A backup system must be highly automated to ensure
that jobs are launched consistently and correctly, but a backup system
also needs to be watched over diligently to make sure it continues
to function reliably. Unfortunately, monitoring the backup system
generally isn't a priority until something goes wrong. By then it's
too late.
People have a strong tendency with a backup system to
set it and forget it. Automation clearly has many benefits, but a
total hands-off approach can be very dangerous if no one is
overseeing the process.
Test and Then Test Again: VERY Important
With any newly installed backup system, don't assume
everything works correctly right out of the box. Even more important,
don't take for granted that your backup system will continue working
indefinitely. You need periodically to restore some folders and files
from your backup media to validate that your system still works.
If your automated backup routine is configured to
include a verify run with each backup job, testing a sample restore
job monthly should be adequate. However, if you have an extremely low
tolerance for risk, you may want to simulate a sample restore job once
a week.
|
Note:
A verify job compares the contents of what
was just backed up to media against what resides on the hard
drive. |
The Hazard of Moving Parts and Open Design
Why do you need to take these precautions if you're
purchasing a reliable, business-class backup system to start with?
Typically, a tape drive or other backup device is one of the few
components in a PC or server that still have moving parts.
As a result, it's prone to mechanical failure. In
addition, because a backup device generally is open, as opposed to the
sealed design of a hard drive, it's easy for the inside of the backup
device to attract a significant dust buildup in a relatively short
period of time.
Sample Restore Jobs and Cleaning Tape Heads
Testing a tape for a sample restore job is also a
great time to clean the heads of the backup drive if your backup
system requires this kind of maintenance. Restoring a few hundred
megabytes (MB) of data to a scratch directory and running a
head-cleaning tape should take no more than 15 to 30 minutes.
When running a test restore job, always restore the
data to an alternate server folder path, so as not to disrupt the use
of any shared folders.
|
Tip:
If you use Microsoft Outlook, a personal
information manager (PIM) or a similar scheduling program,
you may want to set up a recurring monthly or weekly event
in your calendar for testing the backup system and running
your head-cleaning tape. |
Building a Backup and Restore Procedure Checklist
In times of crisis, the most crucial issue becomes how
quickly you can get the data back onto your system, undamaged. So, as
you build your backup system, be sure to document your test procedures
into handy checklists.
This documentation also can be great for
cross-training and crucial for avoiding panic during an emergency. Be
sure you have a hard copy of this documentation next to your system
and stored offsite with your backup media.
Watching the Log Files
In addition to running test restore jobs, you must
inspect your backup system log files daily. When the backup system is
first installed, take time to get familiar with the way log files
look when everything is working. This way, if something goes awry,
you'll be better prepared to pinpoint the nature of the problem
immediately.
As network operating system (NOS) suites and backup
software have become more sophisticated, it's now possible to monitor
backup system log files remotely and more proactively. In most cases,
the backup system log files are just plain text (.txt) files. Many
third-party tools and utilities, as well as those included with
Microsoft BackOffice Small Business Server (SBS) and Microsoft
BackOffice Server, can automatically e-mail or fax a backup system log
file at a preconfigured time.
Automatically and Remotely Monitoring Tape Backup
Log Files
Many computer consultants have their small business
clients' log files automatically e-mailed to them daily, so the
consultants proactively can watch out for potential problems with the
backup system.
However, don't think this proactive monitoring is
limited to professional consultants. If your company has one or more
branch offices you support from a centralized location, you also can
use a similar method to monitor backup system health in remote
locations.
For greater flexibility, you can set up an e-mail
alias so the backup system log file automatically is sent to you, your
second-in-command, and perhaps an external computer consultant - so
you are all kept in the loop. Also, this way, monitoring continues
even when you're out of the office or on vacation.
Get
the Sample Action Items for Tip #81
