Clone server pattern

A common problem for any system administrator is keeping data persistent across servers. The approach for keeping data persistent across systems depends on the type of application, the filesystem used, and many other key factors. For example, a database would be clustered across multiple instances so that the data could be shared (or stored in separate blocks) in an optimal way, or so that indexes could be used properly for data in different locations.

The type of application we will cover in this example should be simple, so we will investigate how to scale horizontally using a simple HTTP server that allows for file uploads. In the previously touched scale out pattern, we covered how to create a server that hosts HTML content via Apache, and how to increase the throughput by cloning the server behind a load balancer. That example assumed no dynamic data because if anything were uploaded to the server, it would be lost when scaled down or not reachable, depending on...