HD cloning

[JP Navarro]

We developed a tool called imgdisk that's part of the City Toolkit 
(www.mcs.anl.gov/systems/software). imgdisk is used to save and restore 
disk images, where a disk image includes:

1) partition information
2) master boot sector (non partition table portion)
3) individual partition boot sectors
4) contents of file-system partitions (in dd, cp, or tar format)

When a disk is "saved" the partition information is saved in a text file 
that can be manually edited, boot sectors are are saved in raw dd form, 
and partition contents in dd, cp or tar format.  A saved image is just a 
collection of files in a directory on a network file-system (typically nfs).

Restoring this image involves booting off of floppy/CD/network in a way 
that the disk you are restoring to isn't active, mounting the network 
file system with the saved image directory on it, and running image disk 
to restore the image. The restore operation will partition your disk, 
restore boot sectors, initialize file systems, and restore file system 
contents.

A couple of features that we believe are particularly useful with imgdisk:

- Linux distribution independant
- the saved and restored disks don't have to be the same size
   (issue with using a raw disk dd)
- the image is a collection of files that can be easily copied about

JP
-- 
John-Paul Navarro (630) 252-1233
Mathematics & Computer Science Division
Argonne National Laboratory navarro at mcs.anl.gov
Argonne, IL 60439 http://www.mcs.anl.gov/~navarro


Rich Grenyer wrote:

> Has anyone had any experience with cloning hard drives for cluster 
> installations? Our cluster needs fully functional disk-based nodes 
> unfortunately, and I don't feel like running through the RedHat 
> installation procedure that many times - any suggestions, or am I just 
> being lazy?
> 
> Cheers,
> 
> Rich