A few weeks ago, I asked if anyone had experience with optical
disks and AFS. I got two answers, which indicated that thrashing
of mounted disks would be a problem and that there were some
problems with the size and number of /vicep* partitions.

Since then I have been in contact with Andataco, a company
selling a line of optical jukeboxes with a 1 GB cache disk. They
use a software program called AMASS.

   "AMASS Multi-Platter, Virtual File Management Software treats the
   jukebox as one big disk. It is unique in its ability to provide
   direct access combined with a single device, single filesystem
   view of the system. This relieves users and applications of the
   burden of tracking media volumes and file locations. ... It is
   compatible with all host network protocols."

   "AMASS is implemented at the virtual file system (VFS) layer of
   the host operating system and is transparent to other software
   programs."

The above are quotes from sales literature I received recently.

This system look very attractive as a location for the /vicep*
partitions. The cache disk looks very atractive as it would cover
up any thrashing problems which might arise. You could get the
best of both worlds, large cheap storage with AFS distributed
caching and security.

But there are some unanswered questions. Since AFS changes fsck,
that implies that there is something different about the way AFS
uses the /vicep* file systems.

My questions are:

  Does anyone have experience with the AMASS software?

  Does anyone see a problem with using a file system implemented at
  the VFS level for AFS partitions?

  I know AFS 3.2 has a limit on the number of partitions, but is
  there a limit on the size of each partition? Somewhere I heard
  there may be a 2GB limit on the partition size?

           Douglas E. Engert
           Systems Programming
           Argonne National Laboratory
           9700 South Cass Avenue
           Argonne, Illinois  60439
           (708) 252-5444

           Internet: [EMAIL PROTECTED]

Reply via email to