SUMMARY: Exabyte 8500C & Sol. 2.4

From: js@cctechnol.com
Date: Sat Jan 14 1995 - 23:57:58 CST


Thanks to the following:
pamela@jupiter.legato.com
jimm@csdc02.orl.mmc.com
jstubbs@utdallas.edu

My original post:

> I just upgraded my system to Solaris 2.4. Now I can't read tapes
> created with 2.3 on my Exabyte 8500C. I read the information that came
> with the upgrade about the large block size change. I've tried it
> either way to no avail. Has anyone been able to get this working with
> an 8500C?
>
> This is my st.conf entry for the 8500C
>
> tape-config-list =
> "EXABYTE\040EXB8500C", "Exabyte 8mm with compression", "EXB_8500C";
> EXB_8500C = 1,0x35,1024,0x1679,4,0x14,0x15,0x8c,0x8c,3;
>
>
> Johnie Stafford
> js@cctechnol.com
> System Administrator
> C & C Technologies, Inc.

Scott Stubbs suggested this:

Solaris 2.3 does NOT directly support the 8500C. One workaround is to
contact Artecon Corp. and ask for their PN 92-16090039 (Exabyte
filename 8ce-0189.bin). This loades EXB-8500XL emulation software
that will eliminate the problem. Solaris 2.4 cures the problem with
the patch 101899-02.

I found this unnecessary in my situation.

Pamela Pledger of Legato Systems, Inc. provided my ultimate solution
in this technical bulletin that her company puts out. I have attached
the bulletin to this message. The answer to my question was that due
to a bug in Solaris 2.3 the bsize value had to be set to 1024 instead
of 0 as the st(7) man page recommends. This bug has been fixed for 2.4
and setting a value of 1024 in the bsize field of the device
specification does exactly as it should do, set the block size to 1024
which means that anybody that needs a block size other than 1024 is
out of luck.

          Johnie Stafford
          js@cctechnol.com
          System Administrator
          C & C Technologies, Inc.

=================== Technical Bulletin 142.txt ====================
PURPOSE
This Technical Bulletin shows you how to set up your kernel to work
with tape devices not directly supported by Sun Microsystems.
This Bulletin includes files for SPARC Solaris 2.3 and SPARC Solaris
2.4.\x11 Use these files to add support for nominally unsupported tape
devices to your SPARC system.

INTRODUCTION
This section contains installation notes and a copy of the files to
reconfigure your SPARC (Solaris 2.3 or later) system for tape drives
not explicitly supported by Sun Microsystems.

SOLARIS 2.3
This information conforms to the recommendations outlined in the st
(7) man page, except that in order to avoid a Solaris 2.3 bug, the
bsize field for variable record length devices is not set to zero as
the man page recommends.

Solaris 2.3 File
Install the following as the file /kernel/drv/st.conf, and then do a
reconfiguration reboot (as root, halt followed by a boot -r):
#
# @(#)st.conf 1.18 94/11/18 Copyright (c) 1994, Legato Systems, Inc.
# Additional Copyright (c) 1994, Legato Systems, Inc.
#
#
# 1. This works only under Solaris 2.3 or later.
#
# 2. this file is intended to be installed as /kernel/drv/st.conf.
# Once installed, either do, as root,
#
# rem_drv st ; add_drv st
#
# or do a reconfiguration reboot (halt, boot -r).
#
# 2. Unqualified density devices for BSD behavior (for example,
# :b, or /dev/rmt/Nb),
# always selects the *middle* density, not the one you specify here. This
# means, effectively, that you always should specify the middle density
# for the default, and that if you have, say, an Exabyte 8500c/8505,
# you really cannot use the middle density for the 8200c like it would
# make sense to, in case someone actually *specifies* the fully qualified
# name (in this case, /dev/rmt/Nmb). This may be fixed in Solaris 2.4.
#
# 3. Do *not* clear the bsize field (3rd field of the numbers list
# below) for Solaris 2.3. There is another bug which causes the
# tape driver to hang if this is clear. I *know* that the st(7)
# man page says to leave this at 0 if you set the VARIABLE mode
# option in the flags word, but....
#
# On the other hand *DO* clear the bsize field (3rd field of the numbers
# list below) for Solaris 2.4. If you don't, then for reasons somewhat
# unclear, the blocksize for the tape device gets *set* to whatever
# this value is, and bad things happen (from an interoperability point
# of view). That is to say, under 2.4, a value in the bsize field seems
# to override the flag in the options field that says "Use VARIABLE
# length" mode.
#
# 4. Do *not* for Solaris 2.4 un-comment the tape-driver-buffering line.
# This *may* cause tape integrity problems with Networker.
#
# 5. The ARCHIVE Python below is for the non-Sun Python DAT drives- the
# ones usually used by Compaq. They don't understand some mode select
# values sent by the Sun driver, so it is best to avoid this entirely
# with a mode select entry for this particular drive. The Sun internal
# capture for ARCHIVE/Python should still then work for all entries
# not specified below.
#
#
#tape-driver-buffering = 8; tape-driver-buf-max-size = 262144;
tape-config-list =
      "Metrum", "Metrum VHS Cartridge", "metD",
      "ARCHIVE Python 25501", "DLI Python DAT drive", "DAT",
      "ARCHIVE 4586XX 28", "Compaq TurboDAT", "DAT",
      "ARCHIVE Python 28", "Archive TurboDAT", "DAT",
      "ARCHIVE IBM4586", "IBM DAT", "DAT",
      "WangDAT Model 2600", "WangDAT 4mm DAT", "DAT",
      "WangDAT Model 3400", "WangDAT 4mm DAT", "DAT",
      "SONY SDT-", "SONY 4mm DAT", "DAT",
      "HP C1553A ", "HP C1553A 4mm DAT", "DAT",
      "HP C1533A ", "HP C1533A 4mm DAT", "DAT",
      "HP HP35480A", "HP 35480A 4mm DAT", "DAT",
      "R-BYTE RB100", "R-Byte Whozits", "DAT",
      "CIPHER L860", "DLT (Cipher)", "DLT",
      "DEC TZ87", "DEC DLT", "DLT",
      "DEC DLT2", "DEC DLT", "DLT",
      "DEC DLT4", "DEC DLT", "DLT",
      "DEC TLZ6", "DEC Python Changer", "DAT",
      "DEC TLZ06", "DEC 4mm DAT", "DAT",
      "DEC TLZ07", "DEC 4mm DAT", "DAT",
      "EXABYTE EXB-2501", "Exabyte 2501 MiniQIC", "WtQIC",
      "EXABYTE EXB-2502", "Exabyte 2502 MiniQIC", "WtQIC",
      "EXABYTE EXB8500C", "Exabyte 8500c", "Exa8500c",
      "EXABYTE EXB-8505", "Exabyte 8505", "Exa8500c",
      "WANGTEK 5525ES SCSI", "Wangtek 525MB QIC", "WtQIC",
      "WANGTEK 51000 SCSI", "Wangtek 1.2GB QIC", "WtQIC",
      "WANGTEK 9500 DC", "Wangtek 5GB QIC", "WtQIC"
metD = 1,0x36,1024,0xd639,1,0x00,0
DAT = 1,0x34,1024,0xd639,1,0x00,0
DLT = 1,0x36,1024,0xd639,1,0x00,0
Exa8500c= 1,0x35,1024,0xd639,4,0x14,0x15,0x15,0x8c,1
WtQIC = 1,0x32, 512,0xc40a,1,0x00,0;

name="st" class="scsi" target=0 lun=0;
name="st" class="scsi" target=1 lun=0;
name="st" class="scsi" target=2 lun=0;
name="st" class="scsi" target=3 lun=0;
name="st" class="scsi" target=4 lun=0;
name="st" class="scsi" target=5 lun=0;
name="st" class="scsi" target=6 lun=0;

SOLARIS 2.4
This information conforms to the recommendations outlined in the st
(7) man page\x11 It is identical to the Solaris 2.3 example, except
that the bsize field for variable record length devices is set to zero
as the man page recommends.
Install the following as the file /kernel/drv/st.conf, and then do a
reconfiguration reboot (as root, halt followed by a boot -r):

#
# @(#)st.conf 1.18 94/11/18 Copyright (c) 1994, Legato Systems, Inc.
# Additional Copyright (c) 1994, Legato Systems, Inc.
#
#
# 1. This works only under Solaris 2.3 or later.
#
# 2. this file is intended to be installed as /kernel/drv/st.conf.
# Once installed, either do, as root,
#
# rem_drv st ; add_drv st
#
# or do a reconfiguration reboot (halt, boot -r).
#
# 2. Unqualified density devices for BSD behavior (for example,
# :b, or /dev/rmt/Nb),
# always selects the *middle* density, not the one you specify here. This
# means, effectively, that you always should specify the middle density
# for the default, and that if you have, say, an Exabyte 8500c/8505,
# you really cannot use the middle density for the 8200c like it would
# make sense to, in case someone actually *specifies* the fully qualified
# name (in this case, /dev/rmt/Nmb). This may be fixed in Solaris 2.4.
#
# 3. Do *not* clear the bsize field (3rd field of the numbers list
# below) for Solaris 2.3. There is another bug which causes the
# tape driver to hang if this is clear. I *know* that the st(7)
# man page says to leave this at 0 if you set the VARIABLE mode
# option in the flags word, but....
#
# On the other hand *DO* clear the bsize field (3rd field of the numbers
# list below) for Solaris 2.4. If you don't, then for reasons somewhat
# unclear, the blocksize for the tape device gets *set* to whatever
# this value is, and bad things happen (from an interoperability point
# of view). That is to say, under 2.4, a value in the bsize field seems
# to override the flag in the options field that says "Use VARIABLE
# length" mode.
#
# 4. Do *not* for Solaris 2.4 un-comment the tape-driver-buffering line.
# This *may* cause tape integrity problems with Networker.
#
# 5. The ARCHIVE Python below is for the non-Sun Python DAT drives- the
# ones usually used by Compaq. They don't understand some mode select
# values sent by the Sun driver, so it is best to avoid this entirely
# with a mode select entry for this particular drive. The Sun internal
# capture for ARCHIVE/Python should still then work for all entries
# not specified below.
#
#
#tape-driver-buffering = 8; tape-driver-buf-max-size = 262144;
tape-config-list =
      "Metrum", "Metrum VHS Cartridge", "metD",
      "ARCHIVE Python 25501", "DLI Python DAT drive", "DAT",
      "ARCHIVE 4586XX 28", "Compaq TurboDAT", "DAT",
      "ARCHIVE Python 28", "Archive TurboDAT", "DAT",
      "ARCHIVE IBM4586", "IBM DAT", "DAT",
      "WangDAT Model 2600", "WangDAT 4mm DAT", "DAT",
      "WangDAT Model 3400", "WangDAT 4mm DAT", "DAT",
      "SONY SDT-", "SONY 4mm DAT", "DAT",
      "HP C1553A ", "HP C1553A 4mm DAT", "DAT",
      "HP C1533A ", "HP C1533A 4mm DAT", "DAT",
      "HP HP35480A", "HP 35480A 4mm DAT", "DAT",
      "R-BYTE RB100", "R-Byte Whozits", "DAT",
      "CIPHER L860", "DLT (Cipher)", "DLT",
      "DEC TZ87", "DEC DLT", "DLT",
      "DEC DLT2", "DEC DLT", "DLT",
      "DEC DLT4", "DEC DLT", "DLT",
      "DEC TLZ6", "DEC Python Changer", "DAT",
      "DEC TLZ06", "DEC 4mm DAT", "DAT",
      "DEC TLZ07", "DEC 4mm DAT", "DAT",
      "EXABYTE EXB-2501", "Exabyte 2501 MiniQIC", "WtQIC",
      "EXABYTE EXB-2502", "Exabyte 2502 MiniQIC", "WtQIC",
      "EXABYTE EXB8500C", "Exabyte 8500c", "Exa8500c",
      "EXABYTE EXB-8505", "Exabyte 8505", "Exa8500c",
      "WANGTEK 5525ES SCSI", "Wangtek 525MB QIC", "WtQIC",
      "WANGTEK 51000 SCSI", "Wangtek 1.2GB QIC", "WtQIC",
      "WANGTEK 9500 DC", "Wangtek 5GB QIC", "WtQIC"
#
# clear the 3rd field below, except for fixed-length device WtQIC
#
metD = 1,0x36,0,0xd639,1,0x00,0
DAT = 1,0x34,0,0xd639,1,0x00,0
DLT = 1,0x36,0,0xd639,1,0x00,0
Exa8500c= 1,0x35,0,0xd639,4,0x14,0x15,0x15,0x8c,1
WtQIC = 1,0x32,512,0xc40a,1,0x00,0;

name="st" class="scsi" target=0 lun=0;
name="st" class="scsi" target=1 lun=0;
name="st" class="scsi" target=2 lun=0;
name="st" class="scsi" target=3 lun=0;
name="st" class="scsi" target=4 lun=0;
name="st" class="scsi" target=5 lun=0;
name="st" class="scsi" target=6 lun=0;

To request an electronic list of all Legato Technical Bulletins,
e-mail your request to request@Legato.COM with a subject line of send
bulletins index.\x11\x11Or download them from CompuServe (go
Legato).\x11\x11For a hard copy subscription, see Technical Bulletin
#025 and FAX your request to (415) 812-6034.\x11\x11
PRINTING HISTORY
First published 6/15/94
Updated 12/9/94
M.J.

TECHNICAL BULLETIN
142:\x11\x11Adding Unsupported Tape Devices to Solaris Systems (UNIX)



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:14 CDT