Problem with SLR SCSI tape

Julian Stacey jhs at berklix.org
Wed Jun 28 09:13:00 UTC 2006


Reference:
> From:		Olivier Nicole <on at cs.ait.ac.th> 
> Date:		Wed, 28 Jun 2006 14:44:14 +0700 (ICT) 
> Message-id:	<200606280744.k5S7iEna073729 at banyan.cs.ait.ac.th> 

Olivier Nicole wrote:
> Hi,
> 
> I have had that tape drive and motherboard working hand in hand for
> many years without a problem.
> 
> Lately it started to give messages like:
> 
> on a amlabel command:
> Jun 27 13:02:10 amanda /kernel: (sa0:sym1:0:2:0): WRITE FILEMARKS. CDB: 10 0 0 0
>  2 0 
> Jun 27 13:02:10 amanda /kernel: (sa0:sym1:0:2:0): ILLEGAL REQUEST asc:50,0
> Jun 27 13:02:10 amanda /kernel: (sa0:sym1:0:2:0): Write append error
> Jun 27 13:02:10 amanda /kernel: (sa0:sym1:0:2:0): failed to write terminating fi
> lemark(s)
> Jun 27 13:02:10 amanda /kernel: (sa0:sym1:0:2:0): tape is now frozen- use an OFF
> LINE, REWIND or MTEOM command to clear this state.
> 
> 
> Jun 27 13:03:18 amanda /kernel: (sa0:sym1:0:2:0): ERASE. CDB: 19 1 0 0 0 0 
> Jun 27 13:03:18 amanda /kernel: (sa0:sym1:0:2:0): ILLEGAL REQUEST asc:2c,0
> Jun 27 13:03:18 amanda /kernel: (sa0:sym1:0:2:0): Command sequence error
> 
> on a dd command:
> Jun 28 13:48:04 amanda /kernel: (sa0:sym1:0:2:0): extraneous data discarded.
> Jun 28 13:48:04 amanda /kernel: (sa0:sym1:0:2:0): COMMAND FAILED (87 0 1).
> 
> But at same time I can do a tar without problem, I could write quite
> some amount of data during a backup at night...
> 
> Error are consistant whatever new or old tape I use in the drive.
> 
> Drive is SLR 100 from Tandberg
> sa0 at sym1 bus 0 target 2 lun 0
> sa0: <TANDBERG SLR100 0402> Removable Sequential Access SCSI-2 device 
> sa0: 40.000MB/s transfers (20.000MHz, offset 15, 16bit)
> 
> SCSI adapter is Symbios
> sym1: <896> port 0xb000-0xb0ff mem 0xf8000000-0xf8001fff,0xf8800000-0xf88003ff irq 10 at device 5.1 on pci1
> sym1: Symbios NVRAM, ID 7, Fast-40, LVD, parity checking
> sym1: open drain IRQ line driver, using on-chip SRAM
> sym1: using LOAD/STORE-based firmware.
> sym1: handling phase mismatch from SCRIPTS.
> 
> Nothing has been changed in hardware and operating system lately.

Just a long shot hardware guess:
  Maybe your hardware is not as innocent as you hope ?
  (my mainboard BIOS chip died last week, unusual things do happen :-)

  When did you last clean tape heads with isopropyl alcohol to remove
  dirt ?  is you .th (Thailand ?) area hot ? humid ? tapes decaying
  ?  Did you try demagnetising the heads ?  In an overnight run you
  wouldnt hear tape, but can you during day ?  Does it stream OK or
  does it seem to run rough, retrying as read head doesnt confirm
  write head ?  Just wild guesses as I dont know amanda software,
  but tar may be using different size blocking (tar & mt can set
  blocking) my QIC-525 also allowed vari sized blocking. Maybe
  different block sizes affect the chance of error recovery on the
  fly ? I've not calculated distance between write & read head, &
  compared to block length & wondered if drive internal electronics
  & tape data standard might let write head write a 2nd short
  correction block again, if bad block & short enough ?.

  All speculation, but if you really havent changed the software that
  used to work, I guess its some way hardware related, termination,
  CPU getting too hot in summer ? (check with /usr/ports/sysutils/xmbmon,
  or BIOS Health Status (though I have a board just gets hotter if
  left in BIOS, whereas I guess FreeBSD in idle loop goes into power
  saving & runs cooler, - so do you have more CPU load now ?).
  
  It can sometimes be non obvious, eg: once I had a  FreeBSD that
  always crashed accessing a CDROM; it was a (back then) high speed
  acceleration drive, sucked lots of power at start, & dipped the
  aged power supply, so it crashed.
Good luck

> Any help or direction for diagnostic is warmly welcome
> 
> Bests,
> 
> Olivier
> 
> _______________________________________________
> freebsd-scsi at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-scsi
> To unsubscribe, send any mail to "freebsd-scsi-unsubscribe at freebsd.org"
> 
> 

-- 
Julian Stacey.  Consultant Unix Net & Sys. Eng., Munich.  http://berklix.com
Mail in Ascii, HTML=spam.     Ihr Rauch = mein allergischer Kopfschmerz.


More information about the freebsd-scsi mailing list