Our production databases provide you with a fault-tolerant platform for storing all your application data. In this post, I just wanted to outline exactly how these databases run and what happens should a failure occur.
All our production databases run on application servers and store all their data on our backend storage platform.
Should anything go wrong with an application server, the database frontend will immediately be restarted on another host and connect to the same persistent storage system.
When your database is restarted, we will automatically restart your whole application to ensure that all processes reconnect and are able to access the database. If this is not desirable, you can disable this through your myViaduct system.
All our storage backends run in pairs with one live and one hot spare which is ready to take over should anything nasty happen to the live host. All the data is also backed up every hour, off site, in case of any serious issues with the cluster or the data centre itself.