> From: Tom Ivar Helbekkmo <tih@nhh.no>
> Date: 26 Oct 1998 19:46:57 +0100
> 
> allisonp@world.std.com (Allison J Parent) writes:
> 
> > Officially 95mb.  depending on how you structure your data and all it 
> > could be as low as 60mb usable. I used to use it to completley image 
> > backup RD53s (71mb) with room to spare.  20mb is way out of line.
> 
> I've been doing 'dump -b 32 -B 90000' here lately.  That is, 32KB
> records and 90000KB expected to fit on the tape.  This was after a
> 'dd bs=32k' that errored out after 92000KB.  On the other, hand, a
> test with 'dd bs=256kb' the other day bombed out after 88832KB; I
> don't know what to think of this.  Aren't all TK50s the same?

All TK50 cartridges  are pretty much the same.  But remember that the
TK50 is a serpentine drive, and switches to another track every time it
gets to the end of the tape.  As the blocks get longer (like 256kB) the
probability of losing one at the end of the tape gets greater.  The
block has to be re-written on the next track.

My TK50 experiments back in 1986 showed that the data capacity of the
tape peaked at a block size around 32kB to 64kB (approx).  Shorter
blocks waste inter-block gaps, longer blocks are lost in track
switching.  Of course, the real data for this is stored somewhere on a
TK50 backup that I can no longer read.

    carl

        carl lowenstein         marine physical lab     u.c. san diego
        {decvax|ucbvax} !ucsd!mpl!cdl                 cdl@mpl.ucsd.edu
                                                  clowenstein@ucsd.edu