Backup strategies for PCs in a SoHo environment should be fully automated and should not be relying on any user initiation or interaction – otherwise they risk to be postponed and forgotten in the daily race. For my everyday backup I additionally prefer techniques that allow me to retrieve different versions of each file, just in case I want to go back to some data that I changed in the meantime.
Disk imaging
Backup strategies based on disk images and incremental/differential snapshots appear somewhat outdated to me. They are a nice complement to a file based backup and should be used every now and then for a complete system image. But they are slow and not very flexible in handling and while the backup process is running there is usually a lot of load on the machine that is backed up, so this never really happens unnoticed in the background.
Second Copy
For many years I used file based backups created with a software called Second Copy running on a separate machine. This system is able to keep older versions of files, but it has no possibility to automatically delete the oldest files when the backup disk is filling up. As it reads the backup data through normal file system or file share access, it also cannot backup locked files. The load on the machine that is backed up is also quite heavy.
Seagate Replica
In 2009 I tested Segate’s Replica
BUT: Replica refuses to backup harddisks larger than the internal Replica backup disk. And I am not talking of used space on your HD but of total disk space of the first HD in the PC you want to back up (even if it spreads over multiple partitions)! As the biggest Replica comes with a built-in backup disk of only 500 Gb, this is by far not enough for most modern PCs. This is a very severe drawback that makes Replica useless for most modern PCs.
Rebit – the magic behind Replica
The software used by Replica has not been written by Seagate. It is available separately (without the harddisk hardware) from a company called Rebit. Combined with a large external HD this is a good solution that circumvents the Replica limitations.
Still there is one heavy usage problem: The backup schedule of files cannot be limited. Every file that changes will go into the backup. Large inbox files of your email program are the best candidates to ruin the great concept: They change with each and every email coming in and lead to constant backup activity. In my case some of the files are of gigabyte size. They lead to constant backup activity, which slows down your system and swamps your backup disk.
Bottom line
Right now I use the Rebit solution and I limit the backup frequency by only attaching the backup disk every second or third day. When a backup run is complete I detach it again. Unfortunately this screws the basic idea of a backup solution with no intervention or interaction. It’s a pitty that this great concept is voided by such a small detail. I am desperately waiting for Rebit to add a configuration option for the backup frequency of single files/folders. Otherwise the Rebit software is a really nice piece of software.
So basically I am still looking for the perfect solution. It is quite hard to believe that after so many years of desktop PC backup evolution, there is still no better way to get this done.