Hello all,

GREAT FORUM - I have already used this forum for my benefit, but now i have a question.

I have a completely unattended install of windows 2003 built using PE as the kick off point.

All runs fantastic except for a small issue which is proving to be quite a pain. I have 2 raid volumes (both mirrored). These disks are incorrectly installed but as i have 180 of these domain controllers to install we won't be physically swapping the disks around. Disk 0 is the larger and needs to be DATA and Disk 1 is smaller and therefore requires the OS (12GB partition) and another partition for other uses.

Before installation they stand as follows:

Operating System - Windows 2000 Server

Disk 0 - 64GB (4GB Recovery partition Drive Letter E: & 60GB Data Drive Letter F:)

Disk 1 - 36GB (12GB Operating System Drive Letter c: & 24GB Partition Drive Letter D:)

Here's my problem.

I need to backup data off these partitions (shared directories) and repartition the drives.

I record the directory path for the required data and many other variables before starting the installation by script and this is read from an answer file for each domain controller during the PE process.

however for example a share that when running Windows 2000 server lies under e:\SMS_DP can only be found under d:\SMS_DP when PE is running.

I thought that PE would read formatted and named drives as they are and therein lies my problem.

My scripts are not backing up the data because PE in not recognising my paths correctly.

This also causes an issue when repartitioning the drives.

Any pointers you can give me in regards to this would be greatly appreciated.

Just off the top of my head and before my first cup of coffee, but could you run diskpart before the rest of your scripts, to correct your drive letters?

That's sounds correct Phurios, but I was under the impression that PE should actually be able to deal with the Drive letters as they are already known.

It seems to be naming them under it's own way.

It sees C: correctly and swaps D: and E: around.

This I don't understand because under 2000 Server C: & D: are on the same physical disk and E: is on another physical disk.

Another point is during the PE section of the install i use diskpart and select disk 0, clean it, partition and name it E:

But once windows server 2003 install is complete it sees this partition as d:

Unbelievable.

Thanks for the reply!

The way Windows is swapping drive letters around makes sense IF the second partition on DRIVE 0 is a logical partition and not a primary partition (See here for a reference.

But once windows server 2003 install is complete it sees this partition as d:

Are you saying PE sees the drive as D: after the install, or once you boot into 2003?

Just re-read your initial post - drinking coffee now - that is odd. If any error in the drive letters occured, I would expect it to be that PE sees drive E: as C:, your C: as D:, and D: as E:. I need a nap.

That's sounds correct Phurios, but I was under the impression that PE should actually be able to deal with the Drive letters as they are already known.

It seems to be naming them under it's own way.

This fact can actually be read in the official Win PE documentation.

The drive letter assignments are NOT read from the OS on the disk, they are assigned like a fresh install would assign them - with one exception: if you boot PE from your hard disk, that partition will get X:, and the rest will get the available letters from C:.