>Hi,
>i have a big problem with dd in UW213 to copy out my dying disk :-((
>Still the disk is running, it is a 8GB IBM SCSI disk only one unix
>partition on it with 4 slices, all of the same size.
>As finding in the man pages i use "dd if=/dev/rdsk/c0d0t0p0
>of=/dev/rmt/ctape1 bs=64k". All slices are unmounted!
>But dd will only return for in/out: 8896+0.
That is not 2GB...
>The device is a IBM DNES-309170W SCSI 9GB 7200 U/min,
>prtvtoc shows 4 valid slices, each of size=4194304 with TAG=USR;
>heads=64, cyls=8748, sectors=32, bpsec=512 for the disk.
>So the dd must have ended without an error(!) not on the end of my
>whole unix partition and when i tried to restore it, on an identical
>disk on a second machine i did only get (a part of?) the first slice,
>needing a fsck!
If you make an image of the disk of a running box you will
always need to do a fsck when using the image the first time,
because the "dirty" bit is still set.
So it's wise to boot another OS.
>The other mount-points are unusable.
>Who can help me to save the still running disk to tape?
I would boot knoppix from CD, and would mount a network drive and
run dd_rescue to make an image to the new "drive" on the network.
If you insistis on using SCO-dd, please set the options to ignore
read errors and to fill them up on the destination drive.
Rainer