IWETHEY v. 0.3.0 | TODO
1,095 registered users | 0 active users | 0 LpH | Statistics
Login | Create New User
IWETHEY Banner

Welcome to IWETHEY!

New Tape operation on Solaris
I want to construct by own header file for each
tape, which should be rewritable without harming
the rest of it. Is there anyway to reserve
10MB at the begginning of each tape, and do
rewrites?

When I try, I always end up destroying the the
true data file, using various dd options such
as conv=notrunc, etc.
New Write a 10meg file to /dev/rmt
write all your other files to /dev/nrmt(no rewind mag tape.) when you change the header just make sure that you use the rmt device and dont overwrite the boundary.
thanx,
bill
TAM ARIS QUAM ARMIPOTENS
New I did. I think
do a 1k file:

dd if=y of=/dev/rmt/0n bs=1k
dd if=data of=/dev/rmt/0n bs=32k
mt -f /dev/rmt/0 rewind

dd if=/dev/rmt/0n of=y2 bs=1k
dd of=data.2 if=/dev/rmt/0n bs=32k
mt -f /dev/rmt/0 rewind

Works find, rereads the data as expected.

dd if=n of=/dev/rmt/0n bs=1k
mt -f /dev/rmt/0 rewind

dd if=/dev/rmt/0n of=n2 bs=1k
dd of=data.3 if=/dev/rmt/0n bs=32k

'n2' comes back, matching 'n', but data.3
fails since the EOT was set to be at the
end of 'n'.

New Couple of ideas
1. Hardware compression? I remember SunOS has numbers tacked-onto the dev names to control this, but I've seen inconsistency between the device numbers and whatever the lights on the drive say it's doing for compression. I trust the lights. Lights is all Spock had.

2. Rather than try to read your data file sequentially after reading the header file, have you tried walking the tape forward with "mt fsf" to see if you can get to the data file? I would try it before rewriting the header file first just to see if it does what you expect.
-Mike
     Tape operation on Solaris - (broomberg) - (3)
         Write a 10meg file to /dev/rmt - (boxley) - (2)
             I did. I think - (broomberg) - (1)
                 Couple of ideas - (morganek)

Why isn't there a shower in other dimensions?
38 ms