NFS & DUMPS & DISKLESS NODES

[R. J. Auletta]

Equipment: Vax3600 & two VS2000 (diskless) Ultrix 3.1


Problem: When file systems on the 3600 are unmounted and then
         mounted again without rebooting either the 3600 or the
         2000's the 2000's refuse to mount the file system
         complaining about a stale NFS handle, and the 3600
         prints the following

    NFS server: unexported fs(9,5) file 2, getattr, client address=129.174.1.31

        while the file system in unmounted. Rebooting the 2000's does
        not fix the problem, it requires that the 3600 be rebooted.
        (Though the 3600 message stops once the file system is mounted.)


Question: Under Ultrix 2.2 this was not a problem. We could umount
          filesystems for fsck and dumps, remount and go on as
          if nothing had happen. I don't see why the vaxstations
          should complain about a stale NFS file handle as the inodes
          have not changed. (?)

	  Is this expected behavior under 3.1 or is the procedure
          fatally flawed?


          Suggestions, comments, ideas would be most welcome and 
          appreciated.


--R J Auletta
  rauletta at gmuvax2.gmu.edu  (newsmachine)
  rauletta at sitevax.gmu.edu  (usual machine)

(BTW, anyone know what the bpi and length is of a TK70?)