Why does the W2000 installation go to the E: drive?

  • Thread starter Thread starter Bo Berglund
  • Start date Start date
B

Bo Berglund

Very strange,
I have two hard disks on our test machine, one big 50 Gb one (FAT32)
to hold the Ghost images of all the operating systems we need to test
with and another (Master drive, 9 Gb) where the operating systems are
to be installed to.
When I run the CD boot setup for a W2000 version and I get to the disk
selection stage I select to delete the existing partition on the
master 9Gb disk. Then I select this for installation of the new
operating system.
When all is done it turns out that Windows has been installed to the
E: drive! I want it to be on the C: of course.
But if I instead select to create a new partition on the empty disk
before selecting to continue the installation, then it finishes as an
installation to the C: drive (and my FAT32 disk is now E:....

What is the reasoning behind this and why does it happen?
Are you supposed to delete existing partition and then create a new
one manually in order to get the system drive to become C:????

/Bo
 
Put very briefly, this has to do with a residual memory of partition
names, and the installer choosing the 'next available' name. If the
system during a session remembers any partition is C: (and it will
instantly reassign names when a partition is deleted), then the name C:
is unavailable.

You can think of this as a modest pain in the neck. The workaround most
people here suggest is to abort then restart the install process
immediately after deleting/recreating/formatting the partition you want
the system in. This interrupts the session and names are reassigned as
you'd expect.

It would be nice (not sure it's trivial, or even possible) to have a
quick "Do you want this partition to be named C:" option during an install.
 
Put very briefly, this has to do with a residual memory of partition
names, and the installer choosing the 'next available' name. If the
system during a session remembers any partition is C: (and it will
instantly reassign names when a partition is deleted), then the name C:
is unavailable.

You can think of this as a modest pain in the neck. The workaround most
people here suggest is to abort then restart the install process
immediately after deleting/recreating/formatting the partition you want
the system in. This interrupts the session and names are reassigned as
you'd expect.

It would be nice (not sure it's trivial, or even possible) to have a
quick "Do you want this partition to be named C:" option during an install.
Click to expand...

But my experience is that if I do it this way then I get the install
finally to C:
- Start from CD
- Select to erase existing partition
- Select to create a new partition on the selected drive
- Then continue installation

Result: The system drive is C: like I want.

But if I do this instead:
- Start from CD
- Select to erase existing partition
- Choose to install to selected *now nonpartitioned* drive, this makes
setup create the partition and format it to NTFS
- Then continue installation

Result in this case: The system is on E: :-(

Last week I have been working to create a test PC with many different
versions of Windows available as Ghost images for our software
delivery testing. Late Friday I struggled with Japanese W2000 setup
without being able to read Japanese....
So I missed the prompt in the crucial point and got the final install
to E:....
Have to redo it on Monday, not fun.

/Bo
 
Bo - If I go further into specific drive letter assignment scenarios in
NT-class MS systems I'll probably get things badly scrambled, get beaten
up politely in this newsgroup, and possibly even have to be sedated and
institutionalized.

However, I've seen some posted traffic about test system situations like
yours, where the folks have suggested that 3rd party boot managers are a
very handy solution. If you're not familiar with such, let me say that
there are several very good and reliable ones available, even freeware,
and they share the virtue of configurability such that each bootable
system (they can be on different drives in some cases) can be isolated
from all the others by making the other bootable partitions "hidden"
when not booted. This is wonderful protection, especially in a test
environment. There's absolutely no chance of errant software corrupting
a hidden system.

The faith in flawless system or app code required for me to put multiple
OSs in one partition or even in different but mutually accessible
partitions is beyond what I can ever hope to achieve. I don't even trust
my own keystrokes much as a simple user. I've been on both sides of the
coder/user fence and been burned too often by too many people including
most of all myself. Murphy's Law is my pole star.

I used the original OS/2 Boot Manager for years. It occupied a 2-4MB
primary partition itself, and allowed up to 3 other OSs each in its own
bootable (hidden when not booted) partition on HDD0 only. The newer BMs
are much more flexible and allow for many more such independently
bootable systems, and on more HDDs.

Apologies for copping out and also, if you know all the above as well,
for stating the obvious! :-)
 
I agree with Dan -- you must use a boot manager such as PowerQuest
Boot Magic to hide any NT-readable partitions from Setup, so that only
the partion you want to install Windows into is exposed. That
partition will then become C.

Setup *really* needs the option to specify the boot drive letter
during the character mode phase of Setup! I am sick to death of
moving/removing disk drive cables and hiding partitions to make new
Windows installations boot as "C:" My current setup has XP, 2003, and
2000 Server triple-boot, and each boot comes up with its home
partition as C:.

I find that once the OS is bootable, you can move it to another
disk/partition using the procedure described in the KB for recovering
from a changed boot drive letter (I forget the KB#). However,
(another real pain) the boot loader can't span multiple IDE hdd
controllers. All the drives mentioned in boot.ini must be on the
controller specified in the BIOS as the boot controller. So, if you
have a Promise PCI card AND on-board IDE, only drives on ONE of these
can be bootable via boot.ini (or other boot managers, AFAIK).
 
Back
Top