Problem with NFS mounted disks over long/slow (9600) line.

[Bob Garwood]

The Situation:

o Two 3/60's separated by a long distance each on a fast local network and
  each network connected by a slow (9600 I think) line.

o Because two programmers working on a project are located at these two
  remote sites, we decided to put the project under SCCS on one machine and
  then have the other machine remotely mount the disk.  Each programmer
  would then set up a directory hierarchy under their own tree with symbolic
  links from the various SCCS directories to the one true SCCS directory.

o Both 3/60's are currently running OS 4.0.3

The Problem:

o For "large" files trying to do almost anything with them at the remote
  site results in 

	NFS server ... not responding still trying

  Even though, on this same remote machine at the very same time that this
  message shows up, one can continue to do "ls" on the servers disk, do
  "stuff" with "small" files, etc.  This "not responding" message is never
  followed by the familiar "NFS server ... OK" message unless an interrupt
  (control-c or control-z) is issued.

o Things which set off this behaviour are: sccs get, vi, cp

o "large" vs "small":   The largest file which has NOT shown these symptoms
  is 293 lines long and 6658 characters.  The smallest file which HAS shown
  these symptoms is 324 lines long and 7278 characters. 

If you have any suggestions on what our problem might be, any pointers to
things we might try and do, anything that might help us, please send me
e-mail.  Thanks.

Bob Garwood
National Radio Astronomy Observatory
bgarwood at sngldsh.cv.nrao.edu	
bgarwood at nrao.edu