[snip NFS lockups]

I'v also had problems with NFS with 1.0 [patched with patches 1 to 5]
on a P90.

I'v been running a test of the usability of SLIP and choose to compile
X11R6 over a 9600 SLIP line.  I do the folllowing:

mount -r -t nfs -o -a1,-r1024,-w1024,-i bigmachine:/usr/local/src/X11R6 /mnt

I am also using a union mount such that the .o and binaries are on a
local disk.  I use this:

mount -t union /usr/local/tmp/X11R6/build /mnt

In order to watch the progress I do a 'stty status ^t' in all the
effected virtual consoles and begin the build.  Everything seems to
progress along fine, until, at some random time, a large amount of
disk access occure.  The cause of these accesses could be expiring
news, unpacking a large tar archive, reading news with GNUS, etc.  In
any case, the build of X11R6 stops, a ^T against the make shows that
the process is hung in a 'vget' call.  In addition, the process that
caused these disk accesses is also hung.  The two processes are
unkillable.  In addition, sometimes, the entire system will mostly
hang, a ^T in a shell, for example, indicates that it is hung in
'netio', even if the shell wasn't involved with the network.  This is
all very particular.


Brad Spencer - bspencer@bassun.cb.att.com [work]
               bspencer@cbnmva.att.com [more work]
               brad@anduiin.eldar.org [home]