This posting is a rant of frustration with little value to you little boys and girls. If you want to maintain your sanity, please leave now...

OK guys the problem: My boot disk, containing my WinXP system is only 100 Gb long, and, being the pack rat I am, keeping all the software I have installed, just in case, I ran out of space in the OS drive.

I decided to buy a brand new 1Tb SATA Hd to move my stuff, and I wanted to make this new ample HD my system one. So I just installed it as a 4th drive (I have very may other drives for other OS and storage) and cloned the original HD into the new one. (simple right?)
When I changed the boot configuration and tried to boot from the newly copied OS in the new drive, windows will stall: It would carry all the way until the winlogon module then hold. . . forever!

I tried several alternatives, Acronis disk clone, easymigrate, Symantec ghost,
Casper, windows backup and restore, you name it, I tried it and always found the same problem: the child new cloned OS will hang.

By sheer chance I tried to clone the OS hard drive and boot to a different disk drive letter, while the original 100Gb boot hard drive was still present in the system into its original letter: Lets say that the old OS HD letter remained C and the newly cloned one was now located at drive D:

Now the new OS booted OK, but it will refuse to boot if I disconnected the old 100 Gb HD of the original OS(??????)
What I learned after much probing was that the windows page file remained in the original C: drive.(The 100 Gb full Hd).
No simple way to establish a new paging file in the new HD: No page file, no boot with new OS. Trying to establish a page file in the new D: drive resulted in no boot form the D: drive base cloned OS .
Plus a lot of drivers and services including the sound system refuse to work in the brand new cloned OS running from drive D:.

Fuck it. I've done OS migrations with Linux and they go smoothly.

Why is Windows so cumbersome??

Oh Grasshopper.........

Intead of asking why is windows so cumbersome ask yourself this:

Why is it not so cumbersome ?

OBC

did you install the SATA driver into your motherboard?

I can't say much for sure, but I've seen a lot of important-looking registry entries referencing drives by their physical addresses, rather than relative paths or even absolute letter-mapped paths. If this is a cause of the problem then you really have your work cut out for you. Also, judging from my recent foray into multi-boot OS-migration I get the impression that Windows insists on keeping the boot record on the first physical drive in the system. I'm not sure which piece of hardware or software is responsible for this constraint but it's a little annoying .

Althought my knowledge is limited on the subject, I am still going to put my two-cents about it. Firstable, did you partition the hardrive, being a 1TB you may have to, at least I have to partition mine when I got a 250GB. It has been a while so I don't remember if it has something to do with windows not being able to handle it or the motherboard.

Also, as someone pointed out, windows expect the bootable drive to be the primary/master not the slave, and the main drive is always C:\. If I am wrong, feel free to correct me. And please let us know if you get it working and what you have to do, I would love eventually of course, to replace my hardrive for a TB.

Thank you for your comprehension and sympathy boys.
My first physical drive, with the MBR contains the Linux system. Works like a charm, no problem.
The old 100 Gb drive is a SATA, installed at location 0,2 directly connected to the MB (It supports 4 IDE and 6 SATA devices). It does not matter if I disconnect the old SATA and connect the new 1Tb SATA in exactly the same physical place 0,2 or in location 0,3: Still Windows wants the old page file located in the 100 Gb located at its original 0,2 physical address. If I connect the 100 Gb at location 0,3 Windows says "NO, Not gonna boot". One Tbyte size hard drive is fine with this computer and version of windows, I installed a brand new Windows XP OS in that drive and it boots just fine. It is the transfer that sucks.
Notice that I have done CLONING of the hard drive, which is a sector by sector copy, but regardless, the page file stubbornly remains at the old drive.
I've resigned myself to reinstall all my software in the new drive (Ohh! I need to find all those licenses and original CDs:devil.

The essence of my commentary is: Windows OS assumes it is the only one in the computer, located in the main hard drive. Setting other OS or migrating/updating require significant set-up and tweaking. I don't know if you have tried Vista in multi-boot, but it blatantly kills the Linux loader and plainly refuses to coexist (multiboot) with Linux.

That's about the size of it, and it's just part of a larger confused mentality held by the designers. I figured that Microsoft would have learned their lesson by now, but it seems that each new release of Windows involves further OS-application entanglement with programs requiring some kind of system registration for even simple functionality. This, of course, makes for a nightmare whenever something goes wrong or when maintenance time comes around, but there's not much we can do about it.

Well, there are those of us trying to be part of the solution by writing applications that keep to their own folders and don't use the registry as a general-purpose database but we're fighting a losing battle .

I recently bought a cheap laptop delivered with Vista basic (yeah sucks, I can't even use the MSDNAA licence I have access to, which is a Vista Business), I successfully installed Linux and it managed to get the GRUB working without problems. Afaik some computers are kindof 'tattooed', installing Linux on them is chasing issues. But afair we always had to install the Microsoft OS before any other one for the reason you raised.

Good luck.

This doesn't help you now, but I found this out the hard way a long time ago: buy one small drive and one large drive, instead of one single very large drive. Use a 80g drive for your system & installed software (ie: C: ) , and nothing else. Use the big 500g/1tb/etc disk for storage, and nothing else. It saves so much hassle...

just an off topic query
but whats the issue with pagefile.sys
iirc you dont need it for restarting or rebooting as you can simply tell windoze not to make any page file
if you have a large ram

or simply use some registry key to wipe it off while shutting down

mmc.exe %systemroot%\secpol.msc scroll down and find clear virtual memory
when shutting down and enable it if you like gui

so nothing important enough is kept there
is my opinion
i could be way off the mark
but any experts opinion on need of pagefile.sys presence for rebooting would be
great

Actually Blabberer, i tried to get rid of the page file all together. You can change the settings on the advanced, advanced, advanced tab of "My Computer" properties context menu. For whatever reason Windows refused to boot if I erased/deactivated the pagefile. I do have plenty of RAM.
I know it does not make more sense, the contents of the pagefile is "volatile" as RAM and should not need to be kept from boot to boot, but hey!

boot somehow, how windows dictates.

clear all swap settings, then select another hd drive (and ensure it stays the same later) as fixed, permanent swap file of suggested size + something. reboot as requested.
Swap hd and place it back as first.

windows will _always create a swap file, you want it or not. When you tell 'no swap file' windows will just make one right after booting.
And it will always recreate one in C if you go out of memory in swap file... (windows getting to critical cases is funnnnnny... )

I had a similar pagefile problem with Nero BackItUp.

Restoring a backup on an empty drive would generate the missing pagefile error but not if it overwrites an already-installed Windows, the page file is only found as corrupted and then repaired on the first boot.

TiGa

This is a familiar problem. Windows, since Windows 2000, is very particular about where it is located in the hard disk's partition scheme and each succesive version seems to get fussier in this regard.

What you are describing corresponds to an incorrect etc/fstab entry on a linux/unix system but whereas this information is transparently stored in a plain text file that is easily corrected in unix based systems, it is inaccesible,as far as I know, to anything other than the disk manager and system install software on Windows systems. During the boot phase of the system startup, the disks partition structure is analyzed and a basic default drive letter assignment is made. Then when Windows is loaded, the disk managers logical drive letter scheme is applied over the basic default scheme. If you migrate to another drive or change the partitioning scheme so that the initial default scheme changes and the disk manager assigns the wrong volume to the system disk, once the logon process starts, the system trys to find information that is not there and then will try to log on and then log off again and again for all eternity. And since you can't logon, you can't get to the Disk Manager to fix this. The only solution I know for this at this point is to do a repair install which will correct the information given to the disk manager and give you more or less the system you had before with some tweaking required and, of course, you have to reapply all the service packs and security updates.You can also figure out what the default drive letter scheme will be on the new disk or new partition scheme and make sure that the logical drive assignment will be correct for the system volume before you do the initial boot on it. If it is not going to be correct, sometimes you can fudge it by creating temporary dummy partitions--not always possible if the volume is a primary partition--or hiding partitions.Once the inital boot happens, you're good and you can remove the dummy partitions with the Disk Manager program and you will be fine from then on--unless you mess up the MBR somehow. Then you will have to recreate the dummy partitions again before you boot the system. And yes, you have to watch the page files too. When migrating systems, I would usally create an extra page file on another volume just to be sure.

Windows sucks. My knowledge of this is somewhat dated. I've passed on Vista and use a Mac these days. Unix in a prettier box with a slicker GUI essentially.

Not exactly true.
Recently i changed my partition scheme, since I washed linux partitions, and recreated them.
To make a long story short, being the work windows stored on D: and not C:, it did not boot anymore, causing me a slight headache... reinstalling properly my work environment takes long days.
Windows remapped the partitions in an unknown way, while still keeping the classic 'letter assignment' . Since I did not want to erase and recreate my partitions all the time hoping for Windows to remap them correctly, I empirically altered the boot menu file

Still have to understand why partition 2 slipped to partition 4 in Windows, but meh... Older Windows were no better in this, as once I had to snoop into MBR with a DOS floppy+MASM5 when '98 fried my HD twice (the 2nd made me write a MBR editor out of frustration...).
Mah....
So, I simply had to test all boot configurations until one catched :P

Funny boot system. Oh well, it's M$.

"default=multi(0)disk(0)rdisk(0)partition(4)\WINNT"

This is not what I'm talking about. This is from the boot.ini file and occurs at a much lower level of the boot process--right after the bios initialization. The windows boot loader(ntdetect.com and ntldr) uses this to know which partition holds the system and if it is incorrect you get a black screen with an error message(This is equivilant to a Unix "kernal panic" error and occurs at the same level of system initialization). If it is correct you go to the startup screen(s) with the little dashed lines. This is when the initial disk analysis is made and Windows is loaded into memory. When you switch from VGA mode to the Windows display mode, Windows is in memory and starting to initialize. The initialzation files don't seem to be drive letter referenced but the logon files are and this is where the problem occurs. In Windows NT4 and earlier, this wasn't so and you would logon with a different drive letter and lots of errors if you put the system on a different partition which could all be fixed with the disk manager but since Windows 2000, you get the eternal logon/logoff. There are reasons why I don't do Windows these days.

!
Yep, indeed what I were talking about happens at boot, the OS won't boot up due to this.

Aaah I think I've understood what you are referencing. It is the problem that happens when your HD letters get 'swapped' like Drive 'E' becoming driver 'F' and such. Yeah, very nasty.

Yeah, it is really nasty and in Windows this essesential information is very secret and esoteric when it should be open and transparent like it is in most operating systems. In Unix/Linux it is written like this:

/dev/hda1 / ext2 auto, 1 1
/dev/hda2 /usr ext2 auto 1 1

So if you move your linux partitions and hda1 becomes hda3, you just go open the fstab file with any text editor and change hda1 to hda3 and you're done with it. Windows is a poorly designed system in many ways. It reflects the competitive, bureaucratic and secretive mentality of the corporation that produces it. And each succesive version just gets more bloated without correcting any of the fundimental architectural flaws of the system. It's so nice to use a system without a registry, with file permissions that are simple, logical and easy to set and work as they should, with security that depends on a basicly sound and time proven architecture instead of a boatload of patches and 3rd party applications that prop up a fundimentially flawed and poorly implimented system.

mmh... i do agree but for one thing (I will vomit later for defending M$, or better I get totally drunk in the process and luckily forget )

The Windows Security is in truth a quite powerful and theoretically well designed system (apart 2 big flaws) -it can really rival unix/linux one, if not even be better on some place.
But it is M$: bugs everywhere -MSDN included- and implemented such a way that a single degree is not enough to understand it -proof the fact that even Microsoft OS guys does not understand it in all its depth... (see Vista).

But yes, Unix Security is well organized, and terribly easy to manage (File Permissions) if compared to M$ one (...including their weird 'default' management of common user permissions over the time, the absurd localized 'standard' directory names, the need to write 100 very complex C lines to get a default path just if you wish to write 'proper' code etc. etc.).

Yes, thats definitely part of the problem. It's way too complicated--like a huge overgrown jungle with lots of places for vermin to hide. Too many details to look after. There's a fundimental security in keeping things simple so when something is wrong, it is easy to spot and take care of.

After rereading the original post, I can suggest a specific remedy: You have to make your initial boot of the system on the cloned drive see the drive as drive C:. That means you have to hide all the primary partitions on the other drives. The initial boot code won't see the hidden partitions but the partition manager and utilities such as Partition Magic will and once you've done the initial logon, unhide all the hidden primary partitions in the newly cloned system and set the drive letters as you wish and you should be fine from then on. You might be able to mount the drives and set drive letters with the partition manager even with them hidden which would keep you safe forever--I seem to remember doing this at one time.

If the new drive was IDE, you could just swap drives.