/meetings/openstack_cyborg/2018/openstack_cyborg.2018-10-17-14.06.log.html15:26
*** Li_Liu has quit IRC15:27
*** shaohe_feng has quit IRC15:29
*** munimeha1 has joined #openstack-cyborg15:37
*** card.freenode.net changes topic to "Pending patches (Meeting topic: openstack-cyborg)"15:37
*** openstackgerrit has joined #openstack-cyborg15:47
openstackgerritRajat Dhasmana proposed openstack/cyborg master: Add cyborg-status upgrade check command framework  https://review.openstack.org/61136815:47
openstackgerritMerged openstack/cyborg master: Added Glance Client for Image downloading  https://review.openstack.org/60990615:47
*** efried is now known as efried_pto16:09
*** Sundar has quit IRC16:21
openstackgerritRajat Dhasmana proposed openstack/cyborg master: Add cyborg-status upgrade check command framework  https://review.openstack.org/61136816:34
*** helenafm has quit IRC16:35
*** persia has left #openstack-cyborg16:52
*** wangzhh has quit IRC17:15
*** xinran has quit IRC21:36
*** munimeha1 has quit IRC22:20

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!