Wolfgang Weisselberg wrote:
>
> Hi, Jim!
>
> Trying to kill the keyboard, Jim Rankin ([EMAIL PROTECTED])
> produced 6,4K in 143 lines:
>
> > I'm executing
> > gunzip -c disktarfile.tgz | dd of=/dev/ftape obs=10k
> > where I have used mt to setblk 0 on rft0 first. Yes, ftape points to
> > rft0.
>
> ftmt setblk 0 is probably only necessary if you do not pad
> the data. dd can do this, if you tell it to (using (ibs or bs)
> and sync). On the other hand, with setblk 0 you need not to
> use dd at all, IIRC. (Can't test it, don't have the hardware
> any more.)
I haven't tried dd with sync yet, and that's worth a shot. Thanks.
>
> > It starts like this when the file is basically all written (I think):
>
> > Dec 5 04:33:10 jrankin kernel: [054] ftape-io.c
> > (ftape_report_error_R999edffd) - errorcode: 39.
>
> ./Doc/html/qic117.h says (and so does ./include/linux/qic117.h):
> /*39*/ { "EOT/BOT System Failure", 1, },\
>
> > Dec 5 04:33:10 jrankin kernel: [056] ftape-io.c
> > (ftape_report_error_R999edffd) - errorcode: 20.
>
> same place:
> /*20*/ { "Self-Diagnostic Failed (cannot be cleared)", 1, },\
>
> > (ftape_wait_segment) - Unknown error.
> > (ftape_start_tape) - panic: location not known.
>
> Guess you have a flaky drive, then, these errorcodes from the
> tape drive are fatal.
After this thread was well under way, I used ftape 2.08 to write the
same archive directly from tar to tape under kernel 2.0.36. It worked
fine. Whatever is going on appears to be new under ftape 3.04, OR
related to use of dd, since I got the same error twice at the same point
with that combination under kernel 2.2.12.
I did find the following comment in the ftape HOWTO under Iomega Ditto
Tape Insider 3200 (my drive):
"Another problem has been reported with writing archives (with dd) to
the tape. It may start fine, but when the driver catches up with dd, it
stops the tape and rewinds it to the beginning. Then it starts winding
on through the tape ad infinitum. It appears to occur when the driver
asks the tape to pause which should cause the tape to move back by 3
segments, but instead is moves back to the beginning of the tape. A bug
fix submitted is reported to not solve the problem."
I will try some other things, maybe including 4.03 eventually. Thanks.
Jim
