What is needed for GPU support is probably a workgroup and not a SIG. As others have mentioned there are already three different groups discussing resource management today and it is proving very difficult to stay in sync.
On Wed, Feb 8, 2017 at 12:05 PM, Quinton Hoole <quin...@hoole.biz> wrote: > +1 to dealing with GPU's with the existing SIG's, rather than spawning a > new SIG (for all the reasons already mentioned). > > On Wed, Feb 8, 2017 at 11:42 AM, Derek Carr <dec...@redhat.com> wrote: > >> I would prefer that we promote the existing resource management group to >> a proper SIG before we elevate creation of a new SIG. >> > +1. > >> Prior to doing that, I would like this group to fulfill its stated >> intention in Q1 of identifying a roadmap for this year, and a mechanism to >> support custom isolation and scheduling of a non-natively managed k8s >> resource. If at that point, we fill a need to continue the group and its >> valuable for all participants, maybe it can be a SIG. >> >> Brendan - I would love to have you participate in our next meeting (+2 >> weeks)? The agenda and meeting notes are captured here: >> https://docs.google.com/document/d/1j3vrG6BgE0hUDs2e-1ZUegKN >> 4W4Adb1B6oJ6j-4kyPU/edit# >> >> In that group, we have already presented a number of workload types (many >> that require GPU) and are trying to define a node isolation and scheduling >> roadmap to support them properly. We are also trying to state more >> succinctly when we choose to manage a particular resource/device in core >> k8s versus look towards a custom pod isolator. >> >> Thanks, >> Derek >> >> >> On Wed, Feb 8, 2017 at 2:32 PM, David Oppenheimer <david...@google.com> >> wrote: >> >>> I agree with Tim, the resource management working group is the best of >>> the three as it has folks from both sig-node and sig-scheduling. >>> >>> >>> On Wed, Feb 8, 2017 at 11:22 AM, Tim St. Clair <timoth...@gmail.com> >>> wrote: >>> >>>> General +1 >>>> >>>> We're already deep into resource isolation extensibility and ... would >>>> love your input ;-) ping decarr on time-date of next mtg. >>>> >>>> On Wed, Feb 8, 2017 at 1:19 PM, Clayton Coleman <ccole...@redhat.com> >>>> wrote: >>>> > Yeah, I don't know how the separate SIG would be able to get any >>>> things done >>>> > without working with all three of those - who are already all working >>>> on >>>> > that. >>>> > >>>> > On Wed, Feb 8, 2017 at 2:11 PM, 'David Oppenheimer' via Kubernetes >>>> > developer/contributor discussion <kubernetes-...@googlegroups.com> >>>> wrote: >>>> >> >>>> >> I think there are three existing forums that would be preferable to >>>> >> creating another SIG: >>>> >> - sig-scheduling >>>> >> - sig-node >>>> >> - resource management working group (comprised of sig-scheduling and >>>> >> sig-node folks) >>>> >> >>>> >> I'm concerned that the number of meetings on closely related topics >>>> is >>>> >> reaching a point where it will be hard to get the right people to >>>> attend all >>>> >> the relevant meetings. >>>> >> >>>> >> >>>> >> On Wed, Feb 8, 2017 at 11:01 AM, Brendan Burns < >>>> brendan.d.bu...@gmail.com> >>>> >> wrote: >>>> >>> >>>> >>> Hey Folks, >>>> >>> I'd like to propose a new Kubernetes SIG, SIG-GPU which focuses on >>>> >>> GPU-based workloads for Kubernetes. I'm seeing increased interest >>>> in these >>>> >>> workloads, and there are some very scheduler/modelling specific >>>> things to >>>> >>> GPUs that I think we want to figure out how to model. Not to >>>> mention that >>>> >>> we need better introspection and reporting for GPU resources coming >>>> up from >>>> >>> the node to the master. >>>> >>> >>>> >>> While there is a bunch of cross-over with SIG-scheduling here, I >>>> think >>>> >>> that GPU has enough domain specific items that its worth breaking >>>> out as its >>>> >>> own SIG. >>>> >>> >>>> >>> Thoughts? >>>> >>> >>>> >>> -- >>>> >>> You received this message because you are subscribed to the Google >>>> Groups >>>> >>> "Kubernetes developer/contributor discussion" group. >>>> >>> To unsubscribe from this group and stop receiving emails from it, >>>> send an >>>> >>> email to kubernetes-dev+unsubscr...@googlegroups.com. >>>> >>> To post to this group, send email to kubernetes-...@googlegroups.co >>>> m. >>>> >>> To view this discussion on the web visit >>>> >>> https://groups.google.com/d/msgid/kubernetes-dev/5b6b606e-60 >>>> 66-4586-be70-5184064f7404%40googlegroups.com. >>>> >>> For more options, visit https://groups.google.com/d/optout. >>>> >> >>>> >> >>>> >> -- >>>> >> You received this message because you are subscribed to the Google >>>> Groups >>>> >> "Kubernetes developer/contributor discussion" group. >>>> >> To unsubscribe from this group and stop receiving emails from it, >>>> send an >>>> >> email to kubernetes-dev+unsubscr...@googlegroups.com. >>>> >> To post to this group, send email to kubernetes-...@googlegroups.com >>>> . >>>> >> To view this discussion on the web visit >>>> >> https://groups.google.com/d/msgid/kubernetes-dev/CAOU1bzfeRs >>>> yrxALS0eVdNM9n_bNGaXQE6sQsUmxe1%3DPnAT9nDQ%40mail.gmail.com. >>>> >> >>>> >> For more options, visit https://groups.google.com/d/optout. >>>> > >>>> > >>>> > -- >>>> > You received this message because you are subscribed to the Google >>>> Groups >>>> > "Kubernetes developer/contributor discussion" group. >>>> > To unsubscribe from this group and stop receiving emails from it, >>>> send an >>>> > email to kubernetes-dev+unsubscr...@googlegroups.com. >>>> > To post to this group, send email to kubernetes-...@googlegroups.com. >>>> > To view this discussion on the web visit >>>> > https://groups.google.com/d/msgid/kubernetes-dev/CAH16Sh%2By >>>> pQ9-bazVaMtuyJ3K%2BzYhF9xaJDbaVi2FALWDECPOTQ%40mail.gmail.com. >>>> > >>>> > For more options, visit https://groups.google.com/d/optout. >>>> >>>> >>>> >>>> -- >>>> Cheers, >>>> Timothy St. Clair >>>> >>>> “Do all the good you can. By all the means you can. In all the ways >>>> you can. In all the places you can. At all the times you can. To all >>>> the people you can. As long as ever you can.” >>>> >>> >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Kubernetes developer/contributor discussion" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to kubernetes-dev+unsubscr...@googlegroups.com. >> To post to this group, send email to kubernetes-...@googlegroups.com. >> To view this discussion on the web visit https://groups.google.com/d/ms >> gid/kubernetes-dev/CAHROWxQn%2Bt9AVrS2Wdk6q3Yt%2B57tjNWMaJ7 >> pqh-V%2BRQaoc5JzA%40mail.gmail.com >> <https://groups.google.com/d/msgid/kubernetes-dev/CAHROWxQn%2Bt9AVrS2Wdk6q3Yt%2B57tjNWMaJ7pqh-V%2BRQaoc5JzA%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> >> For more options, visit https://groups.google.com/d/optout. >> > > > > -- > Quinton Hoole > quin...@hoole.biz > > -- > You received this message because you are subscribed to the Google Groups > "Kubernetes user discussion and Q&A" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to kubernetes-users+unsubscr...@googlegroups.com. > To post to this group, send email to kubernetes-users@googlegroups.com. > Visit this group at https://groups.google.com/group/kubernetes-users. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q&A" group. To unsubscribe from this group and stop receiving emails from it, send an email to kubernetes-users+unsubscr...@googlegroups.com. To post to this group, send email to kubernetes-users@googlegroups.com. Visit this group at https://groups.google.com/group/kubernetes-users. For more options, visit https://groups.google.com/d/optout.
