Disaster Recovery (Backup/Restore)
Probably the most important administrative
task for Exchange admins is ensuring the effectiveness of their
disaster recovery plan and the integrity of the backups
themselves. In this section, we will focus on the Backup or
Restore Wizard that is included with Windows Server 2003 as it
satisfies the scheduling and backup needs for our scenarios.
Companies that require a more centralized management tool or
additional features man want to seek out 3rd party backup
and restore tools for added scalability and additional
features.
Normal backups of
Exchange Server 2003 databases should be considered a
requirement in nearly every scenario. The frequency of this type
of backup is determined by the backup requirements of your
company and the ease in which you would like to restore your
server in the event of a disaster. Normal backups assist
Exchange by allowing it to clean up transaction logs that have
accumulated since the last normal backup because technically they
are not needed anymore assuming the new backup is valid.
Moreover, a normal backup does not require the additional
application of an incremental or differential backup job. Those
jobs are better suited for situations where you require backups
of changes that were made since the last normal backups. The
penalty for this savings is paid during restoration which (in my
opinion) is where time is of the essence.
Note
Circular Logging can be
enabled on an Exchange database and will certainly control the
growth of the disk volume. While this is an effective means of
controlling growth during large migrations, this is not
recommended for day to day operations as all changes made since
the last normal backup will be lost if the database fails and
must be restored.
Setting up a Scheduled Backup of
Exchange
Scheduling the backups is
crucial to maintaining the Exchange
environment. In a branch office configuration, there are design
options as to the media in which the backups are stored. Tapes
must be exchanged rather frequently and in some instances there
may not be a remote administrator who can swap the tapes on a
daily or even weekly basis. Moreover, during a restoration of a
failed database, a human would likely be required to be present at the remote
site to find and insert the appropriate tapes. One strategy for
handling this scenario is to use a physical hard drive or the
drive on another server as the backup media for a remote server. Those backup files
could then be archived at a later date to long-term storage.
Then, a normal backup would be performed nightly with the
backup files stored on a physical disk. One week?s worth
of data stored in this manner would allow for quick and easy
recovery should you have a server or database failure. This
also provides for archiving backup files to tape or other media
for offsite recovery or archiving purposes.
Note
An effective
disaster recovery program should also provide a backup of the AD
environment and other key databases such as the IIS Metabase and
any other key services. For additional information on backing up
Exchange, see Microsoft?s Disaster Recovery Operations Guide:
http://www.microsoft.com/technet/prodtechnol/exchange/2003/library/disrecopgde.mspx
