Ghost questions

[Greg]

I use GHOST Ver. 7.5 for maintenance. I have Windows XP and have 2 problems-

1. If I backup one disk to another one (Disk to Disk) I use the flag -FDSP
and the image copying works fine. Problem is if I boot once from the new
disk (that works fine as a stand alone) but with the original disk plugged
in also, than it wont boot again if the original disk is disconnected.

2. I need to backup only the root partition ( with the boot sector ) but
this doesn't work - I must copy the whole disk to an image.

Thanks in advance. Any reference to my problems would be very appreciated.

Regards

Greg

 

[David H. Lipman]

From: "Greg" <[email protected]>

| I use GHOST Ver. 7.5 for maintenance. I have Windows XP and have 2 problems-
|
| 1. If I backup one disk to another one (Disk to Disk) I use the flag -FDSP
| and the image copying works fine. Problem is if I boot once from the new
| disk (that works fine as a stand alone) but with the original disk plugged
| in also, than it wont boot again if the original disk is disconnected.
|
| 2. I need to backup only the root partition ( with the boot sector ) but
| this doesn't work - I must copy the whole disk to an image.
|
| Thanks in advance. Any reference to my problems would be very appreciated.
|
| Regards
|
| Greg
|

And you couln't post in a Symantec News Group such as; symantec.customerservice.general

Afterall Ghost is NOT a MS product, it is from Symantec.

#1 -- Disk-to-disk is not a backup and does not create an image. It is a clone operation.
Only when creating a Ghost image to GHO/GHS files is it considered imaging and/or a backup.
Additionally, I don't know why you want tpo preserve the disk signature bytes on the
destination. I have cloned and imaged numerous hard disks and I have never need this
command line switch (-FDSP). Maybe this is causing you problems ? You also must explain if
you change jumpers master/slave jumpers or put the drive on another IDE channel (assuming
they are IDE becuase you left out that data).

#2 -- So ? Why is the disk partitioned in the first place. All you need is one large "C:"
in WinXP.

 

[Leythos]

#2 -- So ? Why is the disk partitioned in the first place. All you need is one large "C:"
in WinXP.

Bad move - there are security reasons and/or maintenance reasons to
consider. In all my years I've never just done one large partition for
the OS drive when it was possible to have one for OS and one for DATA.

 

[David H. Lipman]

From: "Leythos" <[email protected]>

|
| Bad move - there are security reasons and/or maintenance reasons to
| consider. In all my years I've never just done one large partition for
| the OS drive when it was possible to have one for OS and one for DATA.
|
| --
|
| (e-mail address removed)
| remove 999 in order to email me

That's a bad move as well ;-) If the Hard disk fails, which is inevitable with cheap
drives, then all partitions go. I use a sparate drive. One completely just for data (and
TEMP files and caches) and another for the OS. The chances of both drives going at the same
time are extremely high. Of course the data drive is backuped up reguarly to a DAT tape and
the system is backed up to DVD periodically.

BTW: All my hard disks are SCSI (wide and narrow). By using separate hard disks one can
take full advantage of multitasking disk read and writes. This is always easy with SCSI but
if all hard disks are master drives on their one IDE channel it can be performed using IDE
drives. There would be no advantage gain if there were two IDE hard disks and they are in a
master/slave relationship.

 

[Leythos]

From: "Leythos" <[email protected]>


|
| Bad move - there are security reasons and/or maintenance reasons to
| consider. In all my years I've never just done one large partition for
| the OS drive when it was possible to have one for OS and one for DATA.
|

That's a bad move as well ;-) If the Hard disk fails, which is inevitable with cheap
drives, then all partitions go. I use a sparate drive. One completely just for data (and
TEMP files and caches) and another for the OS. The chances of both drives going at the same
time are extremely high. Of course the data drive is backuped up reguarly to a DAT tape and
the system is backed up to DVD periodically.

But we were not talking about more than one drive. And while we're at
it, using a non-redundant drive to host the real data on is a bad move
too. Most users only have access to simple methods of online storage -
like purchasing a simple SATA/IDE RAID controller card and a second
drive - with minimal effort it can be installed and up and running in
under 30 minutes.

If you have two drives, not RAID, and use one for the OS and one for
data, you have the same risk of a fault rendering either drive as a
loss.

What you gain, talking about partitioning, is that you can setup secure
areas that would be more difficult on a single partition, you can also
wipe and reinstall the OS partition without impacting the data
partition.

If you want redundancy, you need RAID.

BTW: All my hard disks are SCSI (wide and narrow). By using separate hard disks one can
take full advantage of multitasking disk read and writes. This is always easy with SCSI but
if all hard disks are master drives on their one IDE channel it can be performed using IDE
drives. There would be no advantage gain if there were two IDE hard disks and they are in a
master/slave relationship.

We all know that IDE/EIDE is not as performance minded in a multi-
user/multi-task system, but there are significant advances that come
very close to rivaling SCSI controllers.

The Promise SX6000 handles onboard CACHE, also has it's own controller
to off-load CPU time, and handles 6 IDE drives in a hot-swappable mode.
I've got more than a Dozen of these running on servers with 250GB EIDE
drives attached in a single RAID-5 setup.

There are also some onboard IDE or SATA RAID controllers packaged with
Motherboards that don't include their own CACHE memory and take CPU
time, but, they are much more reliable than a soft-raid solution.

So, we're talking two different things here:

Paritioning: 1 drive with 2 areas (OS and DATA) works well since you
have more options when fixing things and also great for security

Redundancy: RAID - Two or more drives that act as one, so that if one
fails there is no data loss.

 

[David H. Lipman]

From: "Leythos" <[email protected]>


|
| But we were not talking about more than one drive. And while we're at
| it, using a non-redundant drive to host the real data on is a bad move
| too. Most users only have access to simple methods of online storage -
| like purchasing a simple SATA/IDE RAID controller card and a second
| drive - with minimal effort it can be installed and up and running in
| under 30 minutes.
|
| If you have two drives, not RAID, and use one for the OS and one for
| data, you have the same risk of a fault rendering either drive as a
| loss.
|
| What you gain, talking about partitioning, is that you can setup secure
| areas that would be more difficult on a single partition, you can also
| wipe and reinstall the OS partition without impacting the data
| partition.
|
| If you want redundancy, you need RAID.
||
| We all know that IDE/EIDE is not as performance minded in a multi-
| user/multi-task system, but there are significant advances that come
| very close to rivaling SCSI controllers.
|
| The Promise SX6000 handles onboard CACHE, also has it's own controller
| to off-load CPU time, and handles 6 IDE drives in a hot-swappable mode.
| I've got more than a Dozen of these running on servers with 250GB EIDE
| drives attached in a single RAID-5 setup.
|
| There are also some onboard IDE or SATA RAID controllers packaged with
| Motherboards that don't include their own CACHE memory and take CPU
| time, but, they are much more reliable than a soft-raid solution.
|
| So, we're talking two different things here:
|
| Paritioning: 1 drive with 2 areas (OS and DATA) works well since you
| have more options when fixing things and also great for security
|
| Redundancy: RAID - Two or more drives that act as one, so that if one
| fails there is no data loss.
|
| --
|
| (e-mail address removed)
| remove 999 in order to email me

Partitioning effectively only provides separation.

I don't want to get into technical aspects of applying storage solutions. It is outside the
scope of this discussion. However two hard disk is better than one hard disk and cacheing
controlers (unless they have a battery back up) can cause catastrophic data loss and/or data
corruption. Promise constollers in general and the company as a whole SUCKS ! You will
also find that advanced controllers use their own CPU in assymterical multiprocessing
techniques. Older ISA comtrollers used 80186 or 80286 CPUs and many PCI controllers use
80386 CPUs. Like I said, it is out of the csope of the discussion at hand.

 

[Leythos]

Partitioning effectively only provides separation.

Which is a very good thing.