OpenEdge Getting Started:<br />Database Essentials Creating a complete backup and recovery strategy

A great deal of time and money are spent on backup and recovery strategies. However, they are often not tested and revised based on the discoveries made.

A complete backup strategy should try to balance the probability of the problem occurring with the amount of data loss in a given situation and the amount of time and resources spent on backups. A disk failure is a relatively likely event compared to a fire or a flood. This is the reason to have redundancy at the disk level—to reduce the probability of failure. A fire or flood is less likely, and most people understand that they would lose some data should such an event occur.

It is also important to include the users in the disaster-planning process. They are the real owners of the data and can help with the probability/data loss/cost trade-off decision.

It is possible to provide near-100-percent reliability, but there is a large cost in doing so. Take the case of an emergency services organization. It needs as close to 100 percent reliability as possible, since a failure on its part could cost lives. It must have complete, duplicate systems at a second location with available staff in case the primary location is knocked out by a disaster. Your situation might allow you to lose some data and availability of the system in a trade-off for lower cost. It is important to focus on the entire system and not just your databases and applications. You must also weigh the cost/benefit of each strategy. While you are doing this analysis, you must include the people who run the business to determine their requirements. It is sometimes helpful to put a price on lost data and downtime because it makes it easier to do the final cost/benefit analysis.

The following sections identify the key areas to consider when devising a complete backup-and-recovery strategy.

Who does the backup?

In most cases the system administrator performs backups of the system on a regular basis. When the system administrator is unavailable, other personnel must be responsible for the backups. It is important to have a documented process so that backups are created consistently.

What does the backup contain?

An application consists of many diverse components, including databases, application code, user files, input files, third-party applications, and so on. Remember that your application is made up of your OpenEdge installation with its associated service packs, your application source and object code, middleware such as the AppServer, and associated operating system files.

The best way to determine what needs to be backed up is to walk through the vital processes within your organization and note activities and systems such as:

The systems that are involved.

The software application files.

The data that is used throughout the process.

Where does the backup go?

The media that you use for backups must be removable so they can be archived off site to better protect data from natural disaster. Consider the size of your backup in relation to your backup media. For example, tapes with a large storage capacity are a practical and reliable option to back up a 20GB database.

Tape compatibility is also a consideration. You might want to use the backup tapes on more than one system. This will allow you to back up on one system and restore to another system in the case of a system failure. A Digital Linear Tape (DLT) is supported on many platforms and can be used to help move data from one system to another or to retrieve an archive.

Archiving off site is as important as the backup itself. If a fire, flood, or other natural disaster destroys your building, you can limit your data loss by having your backup at a separate location. This can be a formalized service, or as simple as placing the completed backup tapes at a second office. It is important to make sure you have access to your archives 24 hours a day, seven days a week.

How to label a backup

The name of specific items stored on the tape. A tape labeled “nightly backup” has no meaning without the ability to cross reference the items contained in the nightly backup.

The date and time when the tape was created. In situations where multiple tapes are made in one day, you must know which tape is more current.

The name or initials of the person who made the tape, to ensure accountability for the quality of the backup.

Instructions to restore the tape. There should be detailed restore instructions archived with the tape. The instructions should be easy to follow and might include specific command information required to make the backup.

Volume number and total number of volumes in the complete backup set. Labels should always read “Volume n of n.”

When do you do a backup?

Perform a backup as often as practical, balancing the amount of data loss in a failure situation with the interruption to production that a backup causes. To achieve this balance, consider these points:

Static information, like application files that are not being modified, only need to be backed up once a week, or less often.

Most database application data should be backed up at least once a day.

In cases where data is backed up once a day, it is possible to lose an entire day’s work if the disks fail or some natural disaster strikes at the end of the day. If you performed multiple backups throughout the day but only archived once a day, you would be better protected from a hardware, software, or user error, but your protection from most natural disasters would be identical. By moving the intra-day tapes from the machine room to a different portion of your office, you decrease the probability of a fire in the machine room destroying your tapes.