Werner Still

all or nothing - total restore

Blog Post created by Werner Still on Oct 13, 2015

The first professional contact with data protection (at that time we just called it backup) was starting of the 80s with a 1600bpi reel-tape. The immense amount of a single 300MB Hard Drive needed backup. And we had only one of these drives. So a backup was to protect against a loss of this Hard Drive.

 

Total restore was the primary usage for restores. On our systems (BSD 4.2, good old times) we used a system dump utility to allow faster total restores, compared to using tar for single file restores.

 

Today with the advance of storage technology nobody needs to worry about a total loss of data due to hardware failures (at least if they use the available technologies and operate them in a good fashion). This lowers the need for a full restore a lot.


The total restore is still in our minds as a very important restore option. Talking to customers there feedback was completely different. For data protection reasons a total restore is used in less than 10% of the cases. And this is going down even more, the bigger the databases grow. Total restore is just to time consuming and the necessary recovery sometimes takes even more time.

 

But there are still issues that will need a full restore. A major point here is to have a fail back position if structural changes are necessary to a Database. ITIL refers to revert to baseline in their operation handbooks.

 

Another task for total restores was to feed other systems with a copy of the production database. This is a very commonly used method.

 

If you have any other tasks for total restores, please let me know!

Outcomes