jr@sesame.UUCP (Jim Rosenberg) (09/28/85)
I guess sometimes a person doesn't know when they've been spoiled. My first
experience with hard disk controllers on PC-type machines was a Maynard
controller. I had very good luck with it. The only problem was that a
program called the Connector, which runs Venix and DOS concurrently, appeared
to be sick as a dog when mixed with Maynard's XT ROM BIOS. Now I have a DTC
5150BX in a 6300. (A long story.) I understand the DTC is the controller AT&T
itself uses for 6300 hard disk options. The DTC has some *very* nice things I
like a lot. It will do alternate track mapping transparent to the OS, which
the Maynard wouldn't. Trouble is, it comes from DTC with NO SOFTWARE WHATEVER
-- I mean no diskette of any kind! Diagnostics? Head Park? On you own,
buddy.
Now here's the worst part. It has no program to tell it which tracks are
bad. I have the usual sticker with a list of a dozen bad tracks. I'm not
completely out of luck for low-level disk software. Unisource, the company
that sells Venix, sent me a diskette with a program called dmap.exe, written
by Terminals Unlimited SPECIFICALLY for the combo of DTC & Tulin. It will
mark the alternate tracks -- BUT!! Only the tracks *it finds itself*. When I
ran it (3 times) it always missed 4 bad tracks marked on the sticker. It has
no provision to type in head and cylinder numbers of bad tracks by hand. My
Maynard had a utility that would do that! Anybody know of a disk utility that
will do this? These bad tracks are nothing but a time bomb waiting to go off.
They could afflict any DTC controller user! Since DTC is the controller AT&T
uses for its own hard disk options, this is serious. Already after only 24
hours I am developing bad blocks because of this, and Venix (mostly Version
7ish, with Version 6 file system) has no way of dynamically handling bad
blocks. Haalp!
You unix wizards reading this note that mkfs's tries to detect bad blocks
are NOT GOOD ENOUGH. Hard disk manufacturers have very good equipment and can
detect marginal tracks that will pass mkfs. I know, because it's happening to
me!
I called DTC to complain about the lack of a program to mark bad blocks.
Mind, I'd even be happy if it just marked the blocks bad without setting up
alternate tracks. The response of the techie was "Oh that's for the
aftermarket." I was flabbergasted. A maker of disk controllers & hard disk
OEM (Teammate, or something) is telling me that marking bad blocks is for the
"aftermarket". I gave him a piece of my mind. I told him flat out that it
wasn't a disk controller he was selling but a time bomb.
Well, enough of ranting. Does anybody know of some celebrated citizen of
the aforementioned aftermarket who can give me a utility to mark these bad
blocks? Remember, it has to be low-level, outside DOS, since it has to work
for the Venix part of the disk. In a pinch, I hear there's a public domain
version of mkfs. For the Venix part of the disk I could maybe hack that to
ask for bad block numbers after it was all done. I assume Venix maps bad
blocks in inode 0, or somewhere -- I don't really know.
The controller is really very nice -- if only it had the right bundled
utility.
Any help appreciated. -Thanks,
Jim Rosenberg
Voice: (412) 785-2806
uucp: {decvax!idis,allegra}!pitt!amanuen!jr
CIS: 71515,124