Hello,

Here is the problem:
I have 5 identical laptops from HP. I have configured one exactly how I would like for the others to be set up (Let's call this laptop 1). All of these laptops have factory installed XP SP2. I have taken the hard drive out of laptop 1, and attached it to a desktop IDE cable using an adapter.

Here is some history:
In the past, I have always been able to boot to a ghost 2003 diskette, do a disk to disk copy to a spare hard drive, then continue to do disk to disk copys from the spare hard drive to the remaining laptops until they are all identical. Since they are sysprepped I name them accordingly and enter all registration info.

Back to the problem again:
I can successly do the disk to disk copy with laptop 1 to hard drive, then it LOOKS as though I can successfully copy the disk back to laptop 2 (3, 4 & 5), but then when I boot these laptops back up again... nothing. I can get to the bios, I can boot to cd, I can boot to USB floppy, but I cannot get them to boot to the newly copied hard drive. I have searched and searched and cannot find anything that has worked.

First I tried using a different hard drive, then switching the IDE cable, then switching to a different adapter. I thought I'd go back to square one. I decided to use laptop 2 for testing purposes. I used the restore disk that came with the laptops and tried to do the same procedure with it, just in case there was a problem with the way I configured it. The one good thing about it is that it will boot up now. So it's not like the hard drives are damaged now, it does seem that they are in fine working order. That still didn't work, I got the same results. So I did a little more searching... I've tried changing from PC-DOS to MS-DOS on the bootable floppy that I'm using. I've also tried making sure that the newly cloned hard drive is marked as active.

Again, I've done this successfully with Toshiba laptops in the past without a hitch. I thought maybe it had something to do with the new hp laptops. They are HP/Compaq nc6120's. I contacted HP support. They had me run the diagnostics on the hard drives, each of them ran successfully with no errors. What am I missing? Please help, I'm at my witts end!!

crbell

1.  Are you doing *everything* in DOS, and never re-booting to Windows with the laptop HDD still attached to the desktop system?

2.  Can you create an *emergency WinXP boot floppy* using this technique and boot the system after the cloning?:

How to use System files to create a boot disk to guard against being unable to start Windows XP

3.  Use TeraBytes *editbini* program to check *boot.ini* file to make sure it has not been changed from the original:

TeraByte Unlimited Freeware

Do any of the above help?

Hi NightOwl,

1.  I am doing everything in Dos, and not rebooting.  When it is done I exit out of ghost to get to a c: prompt then turn the pc off, and remove the disk immediately.

2.  +  3.  I'll try them and post back.

Thank you thank you for your quick response!

Nightowl,

Wow that was an experience.  I copied the Boot.ini, NTLDR, Ntdetect.com files to a formatted diskette from the source laptop, then tried to boot with that diskette to the 2nd laptop.  It took a LONG time, but it finally booted to windows.  I then copied the files to the c: drive on the 2nd laptop, removed the diskette (I'm using a USB floppy drive btw), and rebooted.  No results, it still flashes in the upper left corner.  So that was strange.  It was the first glimmer of hope I've had in a while though I'll tell ya!

As for your 3rd response, I will download and try that also.  

crbell

The cloned system booted using the floppy boot disk, but after copying the files on the floppy to the HDD on the cloned system--it would not boot!--as you said:


Quote:

So that was strange.

I'm still suspecting the *boot.ini* may be altered by Norton Ghost during the cloning process.

In theory, if the system boots from the floppy, the only file that should need replacing is the *boot.ini* file on the cloned system.  

Are you sure you were successful in transferring the copy of *boot.ini* from the floppy to the HDD?--*boot.ini* may be set as a system, read-only file--you should *open* it with *Notepad* and check the text lines--compare the original on the source to the one on the 2nd system after booting from the floppy.

Using TeraByte's *editbini* should also give you the ability to look at the *boot.ini* file's content as well from DOS without being able to boot the system--but as long as you can boot to Windows, Notepad works just as well.

Just to be sure I cloned it again.  Using the diskette with the files from HD1, I was able to boot HD2.  Once it was booted I compared both of the boot.ini files and they looked absolutely identical.  Both marked system and hidden, both have same dates and times, file size.  Opening them with wordpad and comparing them makes it look as though they are the same.

I am running a scandisk on HD2 right now, and will do a defrag when done.  This has me perplexed!

crbell

I do not use *sysprep* and do not know the steps involved in *re-registering* systems using that technique--so that may be an area that is somehow effecting the results.

Here's another idea--you could try using Ghost 2003 with a command line switch to see if there is something being overlooked by Ghost in the Master Boot Record:

Ghost -IB

From this reference:

Switches: Alphabetical list of switches



Quote:

When creating an image or copying disk to disk, the -IB switch (Image Boot) forces a sector-by-sector copy, and copies the entire boot track; not just the boot sector.

I will give the -ib switch a try.  

Have I told you how much I appreciate your help???  :)

Well, I've tried that and it doesn't work either.  It does work when I put that magic diskette in though.  I just don't get it.  I've been searching these forums because I thought I saw something a while back that said something about how hp/compaq laptops have a separate 8mb section that tells it how to boot.  I may have dreampt that though.  ;D  

Do I have any other options?  I was thinking maybe even creating image cd's.  At this point I'm really willing to try anything.  

Here is something else odd that I just noticed.  I booted the 2 laptops up side by side.  Everything is identical EXCEPT when looking at disk management, HD1 is .01GB larger than HD2.  So I rebooted both to dos, then displayed partition info with fdisk, and hd1 is listed as 38154MB, where hd2 is listed as 38147MB.  Where is the other 7MB?  Both are partition 1, active status, and at 100% usage.  

"I rebooted both to dos, then displayed partition info with fdisk, and hd1 is listed as 38154MB, where hd2 is listed as 38147MB.  Where is the other 7MB?"

To diagnose that, you need to look at the actual partition tables, not at DiskMgmt's or fdisk's reinterpretation of the partition tables.

Visit www.partitionsupport.com and download findpart (either the dos or xp versions).  Run findpart on both systems and cut/paste the reports here for review.  Note you don't need the full findpart search, so "findpart tables" should be sufficient.

crbell

You could try another Ghost command line switch:

ghost -ia -ib


Quote:

The Image All switch forces Ghost to perform a sector-by-sector copy of all partitions. By default, when copying a partition from a disk to an image file or to another disk, Ghost examines the source partition and decides whether to copy just the files and directory structure, or to do a sector-by-sector copy. If it understands the internal format of the partition, it defaults to copying the files and directory structure. Generally this is the best option. However, if a disk has been set up with special hidden security files that are in specific positions on the partition, the only way to reproduce them accurately on the target partition is through a sector-by-sector copy. If you use this switch to create an image of a dynamic disk, then the image must be restored to a disk with identical geometry.

Maybe the laptop's HDD has some *hidden security files that are in specific positions on the partition*.

Happy Monday all!

Dan:
This is from HDD1

OS:  Windows 5.1.2600 Service Pack 2     Partition tables:

Disk: 1   Cylinders: 5168   Heads: 240   Sectors: 63   MB: 38154

-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS  CHS
   0 1*07       63 78140097 38154    0   1  1 5167*239 63 OK   OK


This is HDD2

OS:  Windows 5.1.2600 Service Pack 2     Partition tables:

Disk: 1   Cylinders: 5168   Heads: 240   Sectors: 63   MB: 38154

-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS  CHS
   0 1*07       63 78124032 38146    0   1  1 1023 254 63 OK   NB
                                     0   1  1 5166 224 63  Actual


A little different, eh?

NightOwl:  I'll try the -ia -ib switches and let you know.

Dan,
I see the difference, but what can I do to change it?

Aha.  That's what I half suspected when you said there was a 7MB discrepancy instead of 8MB.  I'd previously experienced this idiosyncracy with IBM Thinkpads, but didn't know HP/Compaq laptops also did this.

---

This is from HDD1

Disk: 1   Cylinders: 5168   Heads: 240   Sectors: 63   MB: 38154

-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS  CHS
   0 1*07  63 78140097 38154    0   1  1 5167*239 63 OK   OK


This is HDD2

Disk: 1   Cylinders: 5168   Heads: 240   Sectors: 63   MB: 38154

-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS  CHS
   0 1*07  63 78124032 38146    0   1  1 1023 254 63 OK   NB

---

I've highlighted the problem area.  The cause is the way the computer's bios autodetects the hard disk parameters.  When any computer boots, one of the first things it does is query the controller firmware on the HDD to find out the disk size and parameters.  Put that HDD in an IBM (or, it seems, a HP/Compaq) laptop and it will report back cylinders/heads/sectors configuration of 5168/240/63.  But put that same HDD in a Dell or Toshiba laptop and it will report back cyl/hd/sec = 4863/255/63.  Note these are fictitious numbers anyway (and have been since we got beyond 528MB disks and started using LBA a decade ago), but it has serious ramifications in the way the partitions and partition table are consequently structured.  When you write your partitions and file systems using one geometry, it will not work if you try to read them using a different geometry.

Note this idiosyncracy is dependent on the bios, not the HDD itself.  Any autodetected HDD will always show 240 heads in a Thinkpad, and always show 255 heads in a Dell.  Since all that really matters is the disk size, you'll notice the cylinder count is adjusted to provide the same disk sizes under either bios, given the head count, sector count (always 63), and sector size (always 512 bytes).

You ran into this problem because you removed the HDD from the laptop and put it in another machine during the cloning operation.  That other machine autodetected the laptop HDD with a geometry of 255 heads and different cylinder count, so when the partition was cloned is was automatically adjusted to fit a 255-head geometry.  When you moved the HDD back to the laptop, which again autodetected 240 heads, the partition that was created with a 255-geometry was out of sync.

So how do you get around this?  In the old days we used to have a bios option to manually tell it what geometry to use.  But on modern computers that option has gone the way of the dodo bird, so you'll just have to restrict your clone-making to a machine that will autodetect 240 heads.  That means your imaging and restoring needs to be done from the laptop.  You should explore your options to image from the laptop direct to/from cd/dvd, external usb drives, a network share, or secondary partitions on the same hard disk.

P.S.: why was a 7MB discrepancy a clue?  Because partitions are always sized in 1-cylinder increments.  One cylinder under 255-head geometry is 8MB.  One cylinder under 240-head geometry is 7MB.

Dan Goodell

Okay, so the disk geometries don't match--but why will the system boot successfully if an *emergency* WinXP floppy boot disk is used?

Does the mis-matched geometry only effect where the system looks for the files used for booting that's on the floppy?

That is absolutely amazing.  There is no way I would have figured that out on my own.  HP couldn't help me, symantec coulden't either.  See how much trial & error & experience counts?

I will see if my boss will purchase a USB hard drive and we'll give that a shot.  

One more question though...
If I do a disk to disk, from laptop to usb, then usb to laptop and laptop and laptop (and so on), could I then do it from usb to desktop hard drive (for storage purposes)?  Then later when I get more identical laptops or have to reimage the ones I have now, I would have to go desktop hard drive to usb drive, then usb drive to laptop?  Yes?  I'm asking because we almost always have a spare imaged hard disk for each model machine we have.  I don't think we'll get more than 1 or 2 USB drives.  I hope this makes sense!

crbell

Create whole drive*images* of the HDD's, not *disk-to-disk* clones--the image will restore the HDD just fine, just as if it was *disk-to-disk*, only now it's *image-to-disk*.

As long as your labtop and external USB HDD recognize each other, you should be fine.

I see what you are saying, that way I can save more than 1 image to the USB drive.  Right?

This will all have to wait until tomorrow.  Boss is gone for the day and I'll have to get it approved before I can buy anything.  I sure hope this will get this resolved!

"If I do a disk to disk, from laptop to usb, then usb to laptop and laptop and laptop (and so on), could I then do it from usb to desktop hard drive (for storage purposes)?"

As NightOwl said,  use images.  Images are files, not partitions, so they can be stored anywhere files can go.  You can burn them to cd/dvd, or store multiple images on another drive or network share.

Just make sure the source disk is seen in its proper geometry while the image is being created, and the target disk is seen in its proper geometry at the time the image is being restored, and it won't matter where the image is stored in between.

Dan

I had a question in *Reply 15* that I think you missed--why will the HDD that has been cloned *disk-to-disk* on a different system that sees the geometry different, and then put on a system (the laptop) that calculates the geometry differently, boot successfully using the WinXP emergency boot floppy disk?

"Does the mis-matched geometry only effect where the system looks for the files used for booting?"

When I said mismatched geometry wouldn't work, that was a broad generalization.  It may work for awhile, but it's a ticking time bomb just waiting to blow up on you.

What won't work is read/writes that are based on geometry.  XP actually uses LBA addressing for most read/write functions, so if it can get past the boot phase, geometry is taken out of loop and becomes immaterial--at least, for typical day-to-day operations, that is.

But it may blow up when you get to one of the geometry-based disk access operations, such as some Disk Management functions.  You can also run into trouble if you have multiple partitions, or when trying to access the disk near the end of a partition because there is ambiguity about exactly where the partition ends.  And for heaven's sake, don't ever let XP try to scan or repair "errors" while you're running with mismatched geometry.

IOW, if it boots you're probably not doing immediate damage to your file system, but in the long run you really want to fix the mismatch.

Dan,
I'm waiting for my USB hard drive to come in. :)  Hopefully that will get all of these laptops taken care of.

In the meantime, I figured I'd mess with my "Ghosting Desktop" and not have it autodetect the 2 laptop drives.  I went into the bios and changed from autodetect to 5168/240/63 for both drives.  Booted to the ghost boot disk, and when it shows the drives to be ghosted, it shows 4863/255/63 instead.  I figured I'd go ahead and try disk to disking it, with an -ib switch.  No go.  Installed both laptop hds into the desktop, changed the bios to 4863/255/63, and proceeded to disk to disk it again.  Same deal.  

Dan, when you said "In the old days we used to have a bios option to manually tell it what geometry to use.  ", where you talking about changing it from auto to user and specifying, because that is what I did.

I'm including another findpart.txt, this one is from the desktop with both laptop hard drives plugged in.

Findpart, version 4.42.
Copyright Svend Olaf Mikkelsen, 1999-2004.

OS:  DOS 7.10     Partition tables:

Disk: 1   Cylinders: 4864   Heads: 255   Sectors: 63   MB: 38154

-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS  CHS
   0 1*07       63 78124977 38146    0   1  1 1023 239 63 OK   NB
                                     0   1  1 4863  14 63  Actual

Disk: 2   Cylinders: 4864   Heads: 255   Sectors: 63   MB: 38154

-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS  CHS
   0 1*07       63 78124032 38146    0   1  1 4862*254 63 OK   OK

Am I doing something wrong?  Which way should I set the bios?  Should I just be patient and wait for the usb drive?

Thanks!

"when you said "In the old days we used to have a bios option to manually tell it what geometry to use.  ", where you talking about changing it from auto to user and specifying, because that is what I did."

Yeah, that's the option I was thinking of.

"I figured I'd mess with my "Ghosting Desktop" and not have it autodetect the 2 laptop drives.  I went into the bios and changed from autodetect to 5168/240/63 for both drives.  Booted to the ghost boot disk, and when it shows the drives to be ghosted, it shows 4863/255/63 instead."

Bummer.  That's what I would have expected to work.

Bios code is one of the least standardized pieces of any computer, so behavior can be wildly variable.  Perhaps the bios works different than we're expecting and is ignoring you.  Or what might be just as likely, perhaps Ghost is bypassing the bios and issuing an ATAPI IDEIDENT command itself directly to the controller on the HDD.  

I wouldn't be surprised to find findpart bypasses the bios because findpart is a "discovery" kind of tool, so by the nature of its job it has to consider that pieces of the puzzle may not be working right.  But it's possible Ghost might also work by bypassing the bios.  

Whether it's the bios or Ghost, though, it doesn't look like it's your fault.  For whatever reason, the hardware just isn't going to let you do what you want.

Just as an aside, crbell, this is one reason I habitually split laptop hard disks into at least two partitions.  This avoids all the hassles of imaging my IBM laptops, with their odd autodetecting.  With two (or more) partitions, I simply leave the HDD in the laptop and image the OS partition into a file on the second partition.  If I want, I can always move the image(s) to cd/dvd or across the network later, after I'm back in Windows.

:D
Update!!

The USB HD worked!!  I used NightOwl's very detailed instructions found here:

http://radified.com/cgi-bin/YaBB/YaBB.cgi?board=general;action=display;num=1095438251

I have finally gotten this to work!!  Thank you both for your continued help on this.  It was definitely a learning experience.

Hi,

I had the problem of cloning an XP NTFS harddisk from a non-name notebook and installing it in an IBM ThinkPad X30.
After re-installing Windows the ThinkPad wouldn't boot from harddisk, only from a boot diskette.
TestDisk showed NTFS to have 255 heads, but the harddrive to have 240 heads. But changing the geometry, writing new a new mbr and all the repair console tools where of no help.

My solution:
I compared the boot sectors of a freshly installed XP on my ThinkPad, that would boot, to the boot sector of the non-booting harddrive.
At hex 7E1A my booting sector hat F0 (= 240), my non-booting sector showed FF (= 255).
I booted the ThinkPad with BartPE and used TinyHex to directly change the value from FF to F0 on the installed non-booting harddrive.
That solved my problem.

Hope it helps.

Thomas