NFS stops responding

[Benjamin Scott]

On Wed, Apr 14, 2010 at 10:02 AM, Michael ODonnell
<michael.odonnell at comcast.net> wrote:
> This seems not to have elicited any response, pro or con.  I know of
> no reasons in principle why two machines can't simultaneously act as
> NFS clients and NFS servers - are there any?

  I know relatively little about NFS (and what I know is old), but
from a protocol standpoint, I believe there is nothing preventing a
single host from being both NFS client and NFS server.

  However, I did observe an almost unstated but very real desire to
avoid doing so.  It was almost superstitious in feel.  I suspect this
evolved mainly because, as I mentioned, with sufficiently gnarly
cross-mounts, you can end up with a network of hosts which (1) will
all simultaneously lock-up solid just because one host went down and
(2) be unable to ever successfully boot from a "power out" situation.
(Host A needs host B to work, B needs C to work, C needs A to work...)

  You may think the answer is "don't do that, then".  You might even
be right.  But it was still a depressingly common failure mode in
real-world networks.

  It doesn't help that, in early implementations at least, NFS's
default error recovery mechanism is apparently "hang the whole machine
until it starts working again".

  Does any of this apply to your particular environment?  I dunno.

-- Ben