When you have identified the high availability features that you plan to use in your topology, you can proceed with the configuration tasks for each of the components.
Prerequisites to configuring high availability
Before you can configure your system for high availability,
you must complete a number of prerequisite tasks and determine the
best approach for installing your product.
Resources: IBM Redbooks and product support sites
During the high availability configuration process, you
must refer to specific configuration information for each of the components
in your high availability topology. The information resources are
listed by component and product for easy reference.
Configuring the database
To adapt the database product or feature that you use in
your highly available topology, you must perform specific configuration
steps. You can choose a database configuration to suit the requirements
of your organization.
Configuring the directory server
You can configure Tivoli® Directory Server for
high availability by installing cluster management software, such
as IBM® Tivoli System
Automation for Multiplatforms.
Proxy servers can be set up to help provide seamless failover when
a directory server node fails.
Configuring the application server WebSphere® Application Server can
be configured in a cluster to achieve high availability. You can select
from various cluster configurations, such as a load-balance cluster,
or a cluster with shared disk.
Configuring the HTTP server
You can configure the HTTP server to make it highly available
and you can set system properties to alleviate the symptoms that result
from failover of the HTTP server.
Configuring attached documents for high availability
It is important to configure your attachments for high
availability, as they can be a critical feature of the product for
many organizations. You must configure a shared disk that can be accessed
by all of the application server and the HTTP server nodes in the
topology. The shared disk must also be highly available, so that
it does not become a point of failure.