Conference decwet::windows-nt

5651.0. "NT4.0 TLZ09 backup problem using hardware compression -- "Unrecognizable data" during verify pass" by CERN::CARNEY (Steve Carney) Fri Feb 07 1997 06:13

When I select verification and hardware compression using the NT4.0
Backup window interface, I get an "Unrecognizable data" error during the
verification pass.  I'm backing up 3 partitions (system and load
partitions included) to a 90m TLZ09 DAT.  Is there a fix, or has anyone
else noticed this problem?

When I do NOT select hardware compression, the verification pass
completes normally.  For now, I'm doing backups without hardware
compression.

When I do select verification and hardware compression, somewhere
between 30-50% of the backups fail during the verification pass
with pop-up window:

  Tape device reported an error on a request to read data from tape.
    Error reported:
    Unrecognizable data.

The corresponding entry in the event log is:
  The device, \Device\Tape0, has a bad block

On a couple of occasions there has also been pop-up window:
  The instruction "0x021acc2c" referenced memory at "0x0000001c".
  The memory could not be "read".

At least once, I've also seen popup-window:
  The instruction "0x021acbbc" referenced memory at "0x00000010".
  The memory could not be "read".

The TLZ09 drive is essentially new (Oct96).  Despite its newness, I used
the cleaning tape, but that didn't change the behavior.

I have tried doing backups using 4 different 90m DATs (2 Digital, 2
Sony).  I get the same error for all tapes; the problem does not seem to
be the tape.

I also swapped the TLZ09 with another one with the same firmware
rev. level.  I haven't checked, but the firmware should be reasonably
new because the replacement TLZ09 arrived in Jan97.  Using the
replacement TLZ09, I got the same "Unrecognizable data" results, on my
1st backup attempt, so I replaced the TLZ09 with the original one.  The
problem does not seem to be intrinsic to the original TLZ09 drive.

There are some disks on the same bus as the DAT drive, but they were not
active while trying to perform the backup.  In all cases, I was backing
up partitions from disks on other SCSI buses to the TLZ09:

Configuration information:

  Machine: Rawhide, 2 400 MHz processors, 2GB memory
  OS: NT4.0 Server SP2 (yes, I know it's not qualified on Rawhide)

  SCSI controllers: KZPSA-BB for all backup disks and the TLZ09
       TLZ09 StorageWorks shelf (BA36R-JA) configuration

       SCSI ID  Device
       -------  ------
       0        DWZZB-VW
       1        TLZ09-VA
       2        empty
       3-6      RZ29B-VW

  TLZ09 Firmware: (c)DEC0165 (from AlphaBIOS console V5.28-1)
  TLZ09 Switch settings: 
    1-3 Off
    4-6 On
    7-8 Off

Thanks,
        Steve

I have the same problem like .0 .

Since our system uses more diskspace (>4GB) we changed the backup-procedure
to enable hardware compression. We use 120m tapes.  
We get errors in the event viewer (system):
      4mmsony--TLZ09, /Device/Tape has a bad block. 
   We also see symb810 error messages. 

We changed the drive and loaded the new firmware revision 169 "A03", 
both without success.

Configuration : Alphaserver 1000, Windows NT 4.0 SP2 

Is there a solution for this problem ?


Thanks in advance

Reiner Mueller /MCS Munich 











                                                                   

    We're going to need a lot more data on the nature of the error (sense
    codes, where is it, same place every time, etc, etc) to have any chance
    of isolating this. I know of no such known problems.
    
    Please formally escalate this if you really want a solution, so someone
    will get assigned to fix it. It won't happen otherwise.
    
    I'm only consulting for the tape group these days... And the TLZ09
    engineers/support folks do not read this notes file regularly.
    
    Chris