Exabyte 8505 and 10 Gb capacity

2007-12-25 8:46:00

  Sun-Managers:

  This was a case of RTFM, but not realizing that compression required

  a special option to enable.

  Chuck Foley said it best:

> You should be using /dev/rmt/0c to write with compression

> turned on....The man page (Solaris 2.3) for "st" indicates this

> (near the end, there is a chart of density specifications and

> the density actually written for different drive types.

  -- Bob

  Thanks to:

  szgyula@skysrv.pha.jhu.edu (Gyula P. Szokoly)

  jpl@allegra.att.com (John P. Linderman)

  cfoley@arsenic.cray.com (Chuck Foley)

---------------------------------------------------------------------------

  My original Posting:

  Sun-managers,

  There has been a lot of discussion about tape drives of late, but

  now I have a query of my own. We have two Exabyte 8505 8mm tape

  drives connected to a SPARCserver 1000 (6 processors, Solaris 2.3)

  which are claimed to have a ~10 Gb capacity using built-in hardware

  compression. However, we have not been able to store more than

  ~5 Gb per tape (this makes no difference whether the tape is new

  or old [I've seen the report about tape drives using the lower

  density instead of the higher one for used tapes, but we're seeing

  this behavior with new tapes as well!]).

  A 'probe-scsi-all' shows the drives as '8505.....', but we see

  8500 behavior. We are using '/dev/rmt/0h' (not 'l' or 'm'). Has

  anyone seen/solved this problem?

  Thanks in advance, and of course I'll summarize.

  -- Bob

    ____________________________________________________________

      ____ ____ ____

     / \ | / \ |/ \ Atmospheric & Environmental

    / \| /______\ | Research, Inc.

    \ /| \ | 840 Memorial Drive

     \____/ | \____/ | Cambridge, MA 02139 USA

      

     Robert D. Worsham (Bob) voice: (617) 547-6207

     email: worsham@aer.com fax: (617) 661-6479

    ____________________________________________________________

Comments

Got something to say?

You must be logged in to post a comment.