Custom Search

Friday, February 1, 2008

How to prevent data loss

Forget aboutmotherboards,processors,memorymodules and graphics cards
- the most important part of any computer system is the data it holds. Should
data loss occur, it could take ages to re-install those bloated applications, let
alone any vital files and documents. But time is money, and in today’s business
environment the information and data a company stores is often the basis of its
competitive advantage. With the temporary loss of such data, a company can
quickly lose its cutting edge. To lose it permanently will almost certainly bring
about severe disruption; so much disruption that many companies will never
recover from such a loss.
While disk drives are more reliable than they have ever been, disk failure is not
the only reason for misfortune to strike. Apart from fire, flood, and other “acts of
God”, the system administrator has to look out for a multitude of potentially
disastrous possibilities, including the loss of data from computer virus, human
error, theft, and deliberate vandalism.

Have A Recovery Plan

How quickly a company gets up and running after such a “disaster” depends
largely on the precautions it has taken beforehand. After all, it is far better to
formulate a recovery plan before the event, rather than later. The type of recovery
plan chosenwill depend not only on the level of failure or downtime the company
is prepared to accept, but on howmuchmoney it is willing to spend on a recovery
strategy.A real-time fault-tolerant system - preferably achieved by remote server
clustering - should ensure continued operation after failure occurs. If a server
were to go down, a real-time fault-tolerant systemwould automatically switch to
another server or system. No data would be lost, and the least amount of
disruption would occur.
But real-time fault tolerance is not cheap, and neither is it always perfect. Typically,
when a file is deleted froma real-timefault-tolerant server, the server cluster
also deletes the file. Recovery, therefore, is hardly “real-time” because it usually
takes several minutes to restore the file from its backup set. Moreover, the file
may not be identical to the lost file; it could be hours, or perhaps even days old.
A short-time fault-tolerance strategy is a far cheaper alternative. This type of
strategy is best for businesses that can survive a downtime of around two or three
hourswithout grinding to a halt. During downtime all company files and records
are handled manually or moved to another server or workstation until the
problem is fixed.

Protection

At least Windows NT offers its own degree of fault tolerance. Some of these
protective features work transparently, while others, like the Emergency Repair
Disk and Last Known Good configuration utilities, are designed to get the system
up and running again with the minimum of downtime. The Emergency Repair
Disk is basically an NT-formatted floppy containing the files found in the %SYSTEMROOT%\
REPAIR folder. Although the first disk will have been created
during the initial NT installation, it can be updated thereafter using the RDISK
utility from the command prompt. It is important to note, however, that for
reasons of size - this could be several megabytes on certain systems - the RDISK
utility does not update the SAM and SECURITY hives. Employing RDISK with

Data loss can be
devastating to any
business, but recovery
is much easier if a
contingency plan is
formulated before the
event, rather than after.
www.pcnetworkadvisor.com

the /S switchwill copy these hives to the repair folder, but they should be backed
up as part of the normal backup procedure as well.
Typically, the Emergency Repair Disk contains enough registry and file setup
information to return the system to a bootable state. However, the disk must be
kept up to date. It should be updated before any major changes are made to the
system, and then updated again after the changes have been made and the system
is fully operational. The Last Known Good feature can also be a quick recovery
tool. It works because most successful NT boots are cloned to the LastKnown-
Good entry in the system registry. Thus, hitting the space bar when NT bootswill
invoke the Last Known Good configuration, effectively ignoring the new driver
or system settings responsible for having caused the problem.
More recovery options can be found in the System Properties Startup/Shutdown
tab, accessed by clicking on the Control Panel system applet. For example, the
system can be set to reboot immediately after a crash and without the need for
anymanual intervention. This is useful should the crash occur at a time when the
server is not being monitored.

RAID

Hard drive integrity is, of course, essential to any system. Hardware RAID in
particular can offer superb protection from disk failure. This is because data is
spread or striped across a set of disks, improving throughput and protecting the
data held on the array against the failure of any individual disk. There are six
different ways to configure an array. Levels 0, 1, 3 and 5 are the most popular,
though each level contains benefits and drawbacks.However, even the bestRAID
is not completely infallible. Two disks could fail together, for instance, or the
controller could develop a fault. Nevertheless, a properly configured RAID
should be considered an invaluable asset, especially for the larger network.

UPS

Data protection hasmany facets, of course, andwhen looking at the larger picture
it is easy to forget the one thing thatmost of us take for granted - the power supply.
Indeed,many experts believe that external power failures and surges account for
up to 40% of all system failures resulting in data loss. The best way of avoiding
the problem is to install an uninterruptible power supply - commonly known as
aUPS. Basically, the average UPS is a box containing lead-acid batteries, complete
with an inverter to convert direct current from the batteries into an alternating
current.
The most basic UPS will protect a system from any sudden and unpredictable
power surges. More expensive units are designed to cope with a total blackout.
In such cases, a good UPS will keep a system up and running long enough to
administer orderly shutdown procedures. For networks that need to be up every
minute of every single day, even more powerful UPS units are available. Typically,
these UPS units will run for several hours, long enough to give an emergency
generator time to kick in.

Backup Strategy

As always, a solid backup strategy is vital to any recovery plan and, for sheer
convenience, tape is still hard to beat. However, because today’s hard disk
capacities are huge it would be extremely time-consuming to perform a full
backup more often than is necessary. Some small businesses, for instance, may
require a full backup only once a week, with differential or incremental backups
making do in between.
A differential backup copies files that have changed since the last full backup.
Although this type of backup is not as time-consuming as creating a full backup,
both the full and differential backup tapes are required to fully restore a system.
Incremental backups are quicker still, since they contain only the files that have
changed since the last backup of any description. When it comes to the restore
process, however, incremental backups are not quite so convenient because
several backup tapes may be needed to get the system back to normal.
Themost popular backup strategy is based on a three-generation system,with 21
tapes used to back up data for 12 months. This strategy is commonly known a
the grandfather, father and son set. Typically, a full backup is taken at the
beginning of each month (grandfather), another full backup at the start of each
week (father), with differential or incremental backups in between (son). By
following this approach it is possible to create day-to-day backups of the current
week,week-to-week backups of the currentmonth, andmonth-to-month backups
of the current year. Backups should always be verified and the restore process
tested frequently.At the end of the year, themonth and year tapes can be archived
before starting afresh with new ones.
At the very least, backups should be stored in a fireproof safe. Even so, this is not
ideal. After all, what good is the latest backup if it cannot be accessed? Typically,
a fire-ravaged building could be so badly damaged that it is off-limits to everyone,
and it could be days or even weeks before access is granted. Hence the best
solution is to keep backup sets safely stored offsite. One way to achieve this is to
back up crucial files across the Internet, and a growing number of providers now
offer this useful alternative. For example, @Backup (www.backup.com) provides
a 30-day free trial and charges a flat US$99 per year for 100 MB of storage.

Disk Imaging

As an extra form of insurance, utilities that transfer the entire contents of a disk -
or specified folders - over to another drive are also worth considering. Power-
Quest’s Drive Image Pro 3.0 (www.powerquest.com), for instance, creates an
exact image of a hard drive or a hard drive partition. This image can then be used
to deploy multiple Windows workstations, upgrade existing workstations, or
back up and restore hard drives. The utility is primarily designed to protect data
from amajor hard disk crash, but since it supports selected file restore, it can also
be used as a belt-and-braces approach to backups. At its highest compression
ratio, Drive Image Pro offers around 50% compression; thismeans that, in theory,
a 22 GB drive that’s three-quarters full could have its entire image saved to an
empty 10 GB drive.

Virus Threats

Thanks to the growth of the Internet and our increased reliance on email, the
threat of losing data fromviral infection isnowgreater than it has ever been. There
are, of course, many different types of virus, but it is the relatively recent threat
of the macro virus, sent as an attachment to email, which is presently causing
Issue 126:January 2001 File: P1528.3
Page 19 Problem Solving:Software
Prevention Is Best
Today’s hard disks should provide many years of trouble-free service. They
rarely fail, butwhen they do the results can be catastrophic. Therefore it pays
to be aware of potential problems before they become a real threat.
l Back up data regularly. Store backup sets in a fireproof safe, and off-site
wherever possible.
l Use anti-virus software and maintain regular updates.
l To help safeguard against boot sector infections, disable the “boot from
floppy” option in the system BIOS.
l Store important files in a single location such as the My Documents folder.
Then use subfolders to help organise data by project or category. Old data
can be moved across to a specially created Archives folder.
l NT users should maintain an up-to-date Emergency Repair Disk.
l Format newNT volumes asNTFS. The NTFS format includes transactionlogging
capabilities to help prevent data errors.
l Train users to report any unusual noises immediately. Tapping, clicking
or humming sounds are early signs of disk or controller failure.
l Avoid excessive heat. The latest hard drives spin almost twice as fast as
their older counterparts, so make sure the system fan is up to the task.
Consider fitting an additional fan.
l Store an emergency installation of the OS on a different drive.
l For optimal server setup, use mirrored boot drives with all data maintained
on a hardware RAID5 system. Depending on the nature of the data
involved, consider RAID 5+0 for increased speed.
“Many experts
believe that external
power failures and
surges account for up
to 40% of all system
failures resulting in
data loss. The best
way of avoiding the
problem is to install
an uninterruptible
power supply.”
PC Network Advisor
www.pcnetworkadvisor.com
How To Prevent Data Loss
most concern. At least most of the latest anti-virus products can examine attachments
as they sit in the post office. Other applications can work in conjunction
with Internet firewalls to search incoming mail for viruses. Alternatively, configuring
the system to save all attachments to a “safe haven” can easily solve the
problem. This could be to another drive, for instance, which is set to automatically
scan for viruses every time a file is written to it.
Using anti-virus software is every bit as important as keeping regular backups,
and the anti-virus software must be updated regularly. Meanwhile, the whole
workforce should receive regular anti-virus training and be reminded constantly
of the dangers of virus infection.
Data Recovery Tools
Whatever the cause, it is just as well that data loss is not always on a large scale.
When it comes to the occasional file deleted by accident, there are few utilities as
useful as Executive Software’s Undelete for Windows NT. Available as a free
download (www.execsoft.com), the utility is part of a fully featured Undelete
package that changes the Recycle Bin to a Recover Bin, making it easier to retrieve
files no matter how they have been deleted.
A number of manufacturers provide do-it-yourself solutions in the form of data
recovery software. But utilities such as these should be treated with caution since,
in the wrong hands, they can be responsible for losing data permanently. Not so
Search & Rescue, the enterprise version of PowerQuest’s Lost & Found data
recovery application (www.powerquest.com). PowerQuest claims that as long
as the disk is still spinning, Search & Rescue can locate and recover almost any
file, anywhere on the disk. Even accidentally formatted disksmay be cajoled into
surrendering lost data, though only as long as the data hasn’t been overwritten
first. Importantly, no additional hard disk damage or data loss can occur because
Search & Rescue never writes to the impaired disk. Once Search & Rescue locates
the data, it is left to the user to decide where to transfer and restore the recovered
files to.
Recovery Services
Recovery utilities often offer a quick and cheap solution to data loss problems,
but they are not always the best option. When a company’s very survival is at
stake, the common sense approach is to call in a data recovery service sooner
rather than later. Professional recovery experts can recover data from virtually
every situation imaginable, and without voiding equipment warranties. Indeed,
many of these organisations claim that around 95% of all inaccessible data can be
recovered. Be aware that most companies offering this type of service advise users
against the do-it-yourself approach, claiming that the use of data recovery utilities
may decrease the chances of a successful retrieval. Others, such as Ontrack Data
International (www.ontrack.com), a company with a long and established track
record, actually provide both DIY and remote solutions aswell as in-lab capabilities.
Conclusion
Unfortunately there is no such thing as a single, foolproof way to protect systems
from data loss. But with a strict backup regime in place and a carefully thought
out disaster contingency plan, the threat of losing data permanently can be
reduced to almost zero.
Issue 126:January 2001 File: P1528.4
Page 20 Problem Solving:Software
Further Information
Data recovery information
and services
www.datarecoverygroup.-
com
Data recovery FAQs
www.cbltech.com
Daily backups over the
Internet
www.netstore.net
Crash protection software
and anti-virus products
www.symantec.com
RAID technology information
www.rising-edge.com
RAID 1 IDEdiskmirroring
www.arcoide.com
High-performance UPS
systems
www.opti-ups.com
Copyright ITP, 2000
PCNA
PC Network Advisor
www.pcnetworkadvisor.com
How To Prevent Data Loss
Click here for more free networking guides

No comments: