LAROKE Sample Backup Procedure

("Lawyer stuff" in a nation with more lawyers than the rest of the world combined, disclaimers are sadly necessary)

NOTICE OF LIABILITY: All information presented herewith is provided on an "As Is" basis, without warranty or the implication thereof. Neither LAROKE Microcomputer Consultants nor the individuals associated with the preparation of the information supplied herein shall have any liability to any person or entity with respect to liability, loss, or damage caused or alleged to be caused directly or indirectly by the instructions contained within this presentation.

INTRODUCTION:

This is the backup procedure I use for my major client. The client's Network consists of ten PCs with a tape drive connected to the main file server but the procedure works as well for one or two PCs. There is no "one correct procedure" for backups. A user, therefore, must adapt a procedure optimized for his/her particular situation. S/he must balance risk of data loss against convenience and find a comfortable medium between the two. At one extreme a user makes no backups at all. This is the highest risk and the most convenient solution, and it only works if nothing ever goes wrong. At the other end of the spectrum, the user backs up the files immediately everytime they are modified to a secure off-site location and medium. Except in "mission-critical", highly-automated systems this solution is impracticable. It removes all risk, but the user wouldn't have any time for real work because s/he would be busy performing backups full-time.

PREPARATION:

Keep a "Backup Log"
Keep spare backup media at hand (blank tapes if you are backing up to a tape drive, etc.)

STEP 1:

Determine the risk/convenience ratio as explained in the INTRODUCTION above. In this example the risk is one week. Backups are made every saturday when the office is closed. This is convenient for the users, but if a "system meltdown" occurs on a friday, a complete week's worth of data is lost and must be reentered. The individual user may elect to make backups bi-weekly or daily as an alternative solution with less risk.

STEP 2:

Setup a backup schedule based on the number of PCs in the system and the risk/convenience ratio determined in STEP 1. Three examples of schedules follow:

Sample Schedule for One PC	FULL Backup	DIFFERENTIAL Backup
Week 1	PC 1
Week 2		PC 1
Week 3		PC 1
Week 4		PC 1
Week 5	PC 1
Week 6		PC 1
Week 7		PC 1
Week 8		PC 1

Sample Schedule for Three PCs	FULL Backup	DIFFERENTIAL Backup
Week 1	PC 1	Network
Week 2		Network
Week 3	PC 2	Network
Week 4		Network
Week 5	PC 3	Network
Week 6		Network
Week 7	PC 1	Network
Week 8		Network
Week 9	PC 2	Network
Week 10		Network
Week 11	PC 3	Network

Sample Schedule for Ten PCs	FULL Backup	DIFFERENTIAL Backup
Week 1	PC 1	Network
Week 2	PC 2	Network
Week 3	PC 3	Network
Week 4	PC 4	Network
Week 5	PC 5	Network
Week 6	PC 6	Network
Week 7	PC 7	Network
Week 8	PC 8	Network
Week 9	PC 9	Network
Week 10	PC 10	Network

A "FULL" backup backs up all the files and resets (toggles off) the file archive attribute so that the backup software knows the the file has not changed since the last full backup. When application programs make a new file or modify a file the file archive attribute is set (toggled on). A "DIFFERENTIAL" backup backs up only files that are new or have been modified (file attribute is on) since the last FULL backup and does not change the file attribute. This may seem somewhat technical, but is mentioned because there are some unfriendly backup programs out there that expect you to know about file archive attributes.

STEP 3:

Start a backup log and record an entry for every backup as it is performed. The purpose for this is to allow for quick restoration of files if it ever becomes necessary without the necessity of searching through many tapes. An example of possible log entries for a three PC Network follows:

DATE	DESCR	BACKUP TYPE	MEDIA SET
9/7/96	Sys 1 complete backup	FULL	Full Set A
9/7/96	Network modified files	DIFFER- ENTIAL	Mod Set A
9/14/96	Network modified files	DIFFER- ENTIAL	Mod Set B
9/21/96	Sys 2 complete backup	FULL	Full Set A
9/21/96	Network modified files	DIFFER- ENTIAL	Mod Set A
9/28/96	Network modified files	DIFFER- ENTIAL	Mod Set B
10/5/96	Sys 3 complete backup	FULL	Full Set A
10/5/96	Network modified files	DIFFER- ENTIAL	Mod Set A

The reason for FULL and DIFFERENTIAL backups in this procedure is that if a backed up disk drive fails or a file(s) is corrupted and has to be restored from tape, only two backup sets have to be accessed for the missing file(s), the last FULL backup and the last DIFFERENTIAL backup.

STEP 4:

Determine how much old backup data you want to keep on hand. This is a "cost/benefit" decision. Backup tapes are not inexpensive, but if the same ones are used over and over, they wear out. If a tape breaks or otherwise malfunctions, and there are no other backups of the same data, you are in the same sad situation as if you had never made backups in the first place. Also if a file has been modified over a period of time or maybe even erased in the distant past, you may be able to retrieve the original file from an old backup set. As a general rule, I do not recycle tapes until the data on them is at least a year old.

STEP 5:

Perform the weekly (or bi-weekly, or daily, etc) backup, fill-in the log book, and store the backup tapes "off-campus" or at least in a fire-proof safe in a different room from where the PCs are located (if fire or theft occurs, the backups are no good if they are destroyed or stolen along with the PCs).

STEP 6:

TEST YOUR BACKUPS! Most tape backup software provides for comparsion of the files on the backup tape against the original files immediately after the backup. This process makes a backup take longer, but it is good insurance against sad surprises later when you try to restore a file. The comparison process is not infallible, however. It sometimes misses faults. For more protection, it is good policy to restore a few files or directories after a backup as a test. One way to do this is to make a test directory on the hard drive being backed up (before the backup, of course), and copy a few good files and subdirectories to it. After the backup (and comparison) is finished, restore this "test" directory, then open the files to determine if they are good.

CONCLUSION:

You could write a college thesis on backup strategy and there is a vast amount of information on the subject. This procedure has been kept simple, and relatively brief. It should be enough to get someone new to backups safely "up and running". If you monitor your backups, it's an excellent time to research procedure refinements =^)