Legato Co-StandbyServer 2000

This software creates a fail-over cluster using two Windows 2000 servers.

Under normal circumstances, one of these machines acts as the primary server

providing access to applications and resources for the clients. If due to some

reason, the primary server goes down, the second server automatically takes over

the primary server’s role. This ensures that network resources are always

available to the clients. This makes it suitable for applications where you

can’t afford to have any downtime.

Some of the resources that can be managed using Co-StandbyServer are data,

applications, IP addresses, services, server aliases, storage devices, and

printer queues. If any of these resources fail on the primary server, they can

be made available on the secondary server. Both servers must have two network

cards each. One pair acts as a dedicated link between the two servers, while the

other connects to the main network. Both the links were 10 Mbps full duplex. The

best part is that the cluster machines don’t need to have similar

configurations.

We tested the server by mirroring a partition on the two servers. Initially,

the server takes a while to mirror the complete data between the two partitions.

While the actual server configuration is not very difficult to do, it took us a

while to have it up and running, because of the documentation. It covers the

prerequisites, installation process, and descriptions of the various menu items.

However, a step-by-step example illustrating its functions would definitely make

things simpler.

Once the cluster was up and running, it worked like a charm. We tried a

number of tests on it right from a manual fail-over to an automatic one from one

server to another. We pulled out the network cable of the primary server, and it

took about 22 seconds to show the secondary server take complete control. We

also switched off the primary server, and the failover to the second server

happened automatically. Another interesting feature is that the failover is

bi-directional, meaning control can be shifted from any server to any server in

the cluster.

The Bottom Line: The beauty of the software is that you don’t need machines

with the same hardware to build the cluster. It does automatic failover to the

second server in case the primary server fails. Overall, a useful software for

high-availability application requirements.

Anil Chopra and Santhosh Unny at PCQ Labs