There are many possible causes of database corruption. These include:
Improper entry of programs or data
Accidental removal of or damage to files
System halts or hangs
Defective backup media
Power outages or natural disasters
To protect against such problems, all files stored on the disk drive should be regularly backed up onto some offline storage medium, and a well-established procedure must exist to permit orderly recovery should the database become corrupted.
Backing up and restoring the system is also the usual method used to reorganize the disk drive space on an overly fragmented system, as well as to reallocate files that have become either too large or too small. A detailed discussion of this topic can be found in System Maintenance.
mvBase must be installed on the system before any mvBase data can be added. If data corruption is severe, the administrator may have to reinstall mvBase before restoring any data files. Your system supplier will inform you when this is necessary.
The system administrator has two main decisions to make about backing up the system:
What is backed up? The whole data area? Certain accounts?
There are two main types of backup: full and partial.
full backup |
Saves the entire data area of the disk drive. |
partial backup |
Saves only a portion of it. Partial backups can either save one complete account or file, or can be incremental, saving only the changes made to the data area or to one account since the last full backup. |
How often is the disk drive or data area backed up?
This depends on several issues:
The individual installation and specific operating procedures.
The amount of use the system receives—the number of files that are created or altered in a specific span of time.
The amount of data the users can afford to lose. The more important the data is, the more care should be taken with the backup.
NOTE |
If you are not sure how frequently to perform backups, consult your system supplier or software vendor. Rocket recommends that a full backup of the system be performed at least once a week, followed up by daily incremental backups. |
Safeguard system backups by using only high-quality backup media. Low-quality tapes may jam, stretch or break. Low-quality tapes also shed their magnetic particles faster than high-quality tapes, increasing the possibility of data errors and tape drive problems. (The particles cause the tape heads to become dirty and less reliable in reading or writing data to tapes.) Similar problems can be caused by low-quality diskettes. Replace tapes and diskettes periodically as a safety precaution.
Store backup media on-site in a cabinet that is both fireproof and safe from powerful magnetic fields. The system administrator may also wish to create a weekly backup that is stored in a safe place off-site. This will help cut down on the possibility of losing important data in the event of a fire, flood, or similar disaster at the computer site.
NOTE |
Backups written to virtual tape files should themselves be backed up on the host operating system. |
See Also