Testing your disaster recovery preparedness.

Being a good administrator for a small to medium size company means being able to restore data when requested. Your first job as an administrator is to make sure that all important data is being backed up and verified per requirement. This should be well documented and approved by your superiors. Making sure someone besides yourself approves the backup plane removes all liabilities from you. Okay, so the data is being successfully backed up and verified per requirement. What about restoring the data? Here are a couple different scenarios you will likely face:

  • A user accidentally moves or deletes a folder
  • This is probably the most common recovery a Windows administrator will face. It can easily be resolved by using the Shadow copy feature of Windows, if the user doesn’t report this before your retention has expired, you may be digging through backups, and worst case archives. If you can’t find them in the archives, hopefully your backup policy will bail you out.

    As a Windows administrator, I have restored data through Shadow copies at least 10 times, or at the frequency of once every couple months.

  • A user deletes a folder in their Inbox
  • A little trickier, but post 2003 Exchange Severs provide an easy way to restore through Outlook. If you’re running a version of Linux server, e.g., Sendmail with IMAP, you will be digging through backups manually, possibly using the tar or unzip Linux command. You should have the acumen of whatever e-mail storage your using to restore a single mailbox or possible a single folder in a mailbox

  • Here are some inexpensive methods to testing your disaster preparedness:
  • Setting up a Hyper-V host and a guest for each critical server provides an easy to test restoring your system images. The system images being restored in this scenario will be coming from a Windows backup, i.e., wbadmin.

    Using Hyper-V, make sure that your Virtual Network Switch as an entry that is private, this way each machine can be isolated from your production network

    To conclude this short, possibly incomplete article, not only is it critical you understand how to restore each of your services at the volume level to the file level, you should also have the acumen to rebuild your RAID. It’s likely you will have a RAID-1 for the boot/system partition and a RAID-5 or RAID-10 for the data portion. Download the manual you familiarize yourself with the CLI, BIOS, and GUI

    When your system goes down or your boss accidentally removes a folder, you will receive adulation after recovering quickly from these mishaps.

    Leave a comment