I had many VXA-1 v17 tapes act this way with amanda with 2 different 
internal VXA drives. I have also tried an older external drive that never 
showed the behavior. I have since replace the internal drive with VXA-2 
(basically, in order to get more capacity out of the tapes I'm reusing) 
and the problem has never manifested with the VXA-2 drive. 

The problem only ever occurred on the first file header; subsequent file 
headers were ok and I have done restores from the tapes, using 

        mt -f /dev/nst0 rewind
        mt -f /dev/nst0 fsf 2
        amrestore ....

I never did manage to dig to the bottom of the problem. It did not seem to 
be localized to specific tapes (overwriting a tape could sometimes make 
the problem go away)


On Mon, 24 Feb 2003 rwong AT mrs.mni.mcgill DOT ca wrote:

> Hum.  I thought the same originally as well.  But I have about a 
> dozen tapes reacting the same way.  Half are new and half are old.  I use 
> vxa-1 v17 tapes.  I find it unlikely that so many are defective/went bad 
> all together in one shot.  What does your experience say?
> 
> robin
> 
> 
> On Thu, 20 Feb 2003, Gene Heskett wrote:
> 
> > On Thu February 20 2003 17:50, rwong AT mrs.mni.mcgill DOT ca wrote:
> > >This is what I did at the shell...
> > >
> > ># Pop in a tape.
> > >% mt -f /dev/nst0 rewind
> > >% mt -f /dev/nst0 fsf 1
> > >% dd if=/dev/nst0 skip=1 bs=32k | tar tvf -
> > >dd: reading `/dev/nst0': Input/output error
> > >0+0 records in
> > >0+0 records out
> > >
> > ># Let's try another file marker on the same tape.
> > >% mt -f /dev/nst0 rewind
> > >% mt -f /dev/nst0 fsf 10   # I just want to look at the file
> > > header. % dd if=/dev/nst0 count=1 bs=32k
> > >AMANDA: FILE 20030213 host1 /boot lev 1 comp N program /bin/gtar
> > >To restore, position tape at start of file and run:
> > >        dd if=<tape> bs=32k skip=1 | /bin/gtar -f... -
> > >
> > >1+0 records in
> > >1+0 records out
> > >
> > >
> > >On the first dd attempt, this showed up in the log:
> > >Feb 20 17:30:24 host1 kernel: st0: Error with sense data: Info
> > > fld=0x8000, Current st09:00: sense key Medium Error Feb 20
> > > 17:30:24 host1 kernel: Additional sense indicates Recorded entity
> > > not found
> > >
> > >
> > >I can read the other file markers...except file marker 1.
> > >
> > >robin