The two main elements of disaster recovery include
fault tolerance and data recovery. A good combination of both is recommended.
We can provide server clustering, hardware raid, redundant power supplies,
and server backups.
To prevent a service interruption due to server
failure, a server cluster can be set up. A server cluster is created
when two or more servers are configured to provide the same service
(Web, SQL Server, etc). This cluster of servers is seen by the outside
world as one logical server, such that failure of one server does not
result in loss of service. In addition to mitigating the risk of server
failure, running a clustered solution eases the task of server maintenance.
Patches and security updates can be made to the individual servers,
even those requiring reboots, without affecting the overall level of
service.
Another type of hardware failure involves the
loss of a physical disk. On normal personal computers with only one
hard drive this event is certainly disastrous. The only steps that can
be taken are to replace the affected drive and to restore the data and/or
operating system from backup. On servers, the option exists to enable
a disk configuration that provides a greater degree of fault tolerance.
An array of physical disks can be set up so that failure of one physical
disk does not cause server failure. The bad drive can then be replaced
with zero downtime.
Finally, a common problem for servers and in
fact for all computers is the risk of power outage. Our data center
mitigates most of this risk by providing multiple city power sources,
conditioning the power feed and providing extensive battery and diesel
backup power supply. However, on most servers there is still a single
point of failure where the electricity enters the server. You can reduce
your risk by providing redundant power supplies. In the event one power
supply fails, the other is sufficient to power the server, while the
bad power supply is replaced.
A necessary piece to every data restoration strategy
is a backup of part or all of the data contained within your server.
Backup copies of your data within the server itself will guard against
accidental deletion and file corruption, but will not alleviate the
risk of data loss due to hardware failure. Therefore, the backup data
needs to be physically removed from the server being backed up. This
can be done using tape media and / or remote hard drive arrays.
