How does fail-over takes place in a cluster definition and what is the FailOver threshold definition?

book

Article ID: 52235

calendar_today

Updated On:

Products

CA Single Sign On Secure Proxy Server (SiteMinder) AXIOMATICS POLICY SERVER CA Single Sign On SOA Security Manager (SiteMinder) CA Single Sign-On

Issue/Introduction

Description:

The Failover Threshold is the number of Policy server that needs to be active for the cluster to be active, if the number of policy server is below this number, the webagent will switch to the other cluster.

Solution:

The FailOver Threashold number that you see in the cluster definition tells the number of Policy Server that needs to be up for the cluster to be active.

For example: 2 policy server with a threashold percentage of 60%, you have 2 as failover threadshold, which means that if 1 of the Policy Server is down, the number of Policy Server available in the cluster is 1, which is lower than 2, so there will be fail over to the second cluster.

If you want the cluster to remain active even if 1 Policy Server is down, you have to set a threshold below 50 % and the number will be 1.

Environment

Release:
Component: SMPLC