🚀 go-pugleaf

rab@bevsun.bev.lbl.gov (Robert Belshe) writes:
  |
  |Has anyone noticed a loss of data capacity when writing
  |model 8500 exabytes at the optional low density?
  |
  |We have noticed that when we write a full tape from an application
  |which writes 8kbyte records at a data rate of around 50kbytes/second
  |that the tape fills up at approx 1 gbyte, instead of the expected 
  |2 gbytes.
  |
  |A test program which writes data at 250kbytes/sec is able to
  |get a full 2 gbytes on a tape.
  |
  |My guess is that this is caused by the tape being unable to stay
  |in the streaming mode.
  |
  |The drive is a TTI model 8510, running on a uVax 3100 thru a
  |TD-systems Viking SCSI controller.
  |
  |The folks at TTI say they never heard of this problem.

My guess is your guess is correct. I was told that on 8200s, 8 KBytes
of space (one turn of the head) is wasted each time the 256 KB buffer
goes empty. Depending on how tape motion is triggered as the buffer
fills, this might explain the 50% capacity reduction you're seeing.
If you can, try increasing the block size written from 8K to, say, 32K
and see if tape capacity goes up proportionately.

----------------------------------
Ed Kubaitis (ejk@ux2.cso.uiuc.edu)
Computing & Communications Services Office - University of Illinois, Urbana

In article <27204@dog.ee.lbl.gov> rab@bevsun.bev.lbl.gov (Robert Belshe) writes:
>
>Has anyone noticed a loss of data capacity when writing
>model 8500 exabytes at the optional low density?
>
>We have noticed that when we write a full tape from an application
>which writes 8kbyte records at a data rate of around 50kbytes/second
>that the tape fills up at approx 1 gbyte, instead of the expected 
>2 gbytes.
>
>A test program which writes data at 250kbytes/sec is able to
>get a full 2 gbytes on a tape.
>
>My guess is that this is caused by the tape being unable to stay
>in the streaming mode.
 Yes, the thing has some state mode where it writes an ignored
 on input track rather than stop streaming.

 I don't have the specifications around anymore, so I don't
 have the details.

 I'm surprised it doesn't work the same way at the higher density.

-- 
Phone: (408) 765-4277  FAX: (408) 765-5719
Mail Stop RN4-36; Pole L10; 2200 Mission College Blvd.,
P.O. Box 58119, Santa Clara, CA 95052-8119
mfineman@scdt.intel.com