[Info-vax] strange backup behavior

[AEF]

On Aug 11, 4:05 pm, glen herrmannsfeldt <g... at ugcs.caltech.edu> wrote:
> AEF <spamsink2... at yahoo.com> wrote:
>
> (snip, I wrote)
>
> <> Yes. ?But given a command to skip to a specific tape mark, the
> <> drive knows which track and approximately where on the tape that
> <> tape mark is. ?The heads are moved to the appropriate track, and
> <> with an average of one half a pass through the tape the spot is found.
>
> < How does it know?
> (snip)
>
> < I don't think we're talking about the same thing. If you are putting
> < multiple save sets on one tape, you don't need to rewind or skip
> < unless you use /verify, which the OP is not.
>
> If you don't rewind, then there is no problem.  If you do then you
> want to be able to find the EOT fast.
>
> < The only time one would need to do this is to find the EOT mark on a
> < tape that already has save sets on it when you put it in the drive.
> < Tell me how the tape drive is going to know what tracks the tape marks
> < are on when I go to my library (a drawer in a cabinet), pick some tape
> < at random, and put it in the drive.
>
> I know if for LTO better than DLT.  As I understand it, when the drive
> rewinds the tape it then writes the known tape mark positions onto
> a special region at the beginning of the tape.  (Before the actual
> data storage starts.)  Then when loading a tape, this information is
> read back.  
>
> Given a command to seek to a specific tape mark, including EOT,
> the drive can go directly to the appropriate track and approximate
> tape position.  Then search for the actual tape mark on the tape.

Nice trick! But I thought you can't write to a tape before its logical
EOT. I guess with this, you can. I suppose it leaves a large gap
between the special area and the normal data storage.

> I presume that this wasn't done for DLT-I, but I don't know which
> version it was implmented on.  

Not on DLT-4 -- at least not on mine!

>
> -- glen

AEF