Re: [nfsv4] IETF 98 - NFSv4 WG meeting (soliciting agenda items)

I'm not sure whether I'd be able to physically attend, but I'm assuming
facilities for remote participation will be available.

I'll need 30-40 minutes for extension-related discussion:

   - One short part will be primarily retrospective, explaining the
   consequences of the new extension model.  The content will not be changed
   substantially by the completion (or not) of an updated draft for umask, or
   the progress of AD Evaluation or IESG discussion of the three
   versioning-related documents.
   - Another part will be a short proposal for an extension that is a bit
   bigger than those that have been done so far.  No document has yet been
   submitted but I expect to submit one before the deadline.  The idea is to
   provide some (privileged) clients the ability to be notified of changes in
   the stateid set of other clients.   By treating pNFS data servers as such
   privileged clients, we could define an optional standard way for metadata
   servers and data servers to interact, in providing pNFS file support.  It
   appears that this could also support the tight coupling mode of flex files.

We will need some time to discuss RDMA-related issues but Chuck and I need
to talk about the length of time needed and who will talk about what.  Two
possible ways of dividing things up:

   - Chuck could talk about stuff related to Version One including
   rfc56667bis, while I cover stuff related to Version two including
   rpcrdma-version-two, rpcrdma-rtrext, and nfsulb (to make the NFS ULBs
   transport-generic).
   - Chuck could talk about stuff related to Version One (including
   rfc56667bis) and rpcrdma-version-two, while I cover stuff related to
   Version Two extensions including rpcrdma-rtrext and nfsulb.

I hope we can hear from Christoph regarding NVMe-related stuff, whether he
is able to be physically present or not.  In addition, I'd like us to her
about the progress of the scsi layout draft.  That took eight months to go
from -00 to working group consensus :-) and over thirteen months after that
and it is not out yet :-(. I think we have a general problem and need to
understand it better before we try to fix it.  Hearing from Christoph would
help us understand what is going on.

Other people may have other items so it appears to me that the issue not
the length of the session but how many.  If you do decide to go with two
sessions, one possible division would be:

   - One session with everything performance-related: RDMA, NVMe, and
   possibly Andy's trunking-relating work, if he wants to talk about that.
   - Everything else including the extension stuff.

On Tue, Jan 31, 2017 at 1:48 PM, spencer shepler <spencer.shepler@gmail.com>
wrote:

>
> Hi.  IETF 98 is coming up (Chicago) and the cutoff date for meeting
> requests is Feb 10th.
>
> Please send in your agenda requests, etc. so that we can determine if we
> will be meeting in Chicago and duration of meeting.
>
> Thanks,
> Spencer S.
>
>
> _______________________________________________
> nfsv4 mailing list
> nfsv4@ietf.org
> https://www.ietf.org/mailman/listinfo/nfsv4
>
>