Here is the situation:

-2 hard drives on the same computer (master and slave). Both IDE drives
-2 primary partitions : Windows Nt Server 4 installed on C:
-disk mirror created between the two drives in Disk Administrator. So, the second drive is an exact replica of the first one.

This is my testserver, so I'm replicating a situation that could happen with my PDC: drive failure.

I unplugged the master drive from the computer to promote (by changing the jumpers) the second drive to master. I thought it was going to boot normally because its an exact replica...but guest what!
I found those little interesting documents at Microsoft site : Q113977 and Q119467

I think my Ntldr and ntdetect.com files are fine. I think my problem comes from Boot.ini

it looks like this...

[boot loader]
timeout=10
default=multi(0)disk(0)rdisk(0)partition(2)\WINNT
[operating systems]
multi(0)disk(0)rdisk(0)partition(2)\WINNT="Windows NT Server Version 4.00"
multi(0)disk(0)rdisk(0)partition(2)\WINNT="Windows NT Server Version 4.00 [VGA mode]" /basevideo /sos

....I think its setup right because I've got 2 partitions, "multi" is there because I've got a IDE drive and my %systemroot” is \WINNT. But then I get this message :

"Windows NT could not start because of a computer disk hardware configuration problem. Could not read from selected boot disk. Check boot path and disk hardware. Please..."

In the second document, at the end, they talk about this Pitfall but I think the meanings doesn't apply to me because its not a SCSI drive.

Any help is appreciated

One configuration change that could significantly ease this problem is to put each hard drive on a seperate controller. If the HDD on the primary controller fails, I believe moving the ribbon cable from the secondary controller to the primary controller will allow you to boot from the drive. This is definitely much easier than dinking around with jumpers in a heated situation. And your performance should increase as well.

If you have an IDE CDROM drive, just place that as a slave on the secondary controller. Some CDROM drives don't like this but that's secondary to being able to recover from a HDD failure.

HTH,

Matt

Almost. Partition(2) is there because the system thinks that the first IDE drive has two partitions, and the system boots NT from the second partition on the drive. If that's not how you're configured, try changing it to one.

HTH,
Chase

Thanks for the correction Chase. It's been a few years since I was forced to resort to WinNT RAID. :)

First : thanks to both of you guys

Let me get this right : When you talk about a controller, is it the 2 IDE connector inside the computer? If so, both disk where on the same controller using the same cord, the first one was master and the second, slave. I just unplugged the master from the cord and change the jumper on the slave to make it master.

Here is a quote from Mastering Windows NT Server 4...
"Another reason to buy SCSI-based storage systems is fault tolerant: disk mirroring and RAID pretty much require SCSI disk subsystems in order to work. You can create a mirrored set of non-SCSI drives, but the drives may not have sector-remapping capatibility (which is an important part of any fault-tolerance scheme), so I would not recommend it."

...could it have anything to do with my problem because the hard drive that came with my compu is brand new but the second one that I use for the mirror (which is the one that I try to boot from) is pretty old (maybe 5 years). That second drive might not have sector-remapping capatibility.

Another thing ; the computer where I'm trying this has IDE drives but my server (PDC) has obviously SCSI drives. Tell me if I'm wrong, but would this be the only differences between the two types of drives?...

I would have to change multi to scsi in the Boot.ini file

About the disk controller again.
Does 1 SCSI card = 1 disk controller or
1 SCSI could = 1 disk controller per connector on the card (either internal or external)

thanks

You lost me. I'm not sure what you're new question even is.

In short, however, NT looks at both IDE and SCSI drives as SCSI devices within the OS. It's just the way it talks to them. But that's the software side of it.

IDE stands for Integrated Drive Electronics, and as you might guess or may know, the actual drive controller is on the drive itself. This makes it easy to move the drives from system to system with relatively little work. The disk (almost) never has to be low-level formatted. SCSI (Small Computer Systems Interface) drives are controller dependent, and the controller is on the SCSI card, not the drive. Consequently, the drive must be low-level formatted when you move it form one controller card to another. Personally, I would format even when moving from one system to another, as minor differences is bus speed or voltages on the bus can cause tracking issues.

Disk mirroring is the process of creating an exact duplicate of an existing drive. In order for mirroring to be successful, the two drives you have should be the same. In every aspect. Regardless of whether you've got SCSI or IDE drives, they should be the same make and model, and you should even check the chip release codes on the drive itself, and in the case of SCSI drives and multiple cards, the SCSI BIOS chip.

I'm not sure if this helps you or not. I think it might clarify a couple things for you. Hope so.

if you've more questions, start a new thread, but ask them. This one's moving kinda far down the list.

Cheers!
Chase