Path: utzoo!attcan!uunet!husc6!think!ames!killer!tness7!tness1!sugar!karl
From: karl@sugar.UUCP (Karl Lehenbauer)
Newsgroups: comp.unix.microport
Subject: Microport System V/386 install woes
Message-ID: <1984@sugar.UUCP>
Date: 14 May 88 12:53:40 GMT
Organization: Sugar Land UNIX - Houston, TX
Lines: 59

I'm helping a friend of mine who owns a computer store try to install $1200
worth of Microport software on a 386/20 clone.  It has three meg of RAM
and an 80 meg Seagate drive.

First off, the surface analysis program doesn't work.  It bombs.  This is
mentioned in the install documentation, but the install script (autobooted
this time, unlike on the 286) valiantly tries to run it anyway.  Without
the surface analysis, this means we *must* type in all the defects by
hand (right?), and that we'll find any new defects by having the system 
screw up rather than having the surface analysis program find it.  If this
is correct, it is a major fubar.

The install docs say that you'll have to convert byte offsets on error
reports over to sectors, tho' they provide a chart for converting sector
numbers to byte offsets and the "enter bad tracks" program asks for byte
offsets.  I assume this is a bug in the documentation and they really
want byte offsets.

Next, the first Seagate 80 meg that we tried had an extraordinarily large
number of bad spots on the disk, like about 100.  After entering almost
all of them, the program bombs back to the root prompt saying "too many
bad tracks."  Boo, hiss.  We had to use a different drive.

Also, has anyone else noticed how much more user friendly DOS install, 
format, etc. programs are than Microport Unix ones?  The Sys V/386 format 
program appears to print a period for every four cylinders that it formats.
(That's based on observation; there's no mention of it in the docs.)_
The DOS format program cursor addresses to continually update, numerically, 
which cylinder and head is being formatted.  It helps one's confidence, 
particularly when the format program seems to slowly drive the heads back to 
track zero after every cylinder, making it seem like something is wrong
since it is constantly reseeking.

Further, you get these dire messages that go along the line of "WARNING!
Couldn't read alternate VTOC.  Couldn't install the boot block!"
Again, the docs say "don't worry about it", but I think it's bullshit
that messages like this be sent if they don't indicate a problem.

The fact is, installing Microport Unix is still for gurus only (unless
my experiences have been very rare rather than, as I suspect, pretty common)
and, although some people at Microport are apparently getting rich off it, 
neither will it be the "next thing" nor will they fix the bugs that have
been killing my system since 9/1986.

To conclude, Sys V/386 install procedures seem to be as painful as Sys
V/AT ones.  To make matters worse, on the AT version they at least put
enough utilities on the boot disk so you can look around and try running
some stuff by hand (divvy, mkfs, etc.) in the process of getting the
system going.  Although this is conceivable on Sys V/386, the lack
of "ls" on the boot disk makes it pretty much impossible.  Meanwhile,
my buddy's not too pleased either.  He had been hoping to sell Unix with
DOSmerge to power users as an alternative to OS/2.  Seeing how much trouble
we've been having, though, he's not too hot on the idea of having to
go through this install procedure very often.  Oh well, he's got a 30 day
return on the software.  I'm going back to the store today to try to finish
installing it on a new Seagate 80 that has fewer errors.  We'll see.
-- 
"Now here's something you're really going to like!" -- Rocket J. Squirrel
..!{bellcore!tness1,uunet!nuchat}!sugar!karl, Unix BBS (713) 438-5018