On Wed, 03 Sep 2003 01:15:54 GMT, Darren Dunham
wrote:

[ Snip ]

Yes, this is actually not hard to believe at all. Remember when
Sun's st driver would not allow you to write in high-density format
on an Exabyte tape (i.e. the 8500's "4.6 GB" format) if you had
previously written data on a the tape in low-density (i.e. the
8200's "2.3 GB" format)? If you erased the tape (and I believe
even "mt erase" would work, but can't remember), then you could
write in high-density format.

I've never heard of this. Sun's st driver should have nothing to do
with this at all.

Shouldn't, but it sometimes does... It comes as a shock to some that
occasionally software is broken.

If the tape is written in the proper density at BOT, the density should
change, otherwise the existing density is used. This is the same thing
that happens with DLT tapes. The st driver is just setting the proper
bits in the density page. Whether the drive chooses to use the density
settings is another matter.

IIRC (and it was a while ago) the driver was basically a the same as
the one that handled the 8200, and it just didn't understand the
subtleties of when you can or cannot change densities.

I don't have an 8500 around, but I do have access to an 8505. I'll try
to write to it in the different densities. Unfortunately, unlike most
standalone DLTs, the 8500s don't have LEDs that show the current density
settings on the tape.

That was a while ago, and I believe the driver has been fixed...

Still, there do exist devices that have initialization procedures
which record out-of-band metadata on the media. Things can get
royally messed up if the metadata is at odds with the attempted I/O
operation (e.g. when the media is partially degaussed; it becomes
nasty!).

Malc.