Re: Fwd: Multi tape problems with cpio

[Tape Help <tape.help@gmail.com>]

On Sun, 16 Jan 2005 20:47:55 +0200 (EET), Kai Makisara
<Kai.Makisara@kolumbus.fi> wrote:
> On Sat, 15 Jan 2005, Tape Help wrote:
> 
> > No one on linux-tape has responded.  I think that list is dead.
> > Can anyone help?
> > 
> linux-tape may not be dead but you have to give us some time ;-)
> 
> > Thanks.
> >
> > ---------- Forwarded message ----------
> > From: Tape Help <tape.help@gmail.com>
> > Date: Fri, 14 Jan 2005 02:53:16 -0500
> > Subject: Multi tape problems with cpio
> > To: linux-tape@vger.kernel.org
> >
> >
> > I am using cpio to backup a system.  The backup does not fit on 1
> > tape, so cpio prompts for another tape.  I change the tape and hit
> > enter.  cpio continues and completes.  cpio does eject the first tape,
> > which is handy.
> >
> > When I try to list the tape, cpio gets an I/O error at the end of the
> > first tape and does not prompt for a second tape.  Same problem if I
> > try to restore data.
> >
> > cpio: read error: Input/output error
> >
> > I have tried 2 different tape drives.
> > Both give the same results.
> > I have done this test about 5 or so times.
> >
> > The backup command:
> > find MyFS -depth|cpio -o --format=newc --block-size=128 -F /dev/st0
> >
> > The command to list the tape(s):
> > cpio -i --only-verify-crc --list --block-size=128< /dev/st0
> >
> > If I use dd to read the tape, it also gets a read error at the end.  I
> > think it should get an end of file and exit cleanly.
> >
> > I am using kernel 2.4.28.  It also fails with 2.4.20-8.
> > 
> I tested cpio and dd in my system with 2.4.28 without finding any problems
> (both variable block mode and fixed block mode with 1024 byte blocks).
> Some additional information may help to solve your problem:
> - Fixed or variable block mode? What blocksize if fixed block mode?
>  (Output of 'mt status' tells these.)
> - Tape drive and SCSI adapter make and model
> - Any tape messages in system log?
> - Debugging output from st. (You can enable debugging in st by editing the
>  driver source (change '#define DEBUG 0' to '#define DEBUG 1') and
>  recompiling the kernel/module.
> --
> Kai
> 

Sorry if I insulted the linux tape list!
I monitored it for almost 1 week, no posts.
The archive has been down for 4+ days.
http://www.geocrawler.com/lists/3/Linux/
The IP pings, but port 80 is closed.
My post also had no response.

Glad it is still there!

Fixed or varible?  I used "--block-size=128"  which gives 64K blocks.
I assume these would be fixed, but not sure.  I think fixed would be a
hardware option.  I have not used any fixed hardware options.  I can't
see that mt satus gives this info.  But output is below.

st0 is an HP LTO-1.
st1 is a DLT-4000 in a 5 slot mini changer (Quantum DLT 4500).

st0 is connected to a SCSI port on the motherboard.  AIC-7860
This is an ultra wide card.  The tape drive is in 20MB/sec mode.

st1 is an Adaptec card, 50 pin, ultra turnned off.  AIC-7880U

>From /var/log/messages (in not order)
kernel: Attached scsi tape st0 at scsi1, channel 0, id 2, lun 0
kernel: Attached scsi tape st1 at scsi4, channel 0, id 5, lun 0
kernel: st0: Block limits 1 - 16777215 bytes.
kernel: st1: Block limits 1 - 16777215 bytes.

>From /var/log/dmesg:
st0:
(scsi1:A:2): 20.000MB/s transfers (10.000MHz, offset 8, 16bit)
st1:
(scsi4:A:5): 10.000MB/s transfers (10.000MHz, offset 15)


# mt -f /dev/st0 status
SCSI 2 tape drive:
File number=0, block number=0, partition=0.
Tape block size 0 bytes. Density code 0x40 (DLT1 40 GB, or Ultrium).
Soft error count since last status=0
General status bits on (41010000):
 BOT ONLINE IM_REP_EN

# mt -f /dev/st1 status
SCSI 2 tape drive:
File number=0, block number=286270, partition=0.
Tape block size 0 bytes. Density code 0x1a (DLT 20GB).
Soft error count since last status=0
General status bits on (21010000):
 EOT ONLINE IM_REP_EN

I will attempt '#define DEBUG 1'.  I have never done any custom Kernel
stuff.  Or generated debug output before.  My current Kernel was built
by me, so I do stand a chance! :)

Thanks.