Redirection and connectivity for data-replication clients

When any server in a high-availability cluster becomes unavailable, any client connections to it should be redirected to an available server. The best way to handle connection redirection is to use Connection Manager to automatically reconnect client connections to the servers that you specify in service level agreements. Alternatively, you can control connection redirection with environment variables or connection information.

If you do not use Connection Manager, you can automatically redirect clients to different database servers in a cluster by configuring applications to connect to the server group to which the servers belong. When you create a connection to a server group, by default the connection is made to the current primary server in the group. If replication is down because one of the servers failed, the connection is made to the server which is online (in Standard mode or Primary mode without a secondary server). You can also automate this action from within the application. Some of the client connectivity drivers included in the have specific mechanisms for automating redirection. For details, see the documentation.

When you design client applications, you must make some decisions on redirection strategies. Specifically, you must decide whether to handle redirection within the application and which redirection mechanism to use. The three different redirection mechanisms are as follows:
  • Automatic redirection with the DBPATH environment variable
  • Administrator-controlled redirection with the connectivity information
  • User-controlled redirection with the INFORMIXSERVER environment variable

The mechanism that you employ determines which CONNECT syntax you can use in your application.