[pnfs] pNFS and sessions testing at connectathon

[Iyer, Rahul]

> 
> > Hi Lisa,
> > I had a few questions about your implementation for connectathon.
> >  
> > 1. Will you hae a complete sessions implementation? Our 
> implementation 
> > of sessions is incomplete in that we do not do proper 
> sequence number 
> > accounting and/or flow control. So, it would be nice if you 
> could turn 
> > off sequence # checking on your server.
> 
> Yes, the sessions implementation is pretty complete.  No 
> matter what, I think we will be able to turn off the sequence 
> id checking on the server if needed.

And does your server mandate that we must be able to answer callbacks in
order to get layouts? We do not have callbacks working so that will be a
problem.

> 
> >  
> > 2. Would you be dealing with the case of multiple sessions 
> created to 
> > the data server?
> 
> I think I am a little confused by this.  Do you mean multiple 
> sessions created by the same client?  Or multiple clients 
> creating a session?
> 
> If it is either of those, we have a little work to do prior 
> to Cthon, but should be able to handle this.
> 
Never mind. I had meant multiple sessions created for a given clientid.
Doesn't seem like you're doing it. The way the our servers (Linux,
Ontap) are coded, they won't handle multiple sessions for a given
clientid.

> >  
> > 3. What days do you plan to be at connectathon? I have to attend a 
> > Netapp meeting on the 6th (Tuesday) so I won't be at 
> Connectathon. I 
> > should be able to attend all other days. hat days do you think are 
> > convenient for you?
> 
> 
> We plan to be there from the beginning (Thursday, 2/1)  to 
> the end (Thursday, 2/9).  How about we start up testing as 
> soon as possible, say Thursday, 2/1?
> 
I agree... We'll be there starting 2/1 too. We can start as soon as we
get set up.
Regards
Rahul