DLT tape errors -- UPDATE

Bill Gurley bgurley at utk.edu
Wed Aug 6 09:30:01 CDT 2003


Steve_Boley at Dell.com wrote:
> Flash the firmware on the 39160 to latest available and go to 
> scsi settings and disable ultra wide negotiation on the id that
> the tape unit is running on.  This will obviate some problems with
> scsi noise on the bus and will actually increase speed on your
> backup as well.  Also specify -b 128 switch and will set your
> blocksize to 64k and will also cause the backup to run faster.
> This should help prevent the sensekey errors you are having on
> the system as well.  Make sure your raid is 2.27 firmware also.
> Steve


Thank you very much for your suggestions, Steve.

I flashed the firmware of the 39160 to the latest version (v3.10, A03).

I disabled wide negotiation for ID 6, which is where the tape drive is 
connected.

I verified that the RAID firmware is 2.27.

I tried a new backup with the command:  tar cvbf 128 /dev/nst0 *

...from the / directory.

It fails with the exact same errors as before, after writing a couple 
dozen files in the 'bin' directory.

I tried again with out the -b switch, and it does the same thing.


Any other suggestions, please??



> -----Original Message-----
> From: Bill Gurley [mailto:bgurley at utk.edu]
> Sent: Tuesday, August 05, 2003 1:15 PM
> To: linux-poweredge at dell.com
> Subject: Re: DLT tape errors -- UPDATE
> 
> 
> An update on my DLT problems...
> 
> I finally got desperate and blew away my RH 9 installation.  I formatted 
> partitions and did a quick-and-dirty RH 8 install, using the default RH 
> kernel.
> 
> I just now did a 'tar cvf /dev/nst0 *' from the / directory.
> 
> I THOUGHT it was actually going to work, because it went for a long 
> time;  longer than before.  But eventually it again stopped with:
> 
> tar: /dev/nst0: Wrote only 0 of 10240 bytes
> tar: Error is not recoverable: exiting now
> 
> 
> And syslog (again) reports:
> 
> Aug  5 13:13:20 nucleus kernel: st0: Error with sense data: Info 
> fld=0x0, Deferred st09:00: sense key Medium Error
> Aug  5 13:13:20 nucleus kernel: st0: Error with sense data: Info 
> fld=0x1, Current st09:00: sense key Medium Error
> Aug  5 13:13:20 nucleus kernel: Additional sense indicates Write error
> Aug  5 13:13:20 nucleus kernel: st0: Error on write filemark.
> 
> 
> Ultimately, I intend to use Amanda for backups, but I haven't set that 
> up yet.  It seems that this simple tar backup should work.  I just did a 
> test backup of a differrent RH8 (non-RAID) machine with an external HP 
> Surestor DAT drive attached to an Adaptec controller.  I used the same 
> command 'tar cvf /dev/nst0 *'.  It completed with only an expected error 
> because I neglected to exclude /proc.  The backup on that machine 
> completed, and I can do 'tar -tf /dev/nst0 > test.txt' to generate a 
> list from the tape.
> 
> Do any of the Dell folks have a suggestion for me?  I have just had this 
> machine a few days.  I'm worried that if I call Dell support on this, 
> they will want to talk in terms of Windows backup software, which I 
> don't have (and don't want).
> 
> Thanks for all of your suggestions.
> 
> 
> 
> Bill Gurley wrote:
> 
>>Hello Folks:
>>
>>I'm new to this list.  I have a brand new PowerEdge 2600 set up with 
>>RAID5 and a PowerVault 110T DLT VS 80 internal tape drive.
>>
>>I have set up Redhat 9, and wanted to backup some stuff to tape.  I did 
>>one little test backup of /tmp, but that's about all that has worked. 
>>Every time I try to back up using tar, it begins to work, but then fails.
>>
>>Here is some info:
>>
>>The tape drive is the sole device on an Adaptec AHA-3960D controller.
>>
>>'mt status' gives:
>>
>>[root at nucleus root]# mt 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
>>
>>
>>Here is a typical command I've been using:
>>
>>tar cvf /dev/st0 etc
>>
>>
>>With this command, the contents of /etc begin rolling across the screen, 
>>until:
>>
>>tar: /dev/st0: Wrote only 0 of 10240 bytes
>>tar: Error is not recoverable: exiting now
>>
>>
>>Syslog reveals:
>>
>>Aug  4 17:02:22 nucleus kernel: st0: Error with sense data: Info 
>>fld=0x0, Deferred st09:00: sense key Medium Error
>>Aug  4 17:02:22 nucleus kernel: st0: Error with sense data: Info 
>>fld=0x1, Current st09:00: sense key Medium Error
>>Aug  4 17:02:22 nucleus kernel: Additional sense indicates Write error
>>Aug  4 17:02:22 nucleus kernel: st0: Error on write filemark.
>>
>>
>>This leaves the tape drive flashing its "Clean/Media" LED, which clears 
>>after ejecting the tape and re-inserting it or another tape.  Yes, I 
>>have tried using the cleaning cartridge.  (Remember:  This is a BRAND 
>>new machine, but I cleaned it anyway.)  I have tried three different 
>>brand new tapes.
>>
>>What gives?  I have tried forcing various block sizes with the tar 
>>command, and it doesn't help.
>>
>>
>>Thanks for any assistance,
>>
>>
>>-Bill-
>>
>>-------------------------------------------------------------
>> Bill Gurley, Technical Director  |
>> Department of Chemistry          |    Consider Linux and
>> Univ. of Tennessee, Knoxville    |    Open Source Software!
>>
> 
> 
> 


-- 
-Bill-

-------------------------------------------------------------
  Bill Gurley, Technical Director  |
  Department of Chemistry          |    Consider Linux and
  Univ. of Tennessee, Knoxville    |    Open Source Software!




More information about the Linux-PowerEdge mailing list