Problem:

How does fail-over work in EnterpriseLink? What happens when a user is unable to get to the web server because the server is down or not available? Will the software redirect itself to find another server?

Resolution:

EnterpriseLink was designed from the outset to be cluster-aware. It does not rely on Microsoft Wolfpack (a.k.a. Microsoft Cluster Server, MCS) or on Sun components.

The EnterpriseLink Remote Administration screen is used to configure EnterpriseLink Cluster Service. The EnterpriseLink Servers hyperlink contains the List tab, which shows all of the EnterpriseLink Servers in a cluster. Additional nodes are added to the cluster with the Add... button.

The EnterpriseLink Cluster Service also detects when a node is off-line and will redirect new sessions to another node. Using the Max Session Limit on the Remote Administration screen is a way to force a node to quiesce state. Setting these configuration parameters to zero means the node no longer accepts sessions, but will allow existing sessions to finish. Thus, a rolling outage allows upgrade of hardware or software without the cluster as a whole being taken down.