It is called a Disaster Recovery Plan for a reason. If you’ve ever experienced a system crash or heard the horror story of someone who has been through one without a Disaster Recovery Plan, you probably understand how aptly it is named. Such a plan can be as simple as backup and restore procedures or as complex as off-site data storage and alternate hardware. With an adequate plan, a system crash becomes no more than a serious inconvenience.
A good Disaster Recovery Plan is as much about prevention as it is about recovery. Most all businesses running on a computer use backup power supplies and make backup copies of their data. However, batteries and backup tapes alone don’t make a Disaster Recovery Plan. Ask someone whose system crashed because their UPS (Uninterruptible Power Supply) had a dead battery and then couldn’t read their backup tapes because they had been recycled beyond their useful life. The backup hardware can be useless for recovery without procedures for proper equipment operation, steps for creating and verifying backup media, and the process to restore the backup to an operational state.
Make someone responsible for the daily backups, even if they are automatic. Have them log the media ID and results of each backup. And since they might find it easy to justify skipping an occasional backup, have someone else check the daily backup log to assure compliance. The only way to tell if the backup tapes and batteries are protecting your system is to test them. Read back or verify the backup tapes. Stop all of your applications and pull the power plug. The UPS battery should keep the system running. Make sure it can run for at least enough time to take the system down. Most UPS devices have an audible beep when on battery and the beeps become faster as the battery runs down.
Watch for other pitfalls such as the Restore software. If it runs as an application program, you won't be able to restore any data unless you can boot the system and you won't be able to restore system files that are in use when the system is running. Make sure you have enough media to backup more than just the last few days. Some files can become damaged without causing a system crash. You may not discover the corruption until you've backed up the bad file on all of your media and no longer have a good copy.
Here are some things to consider in making your disaster recovery plan:
- Have procedures for periodically testing your backup power supplies and data backup media.
- Designate a person who will be responsible for making daily backups. Have another person to check the daily backup logs and perform the periodic recovery test.
- Don't rely on restore software that runs as an application program. Have a bootable recovery program on removable media.
- Replace components within their normal life cycle even if they test OK. Both batteries and tapes have an expected life span for normal use.
- Simulate a disaster recovery at least once a year. Restore to a surplus or backup server. This is the perffered time to learn about a problem with your Disaster Recovery Plan.
- Alternate your daily media backups to include a month-end, quarter-end, and year-end cycle. When making backups of accounting or other transaction based systems that are not on a managed database, bring the application down while backing up to insure data integrity.
- If you use an outside service to support your network and computer infrastructure find out about their Disaster Recovery Plan for your site and what you can expect of them.