We have just completed the last step in our annual disaster recovery test and we are happy to report that it was very successful.
The test simulates a situation where we have lost all of our 6 servers, lost the system code and lost our database due to a catastrophic event.
We managed to recreate a working environment from scratch and tested that we had no loss of data in less than 6 hours.
We have proved that our backups are properly created, that they are readable and can be restored onto a fresh set of servers to resume operations.
Full disclosure: we are not resilient to the loss of Amazon’s server infrastructure in US West (Oregon). At this stage of our growth, we cannot yet afford a multi-site infrastructure.