SUMMARY: Sun ST drivers

From: Jean Goodrich (jean@noao.edu)
Date: Wed Apr 21 1993 - 15:37:32 CDT


Much thanks to all who responded. My original post was as follows:
 
I am using the Sun ST drivers on a Sparc 10/41 and SunOS 4.1.3.
  
I need to track read/write retry statistics for 8mm (exabyte)
tapes running on my exabyte drives.
   
I have been printing out the mtget structure for MTIOCGET as defined
in sys/mtio.h. There are two fields in this structure reporting the
values in registers named mt_dsreg and mt_erreg, which are "drive
status" and "error" status registers, respectively.
    
I have taken several tapes, read several blocks of data from each
(with no rewind), and then printed out these statistics. Each time,
the values in these registers were zero. I find it very hard to believe,
given the age and conditions of some of the tapes I used to test this,
that there would be *no* read retries whatsoever.
     
Is anyone out there familiar with the Sun ST drivers and this mtget
structure? Am I looking at the correct registers for the numbers I'm
seeking? Are these error values bogus, or is this a known bug?
      

      The answers were appreciated but discouraging.

      The Sun ST drivers do not return the read or write retry count. The
      MTIOGET structure returns some error, but this error count is the
      number of driver detected errors, not tape drive detected errors,
      and thus not the read/write retry count.

      The same phenomenon can be found when doing the mt -f /dev/rstxx status
      command. The read retries field always comes back 0 or empty.

      There are rumors that this ability was available in SunOS 4.1.1. It
      may or may not be available in the Apunix drivers. It is definitely
      available through hardware, if you purchase a tape drive with the
      LED panel on the front that displays these numbers. There may also
      exist third party software that makes this info available on-line,
      though I haven't had any luck myself in finding what I need.

      My only solution so far has been to use the hardware. The Apunix drivers
      provide a software solution but are not compatible for other reasons
      with the requirements of my project.

      As I said, thanks to:

      guy@auspex.com Guy Harris
      eckhard@ts.go.dlr.de Eckhard Rueggeberg
      don@mars.dgrc.doc.c Donald McLachlan
      wolfgang@wsrcc.com Wolfgang Rupprecht
      ted@ssl.berkeley.edu Ted Rodriguez-Bell

Jean Goodrich
National Solar Observatory
GONG



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:45 CDT