Best Practices for Database Backups

"Let our advance worrying become advance thinking and planning"
—Winston Churchill

The database backup tasks that your CBORD systems complete on a nightly basis are arguably the most important task that your system performs. No matter which system you use, or how you use it, your database is your single greatest asset. The guidelines below are presented to make sure that your data remains safe and secure when the unexpected happens.

Do not rely exclusively on local backups

This is perhaps the most common mistake that our users make. CBORD applications all have the capability of performing a local backup. While this task is important and should be monitored, it's not sufficient. Too often CBORD's customers experience drive failure and are left with a non-recoverable database.

The solution is simple. A task can be configured to move the backup from the local disk to a network location. CBORD recommends that this process happens at least daily.

VM snapshots should not be used to replace a database backup process

Many of our customers use virtual infrastructure for a variety of very good reasons. While CBORD encourages all of its customers to maintain regular snapshots, these should not be seen as sufficient system backups. Occasionally problems with the database are not identified until a snapshot is restored. While CBORD can often work to correct the issues, the work significantly increases the outage period.

Practice, practice, practice

When was the last time you or your site performed a test restore of a backed-up database? You might be doing regular, off-server backups, but when was the last time you verified to make sure that the database integrity was sufficient to perform a rapid restore in the event of an outage? Unfortunately, there is no perfect backup process. While the tasks may all show success, the proof, as they say, is in the pudding.

You should perform a validation or test refresh at least monthly. Your site may have more specific guidelines that you should follow in order to maintain internal compliance. The restore process has several benefits, especially if you license a test system. The benefits include:

  • Regular restore procedures into test will keep the test system up to date with production and assist in the rollout of a test/production deployment cycle.
  • Regular restoration of the CBORD database helps ensure that onsite personnel are able, capable, and comfortable restoring your database should there ever be an emergency.
  • Regular restore procedures facilitate the creation of site-specific documentation. Remember that the backup process is all about protecting yourself. The difference between a speedy disaster recovery and an extended outage has a lot to do with your team's familiarity with the process.

A few quick questions

The answers to the questions below should guide you and your team toward defining and establishing a comprehensive backup and restore plan.

  1. Who is responsible to monitor your local and off-server backups? How do they know that they're successful?
  2. How much data do you lose if you lose an hour of data? How about a day, a week, etc.?
  3. What's the service impact on your staff and your user community if your system goes down? Given that impact, what should be your restore timeline?
  4. How much data can you afford to lose?