Disk image files vs. physical disk clones

In every cloning project, there is a choice between acquiring the image file or making a clone to a physical drive. I will list the advantages and disadvantages of each, in no particular order. You can review these and decide which way to go for each case you have.

Disk image file

  • Raw image (bitwise copy) is compatible with any data recovery software.
  • Raw image files lack metadata. For example, physical sector size cannot be stored in a raw image.
  • VHD or VHDX images do not store zero data and occupy less storage.
  • VHD and VHDX images are allocated in blocks, so they can't have arbitrary sizes. Their size must be an integral multiple of the block size (2 MB in most cases). Therefore, images may be slightly larger than their source drives.
  • Disk image files can be further compressed (either zip-like or by native filesystem compression) or deduplicated for archival purposes.
  • Disk images are versatile in long-term storage. You can mix and match physical drives or use RAIDs to put, for example, ten compressed disk image files onto three physical drives arranged in a RAID5.

Clone on a physical disk

  • If made on the same model/revision hard drive, a physical clone is as close to the original as you can ever get. The only thing different (save for bad blocks) will be a clone drive serial number. You can try to put it back into whatever original environment, and it will often work.
  • A physical clone requires a drive to be the same size as the original. One may think using larger drives to hold clones is safe, but that's not exactly true.
    1. If you use a larger drive, the data past the end of the original must be zeroed to avoid contaminating the clone with whatever residual data is on the target drive. This is usually achieved by zero-filing the target drive before cloning.
    2. Now, using a larger drive breaks all address calculations and pointers computed using the end of the drive as the reference point. This does not affect common simple partitioning schemes but often affects RAID metadata.
  • Physical clones occupy more physical space per clone if you happen to be concerned.
  • If you need redundancy for physical clones, the cost increases by a factor of two. You can put image files on RAIDs at a lower cost.
  • Physical drives must be carefully selected to match sector size with the source, or else undesired side effects will occur in subsequent processing (see sector size with USB adapters).

Created Monday, April 24, 2017

Updated 22 May 2018