That of course depends on what scenario (for what disasters) you planned.
Testing the recovery of a user borked file is of course easy. Create one as normal user. delete/change it after a backup has been made. Wait some time and see if you can restore wthout meddling to much around. Check ownership and pemission bits after restore. In short gather experience (and write it down somwhere, but not on the to be recovered system).
Same for complete rcovery. We realy tested this on live systems (again use an old system from a second hand store or so, as said, it is a balance between your needs and your resources). Grab the disaster recovery book wich will tell you things like:
. where are the backups, how to get them asap (telephone numbers at midnight, taxi drivers for transport, and … and …)
. install on the new system, with the list of parameters.
. partition data disks, create Logical Volumes, etc.
. restore from your backups (monthly first, then the up to three weekly, then the up to 7 daily, but this again depends very much, the rsync/cp hardlinks method only requires restore of the last backup.).
. recovery of databases.
.
.
.
.
Can be as long as a 24 hourse of hard work. So organise for food and driink.
When you do not practise, you will loose customers when it comes to the real thing rotfl!rotfl!