Some Cloud Servers use an architecture that pairs a system disk with a separate data disks. The benefits of implementing this architecture include improved provisioning times and more flexibility for local data storage and disaster recovery. Server images don’t include data disks, but they do include local system disks. This means that you need to use a separate method for saving information located on the data disks.

You can retain the operating system and configuration information from your system disk by using our Cloud Servers imaging feature. However, this process does not save any information from the data disks. To save your data disks, you can use Rackspace Cloud Backup, which is a granular file-level backup system that you can configure to save the files and folders you want to keep.

Because system images save data from only the system disk, you need to use a backup solution like the Rackspace Cloud Backup service to retain information from the data disks.

You can scale your servers by hosting your application on a load-balanced cluster (horizontal scaling) or recreating the server with a different server flavor.

Horizontal scaling is the most flexible approach, but your application must work in a clustered environment. With horizontal scaling in place, you can add and remove servers on-demand, either manually or with our Auto Scale features.

To recreate the server, you can take an image of your server and build a new server with a different size. Alternately, you can migrate data to a new server by using a tool like rsync. If you use a Cloud Block Storage volume as your server’s boot device, you can also clone your boot volume and use the clone to create a new server (see Boot a server from a Cloud Block Storage volume). Note that none of these approaches preserves the server IP address. If you require a persistent IP address, consider a load balancer or other proxying solution that maintains the IP address separately from your application servers.