> Juergen Marenda writes:
> > Thus spoke Matthias Scheler:
> >> I suppose SunOS does this via NIS (it's hard to get a Sun working without
> >> NIS) which you probably don't use.
> > 
> > SunOS (4.x) shutdown
> > uses /etc/xtab to get a list of hosts who have remote mounts.
> 
> No, it makes an RPC call (clnt_create()) to the portmapper to get the mountd
> program, and then it makes another call (clnt_call()) to mountd to get the
> list of hosts (MOUNTPROC_DUMP) and it then builds a linked list of the hosts
> to notify from that.

  Sorry to flog a potentially dead horse, but I'd recently noticed (while
trying to investigate this shutdown stuff for the original poster, my
advisor), that mountd's MOUNTPROC_DUMP list is not always up-to-date or
correct.  For example, one time it listed 3 hosts, even though there were
active users at >3 different machines (and not X-terminals -- real,
independent machines) with their home directories NFS mounted (ergo,
there's a problem somewhere).  Anybody else had any experience with this
(incorrect mountd data) before I delve into it in my spare time (yeah,
right!)????   Is there any better way of seeing what machines are using
NFS stuff besides asking mountd (which seems to be the ideal source)?

  Part of our problem may be that the NFS server was rebooted, and the
clients picked up again when the server came back up, without manually
re-mounting, so it could be simply a matter of mountd not recognizing
these 'inherited' clients, but I believe that didn't quite hold water,
either, due to the machines that were and weren't listed.  And if that were
the case, it's probably a 'bug' in mountd, anyway, and should be looked at.

  I'll try to get a little more data on how mountd gets out of touch with
reality in my free time.

  Brian
-- 
Brian Grayson (bgrayson@ece.utexas.edu)
Graduate Student, Electrical and Computer Engineering
The University of Texas at Austin
Office:  ENS 406       (512) 471-8011
Finger bgrayson@pine.ece.utexas.edu for PGP key.