Since Monday, I've been testing our disaster recovery procedures. I wrote the procedures for this company while I was on contract for them back in 2009, but unfortunately I never got to test them.
The procedure is based on a disk configuration extract that is converted to an SSSU script that can be played at a blank EVA storage array to configure the physical storage. After the storage is configured and confirmed correct by another automatically generated command procedure, a standard restore is run to get the data onto the machines.
Due to the fact that this wasn't a tested procedure, we had the boxes up and running after a number of hours debugging hardware configuration issues and issues with the automated scripts. All in all, we took about 20 hours, which is well over our current target of 8 hours.
However, I'm pretty confident that we can easily achieve restoration in far less than 8 hours if everything goes right. And hopefully, in the near future, we will be switching to a near hot site solution which will remove the requirement for a restore from tape. In the event of a disaster, it should be boot the DR machines and go.
Posted at June 3, 2010 6:04 PMI'd say for a first run, that's a success. SSSU can be a serious bottleneck as far as the performance of the entire process.
Posted by: Kel at June 6, 2010 12:45 PM
Comments are closed