On Wed, Dec 20, 2006 at 01:09:47PM -0500, up AT 3 DOT am wrote:
> >
> 
> > Also, chg-disk needs a "tapedev" which points to the root
> > of a directory tree, containing "slotN" (N = 1... up-to-tapecycle)
> > subdirectories.
> 
> I used:
> 
> tapedev "file:/mnt/usb/dumps"
> 
> The example page for chg-disk uses:
> 
> tapedev "file:/amandatapes/test/slots"
> 
> I assume either should work fine, but still find the distinction between
> the "slot" name and the (virtual) tape name a little confusing.  Is there
> any point in using:
> 
> labelstr "^FULL[0-9][0-9]*$"
> 
> Like I did for "real tapes", when amanda is going ahead and using "slotN"
> for the names of each dmp directory anyway?
> 

chg-disk tries to mimic a real-world, physical tape drive,
tape changer, and tapes.  So the "slots" could be occuppied
by any tape though with virtual tapes this is typically
a static organization.

The physical tapes can have a human readable label.  Typically
this will match the labelstr, but need not.  Amanda can't read
that external, human readable label, so the amlabel command
writes a file containing the machine readable label.  This
must match the labelstr.

For vtapes the analog to the human readable, external label
is the directory name for the vtape.  When the vtape is
amlabel'ed, a file, containing the tape label, is written to
the directory (i.e. to the tape) just like a physical tape.
And again, the directory name and the contents of the label
file need not match.

My vtapes are scattered across multiple drives.  Because of
this my slots are not directories, but symlinks to the actual
vtape directories wherever they are located.

-- 
Jon H. LaBadie                  jon AT jgcomp DOT com
 JG Computing
 4455 Province Line Road        (609) 252-0159
 Princeton, NJ  08540-4322      (609) 683-7220 (fax)