[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Downgrading the priority of nfs-utils

On Wed, Nov 08, 2006 at 12:35:37PM +1100, Brian May wrote:
> >>>>> "Miles" == Miles Bader <miles@gnu.org> writes:
>     Miles> [Isn't nfs4 rather different than previous versions, in
>     Miles> that it's fixed some of the most egregious "nfs
>     Miles> bogosities"?]
> I have been told NFS 4 has nothing in common with NFS except the name,

This may be because early versions of the Linux implementation for NFSv4
made it impossible to connect to v2/v3 and v4 servers at the same time
from the same client. This is no longer true.

NFSv4 is still an RPC-based protocol; as with any RPC-based protocol,
the server is listed on the portmapper along with a protocol version, so
an NFSv4 server will not communicate with an NFSv3 client (you would
need an NFSv3 server running on the same host to be able to communicate
with an NFSv3 client, as is the case for v2 vs v3, too). The way you
export your filesystems to the client is also fairly different, and
pretty strange if you know how NFSv2 and v3 work. But apart from that,
it's still pretty similar.

> and its reputation for being insecure (even if this reputation in
> unfair...).

It most likely is.

>     Miles> All things considered I'd rather have nfs, even in it's
>     Miles> horrid traditional form, than nothing.
> There are still times when traditional NFS is still the best solution
> (disclaimer: I haven't user NFS 4).

Perhaps you should give it a try.

> Does nfs-kernel-server support v4 yet?

Yes, though you need a fairly recent kernel to be able to do so,
especially if you want the kerberos goodies. The one in sarge isn't able
to; the one in etch will be.

<Lo-lan-do> Home is where you have to wash the dishes.
  -- #debian-devel, Freenode, 2004-09-22

Reply to: