[Gllug] [OTish] SureStore DAT24

Simon A. Boggis simon at dcs.qmul.ac.uk
Wed Jul 9 01:26:49 UTC 2003


Taylor James wrote:
> Well, my SCSI card arrived and I plugged in the tape drive and it worked,
> straight off! For a bit :-(
> 
> I can cat a file to and from /dev/st0 quite happily, I can do a 'mt -f
> /dev/st0 erase' and it erases the tape fine. When I try a 'tar cvpf /dev/st0
> /mnt/mp3' however, I get the following:

AFAIK you are "supposed" to use the non-rewinding device, /dev/nst0 
rather than st0, for this kind of job - try that too?

>     deathstar:/home/james# tar cvpf /dev/st0 /mnt/mp3/
[snip]
>     tar: /dev/st0: Wrote only 0 of 10240 bytes
                      ^^^^^^^^^^^^

So it actually didn't write anything ...
>     tar: Error is not recoverable: exiting now

> /var/log/messages gives the following:
> 
>     Jun 20 18:29:46 deathstar kernel: scsi : aborting 
>     command due to timeout : pid 35470, scsi0, channel 
>     0, id 3, lun 0 Write (6) 00 00 20 00 00
>     Jun 20 18:29:47 deathstar kernel: SCSI host 0 abort 
>     (pid 35470) timed out - resetting
>     Jun 20 18:29:47 deathstar kernel: SCSI bus is being 
>     reset for host 0 channel 0.
>     Jun 20 18:29:47 deathstar kernel: (scsi0:0:3:0) 
>     Synchronous at 10.0 Mbyte/sec, offset 15.
>     Jun 20 18:29:48 deathstar kernel: st0: Error with 
>     sense data: [valid=0] Info fld=0x0, Current st09:00: 
>     sense key Unit Attention
       ^^^^^^^^^^^^^^^^^^^^^^^^

... and given this message, I'm guessing at wrong tape device (see 
above) or failure to set up properly - try rewinding the tape before 
beginning:
   mt -f /dev/nst0 rewind
and checking the status - may take a little while to settle:
   mt -f /dev/nst0 status

I assume you've done the usual and carefully checked SCSI IDs and 
termination (some devices are capable of terminating the chain, so you 
don't want to do it twice or not at all)? Seems unlikely it is that from 
your messages, but always worth a second look.

> I have googled for this error, and found loads of people with the same
> problem but no solution. A lot of people have suggested setting the
> blocksize - I've done this on the drive and from the tar command to no
> avail. I've also tried turning compression on the drive off with 'mt -f
> /dev/st0 compression 0' which does nothing. Perhaps there's a different way

You can also do this through mt - see "setblk" and "setdensity", but 
normally higher level software takes care of that stuff for me (amanda),
so I have limited experience of those.

> to turn compression off with this paticular drive?

mt -f /dev/nst0 datcompress 0

is how I do it, but I think your way is probably OK too if it gives no 
error.

> Hopefully, the thing just needs a good clean, but until my cleaning tape
> arrives I can't try that.

Usually, there is an indication on the front panel when the drive 
requires cleaning, but as you say, worth a try.

The fact you can erase and such is encouraging - I'd keep fiddling - 
best of luck.

Simon


-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug




More information about the GLLUG mailing list