You have got a backup plan, all the data are backed up locally as well as remotely according to their priority The backup is automated. You are sure that the the backups run at the pre-determined time slots.
However, there is one more important step that you should be doing.
That is to do regular backup-restore drills. Overseeing the need for testing the backups can have disastrous results when recovery is essential
Restore drills can expose the glitches in your backups
Could you locate the right backup in short time? If this is a remote backup, could you get to the remote backup service provider within the time limits you expected?
There could be a number of issues that you could face when you or on to the restore process.
When you do the restore drill, document each issue that you faced and the solutions for each. Keep the document safe at a well known location. You will be glad that you did it when an actual restore scenario occurs.
You will identify pieces of data that you missed in your original backup plan
It is common to miss certain pieces of data when you originally plan your backup (See categorizing the data for better backups). Restore drills can help you identify the missing pieces and make your backups more complete
The level of support from the remote backup service provider
As you face problems in restoring the data from the remote servers of your remote backup service provider, keep calling them for support. This will give you a good understanding of the level and the capability of the service the service provider is capable of providing. If it is not satisfactory, you can consider switching to another service provider.
The time to restore
The time to restore is more important than the time to backup. The time to restore equals the time of disruption for the business. So is costly. You could measure the time to restore and find the bottle necks during restore drills. This will help you tune up the process and in turn, reduce the time to backup. Of course, do document the bottle necks that you face and fixes for them as you progress.
Incremental backups can cause long time to restore If this is the case, then you have to increase the frequency of full backups
How frequently should I test?
You should run backup restore drills whenever you make changes to the backup system/process.
In addition, the restore drills should be conducted in regular intervals. Depending on the criticality of the data the interval could vary (once in a month, twice in a month ) but do not make the interval too long.
Documentation is essential
One important outcome of the restore drill will essentially be the documentation. (Lets call it “The backup restore instruction manual”). Note down every single issue that you faced and the fixes for them. It should be clear enough for another person to take the document and restore herself without your involvement.
1. The first page should list the steps for restore clearly
2. There should be a section that covers the issues that can occur and the fixes for them
You should note down even the most minor issues and their fixes - like “if the customer support number of the remote backup service provider is not working , then call this alternative direct number.”
3. Take a print out of the document and keep it in a safe place and let the key persons know the location of the document. Having the printout is important. This is to restore a damaged machine after all : )
Restore test for Live PCs
One most common question is how to conduct restore drills if the machines are in use. You should either keep a spare machine to test the restore drills or grab a spare HDD to run the tests. Either way, it is important that you run the restore drills regularly, for every piece of data you backup.
