[NTLUG:Discuss] Re: A UNIX/Linux Staple: The Automounter -- I think I see where we differ ...

[Bryan J. Smith]

Before I let our tangent go any farther, I think I can sum up where we
are differing.  My specifically, I'm talking from the viewpoint of:

1.  SMB v. NFS

If you start looking at more "enterprise" network filesystems, there are
far better and "more stateful" capabilities in other solutions than
NFS.  But compared to SMB, NFS is "more stateful."

If we look outside of SMB v. NFS, yes, I agree with many of your
statements when comparing NFS to a variety of other network
filesystems.  I was just comparing to SMB, since that is what most
Windows sys admins are used to.

2.  NFS as IETF/POSIX spec

Understand I did _not_ mean NFS, especially not the client, as
implemented in the Linux kernel.  You are completely correct when you
say that the Linux kernel NFS client makes assumptions of NFS servers. 
It does all sorts of stuff that causes problems with very compliant NFS
servers.  But I'm not referring to those.

I'm talking about how meta-data is presented to the NFS client compared
to how it is actually on the underlying, local filesystem of the NFS
server -- especially when, again referring back to #1 -- compared to how
the SMB client is presented meta-data versus the actual data on the
underlying, local filesystem of the server.

NFS is far more "raw" in comparison than to SMB.


-- 
Bryan J. Smith                                  b.j.smith at ieee.org 
------------------------------------------------------------------ 
"Communities don't have rights. Only individuals in the community
 have rights. ... That idea of community rights is firmly rooted
 in the 'Communist Manifesto.'" -- Michael Badnarik