All,

Interesting development, was able to dump 3.2Gb of raw data from SCSI tape drive. Used the follow commands:

mt -f /dev/st0 setblk 0 (this sets the blocksize of the tape drive to variable on SCSI Tape drive Zero)

dd if=/dev/st0 of=(filename) bs=64k (This reads the tape, and buffers/writes it in 64k blocks.)

You can read the whole data stream on a clean tape by not including a count of the blocks desired. If you add count=xxx to the dd command you will dump to file that many 64k blocks. Turns out the key to the whole thing is the mis-match in block sizes. The drive needs to be set to a variable block size on startup in Mandrake. This is where the "mt" command comes into play. Once that is done, dd needs to know how much data to transfer each read sequence. We write data to the tape drive in the VME unit in 64k blocks, so it needs to read in that size. The confusion for me came in that the you can't just use 64k blocks. In case you are wondering, the drive being used is a Exebyte 89001 0-025 drive. I am hoping that with the cost of solid state drives, and reliability of hard drives we can replace them in these units like we have in others.
Just waiting on the capacity to get up there. We have a PCMCIA solution that handles 10Gb of data, but the Exebyte stores 40Gb.


Sorry to have troubled you all on this.

BTW, 9.1 has handled this wonderfully on my 1Ghz Powerbook, with the Radeon Mobility Video. Thanks Guys!!


Charlie Carroll



Reply via email to