SUMMARY: Using Sun 4mm DAT drive with SunOS 4.1.3C

From: Nick Sales (cyber@ee.ic.ac.uk)
Date: Wed Dec 14 1994 - 04:34:41 CST


I posted a question recently about getting the Sun 4mm DAT backup unit to
work with SunOS 4.1.3C (sun4m).

It _is_ perfectly possile to use the drive with 4.1.3C, but some minor
additions to the kernel are required, which will then necessitate
rebuilding. Two people - Woobin Lee, and Dick Gray - sent me the
information. (Thanks guys!)
As Woobin Lee notes, you need to change /sys/scsi/targets/stdef.h and
.../st_conf.c for things to work properly.

I also asked about correct tape length, density and blocking factors. Dick
Gray suggests 126 as the blocking factor, while Sun suggested 96. A recent
post said words to the effect that, since DATs and more modern tape drives
have a variable blocking size used mainly to keep the tape writing at full
speed instead of waiting for the disk to fill the I/O channels, blocking
size is more cosmetic than anything else. Since I don't know enough, I
would suggest either 96 or 126 as a blocking factor.

Tape length is really 295 feet, but as Dick Gray says, if length is used
with dump(8) and dump derivatives, just choose something big!
Density is an unbelievable 610000 (six hundred and ten thousand!) BPI, but
since this is only used by dump as part of its tape length calculations - I
looked at the source - you can again get away with practically anything.
My advice is to use values corresponding to the "correct" ones however.

Thanks for all the help.

Incidentally, the 4mm DAT drive will only run with in-line compression
under Solaris =>2.3. Sun sell an optional compressing device driver, for
250 UK Pounds, if you want to use in-line compression with SunOS 4.1.3.

The rest of the post is the relevant info about kernel mods, etc.
Nick

From: Woobin Lee <woobin@fizzle.ee.washington.edu>
--------------------------------------------------
You need to add the following lines to your kernel config files and then
rebuild the kernel:

To /sys/scsi/targets/stdef.h:

#define ST_TYPE_PYTHON 0x2c /* Archive Python DAT */

To /sys/scsi/targets/st_conf.c:

/* Archive Python 4mm 2GB drive */
{
        "Archive Python 4mm Helical Scan", 14, "ARCHIVE Python",
        ST_TYPE_PYTHON, 1024,
        (ST_VARIABLE | ST_BSF | ST_BSR | ST_LONG_ERASE),
        5000, 5000,
        { 0x00, 0x8C, 0x8C, 0x8C },
        { 0, 0, 0, 0 }
},

They are incorporated into SunOS 4.1.3_U1B.

-----------------------------------------------------------------
From: gray@dialog-semiconductor.co.uk (Dick Gray)
-------------------------------------------------

The dump command I use is:

        /etc/dump 0uctbsdf 1 126 5000 61000 /dev/nrst10 /dev/sd1h

so the blocking factor is 126 (This is the recommended figure as far as I can
tell). the other figures (5000 and 61000) are not really significant as long
as dump (or whatever) doesn't think it's got to the end too soon. It's best to
make sure they are too big, to ensure this doesn't happen.
-----------------------------------------------------------------

-- 
=============================================================================
Nick Sales                          |
cyber@ic.ac.uk                      |      No wife, no horse, no mustache
+44 71 594 6218                     |



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:09:17 CDT