Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I tried looking through old questions, but couldn't find one that hit this exact question.

In VmWare's documentation on Snapshots, they mention utilizing them for a "process tree" setup. This would allow spawning multiple VMs from a base .vmdk disk. The idea is that you create a base disk and then snapshot it. This creates a "differencing" disk that is used for any changes made to the VM. I've tested this by using the following steps:

  1. Create a base VM with a new .vmdk file, and configure it as needed
  2. Shut down the VM
  3. Create a new VM and choose to use the existing .vmdk disk.
  4. Before Powering on the new VM, take a snapshot of the new VM.
    • This will create a differencing disk in the folder with the new VM
    • The new VM will write any deviations from the base, to this new .vmdk file.
  5. Power on the VM
  6. Repeat Steps 4-5 for additional VMs

My thought is that this could save on the amount of disk space needed to run multiple VMs on shared storage (iSCSI SAN).

The Question:

I wanted to see if anyone has run this setup and what type of performance implications can be expected (additional disk I/O, memory, etc) or any other snafus to be aware of? Are there any situations where this would cause more issues than its worth?

share|improve this question
if your backend doesn't dedupe then this is the next best thing. – tony roth Apr 22 '12 at 3:58

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.