Everyone either knows (or suspects)
that they’re supposed to
do backups, yet somehow they don’t always get done, or if they
do, they are very spasmodic. Hopefully, after you’ve read this
article, you’ll “Get religion” about doing your backups.
Some day you will be thankful you did!
Data Protection. Most computer consultants (including those in the
business of selling antivirus programs) will tell you that the best
data protection you can have is a good backup system. This article
will describe this most important part of the data protection process:
Backups. This is not to imply that you should not also be considering
virus protection, firewalls, data encryption, and physical security
for your computing systems. But it is to say that your first line of
defense of your data is a good backup system.
Backups. This article will describe what backups are, why to do them,
how to do them, and when to do them, and will give a recommendation
as to the backup program to use, and the media.
What is a backup? Simply put, a backup
is a copy of a computer program or data. This copy might be on the
same hard disk
drive that the original
data is on, another hard disk drive either in another computer on a
network, or on a removable hard disk drive, on zip disks, optical disks,
CD-ROMs, DVDs, magnetic tapes, or “floppy” disks. In the “network” realm,
there are even companies that (for a fee) will let you backup your
data to their hard drives. For purposes of this article, we will refer
to the class of all backup devices as “media.”
Why should I do backups at
all? Well
first, let’s ask the obvious:
Aren’t personal computers much more reliable now than they used
to be? Well, yes, they are more reliable. But are they indestructible,
invincible to such interruptions as power failures and less-than-perfect
operating systems, impervious to fire and mischief (like viruses),
foolproof to human tinkering? Of course not! So we do backups because
there is a probability that at some point in time we will need to restore
our irreplaceable data from a copy we have made (hopefully recently!)
What do I need to backup? Do you need
to backup your programs? No. Anything that you have original copy
of, and can restore
from that
source does not need to be backed up. So that leaves mainly the information
that you add to the computer. This could include your medical billing
information, of course. But it also might include letters you have
written, notes you have made, in short, anything that you have put
into the computer that would be difficult or too time-consuming for
your to re-create, if you were to lose it. Another way of looking at
it: If I were able to choose which data in the computer I would lose,
and which I would keep, what could I least afford to lose? That’s
(at least) what you should be backing up!
How often do I backup? The rule of
thumb to use here, is to ask yourself the question: How many day’s
worth of work am I willing to re-do? Our recommendation is that
you do a backup at least daily.
“How
do I backup?”—Our recommendation for a backup
protocol: (But before we get to that, make sure you are backing up
the correct data files. Yes, we know this sounds silly, but you’d
be surprised how many times we’ve gone to help someone restore
their data files, only to find out they had been backing up an old
data set, instead of the current one, or just the program files,
instead the crucial data files! So check and double-check your backup
specifications to make sure you are getting your current data files.)
The protocol that we’ve found over the years to offer the
most in covering all reasonable possibilities of data loss, without
going to extremes is as follows: Let’s say for the sake of
generality that your office is open 5 days a week. And let’s
also say that the last day of your workweek is Friday, and the first
day is Monday. We would recommend that you have six (6) backup media.
Four (4) of the media we would label as a “daily” backup
with the name of the day we will use it: Monday through Thursday.
The other two media, which we will call rotating, we would label “Friday
Full Set 1” and “Friday Full Set 2.” By “rotating,” we
mean that we will not use the same media on this Friday that we used
last Friday, but will alternate. All of the media we use will be
reusable. By this we mean that as Monday rolls around, we will pull
out the Monday media we used last week, and use it today. The same
goes for Tuesday through Thursday media.
We start the backup process the first time we do
a backup, by running the “Friday Full Set 1,” even if it’s Wednesday
morning. The Friday sets are what we call “Full” backup
types, meaning everything we select to be backed up will be backed
up, regardless of whether anything has changed since the last time
we ran a full backup. The other kind of backup type we recommend
is a “Differential” backup. This means the backup will
only backup files that have changed since the last time a full backup
was done. We recommend using a backup program that is capable of
doing both “Full” and “Differential” backups.
Such programs will have the capability of “marking” all
the files backed up in a Full backup session as having been backed
up.
For each of the “daily” media we will backup using a “differential” backup
mode.
Appending vs. Overwriting backup media: Most backup
programs capable of doing Full and Differential backups are also
capable of giving
you the choice of “appending” one backup to other backups
on a media, or “overwriting” the media with the new backup. “Appending” a
backup is like the word implies: tacking a new backup at the end
of the previous backup(s), while “overwriting” is replacing
the existing backup on a media with the new backup. Whenever possible
(given sufficient space on your media) we recommend “Appending” backups,
for two very important reasons: (1) It maximizes the number of backups
you can have at your disposal at any given time, and (2) you more
fully utilize the backup media by writing over more of the surface,
thus extending the life of your media. When you do an “appending” style
of backup, sooner or later the media will become full. Your backup
program will inform you of this, and ask you to insert a new media. At that point, you would cancel the backup, and re-start it
this time choosing the “Overwrite” option.
You would then have a mostly-empty media to which to begin again
the appending process.
(If your office is open more than five days a week, you would adjust
the above figures we give here by adding one media for each extra
day you are open, and label them accordingly.)
Should I use an automatic
scheduling program for doing backups? Many backup programs give you the capability
of
backing up your data
automatically at a certain time of day (usually late at night). All
you need to do before you go home for the evening is insert the appropriate
media for the day of the week. Sounds simple, and it is simple. However,
our experience over the years with offices that have used the automatic
scheduling is that more often than we care to remember, those automatic
backups have not been as foolproof as a manual backup (i.e. one where
you, the user inserts the media and starts the backup program). What
often happens, is that the user becomes complacent, and doesn’t
do the checking of the backup each morning, assuming that everything
is running fine.
Can I backup to the same
media over and over? The answer is a cautious, “yes,
but.” You can write to one media over and over, which is better
than not backing up at all, but not much! Just as your hard disk
drive is not invincible, any backup media you decide to use will
likely be just as vulnerable, if not more so than your hard drive.
Our recommendation is that you have a separate media for every day
that your office is open for putting data into your computer, and
one extra for the end of the week. And this includes weekends, if
sometimes someone works on the weekends inputting or changing data.
So
what “media” should
I use? Here’s
our list in order of our preferences for most offices:
Zip or Optical Disks—Advantages: Relatively inexpensive,
reliable, fast to use, and capable of being taken “off-site.” Also
being essentially another “hard drive” on your computer,
you can use any backup software to backup to them. Disadvantages:
Somewhat limited in storage space (although the largest Zip Disk
at 750MB is more than adequate to store most practice’s data,
especially if you use the compression option offered by better backup
software.)
Hard Disk Drives—Advantages: They are fast,
have large capacity, and can use most backup software. Disadvantages:
Relatively expensive
(though not as much as tape systems), and not as portable as Zip
Disks for taking off site.
Magnetic Tapes—Advantages: They can store huge quantities
of data, and can be “linked” so if one gets full another
can be loaded and continued to write to. The software that generally
comes with a tape backup system tends to be robust, and full-featured.
Disadvantages: Tape systems can be costly, as are the media, and
the backup time is relatively slow compared other media.
CD-ROMs and DVDs—Advantages: They are cheap
and have large capacity. Disadvantages: They are relatively slow,
and the software
provided to do the backups is rather limited in capabilities. Also
they are less reliable than either zip disks, optical disks, tapes,
or hard drives. They are fine for archival purposes, but not for
your first line of defense against data loss!
Internet Backup Services—Advantages: They are off-site, meaning
you are protected against natural and man-made disasters to your
data. Disadvantages: Most impose monthly fees, and the backup times
are not as fast as some other media—you are limited to the
upload speed (and download if you need to restore) of your Internet
connection. They are also out of your control. What happens if they
are not available when you need them?
Floppy Disks—Advantages: They are cheap. Disadvantages: They
are limited by their small capacity (1.2MB), and hence are not suited
to backing up today’s large volume of data. Also many new computers
are coming without floppy disk drives. Their reliability has also
become more and more suspect.
What about Backup Software? Many backup media systems come with
their own backup software. And you are also able to use your own
favorite backup software with many media, especially Zip Disks and
Hard Disks. What you should be looking for as a minimum in your backup
software is: 1) capability of doing full backups and marking the
archive bits of the files it has backed up that way; 2) capability
of doing Differential backups and not marking the archive bits; 3)
capability of compressing data files while it backs them up.
Restoring your data from a backup:
A backup program capable of doing Full and Differential backups
will also be capable
of doing a “Restore” of
the data, when you need to do that.
When to restore: Whenever your current
data is no longer accessible, or is not correctly accessible.
In the first
situation you might
attempt to open a letter you have written, only to find out that
either it’s not there, or you get an error message saying
it can’t be opened. In another scenario (one we all hate
to even contemplate) you lose your computer through theft or fire
or some
other natural disaster. In the latter situation, where the information
is not correct: Let’s say you have been using your medical
program, and when you go to retrieve some information that you
know you’ve entered, you get an error, saying the information
is not there. (There are other steps you would take here first,
probably,
like calling your computer-support personnel, to see if the information
was recoverable. But for the sake of this article, we’ll
assume that it was not, and you have decided it’s time to
restore your data from a backup.)
How to restore: The general rule
to follow here is:
(1) Restore with the latest Full
backup you have first, and if you
have been doing Differential backups, follow that Full backup
by the latest Differential backup you have (that was dated after
the
Full backup you have restored). For example, let’s say it
is Thursday, and you’ve decided to restore your data. The
last full backup you did was the preceding Friday, so you restore
that first. Then you restore yesterdays, or Wednesday’s Differential
backup.
(2) Test to see if the problem is corrected.
(3) If it is not corrected, retreat to the next latest Differential
backup. Check to see if the date of that backup is still later
than the Full backup you have restored. If not, then retreat
to the next
Full backup you have, and restore any Differential backups you
have that are newer than that, and that you have not already
restored.
(4) Repeat this process until the
problem is solved.
(5) Finally identify what data is missing from your restored
data set, and begin the process of re-entering that information
back
into your data files. Hopefully, if everything goes well,
you will have
only part of a day, or a full day at the most to re-enter.