The RCM is an OpenSwitch component that coordinates the failover of a high availability, warm standby environment.
The term “failover” in this document refers to automatically
switching to a redundant or standby server when the currently active
server fails or terminates abnormally. It does not refer to Sybase
Failover, which is a specific feature of Adaptive Server Enterprise.
A high availability, warm standby environment minimally consists of:
A Replication Server configured for warm standby replication
Two Adaptive Server Enterprise servers and corresponding databases
One OpenSwitch server
One RCM instance, configured to coordinate failover through the OpenSwitch server
A redundant high availability, warm standby environment includes a backup and secondary OpenSwitch, and a backup and redundant RCM. A redundant system minimally consists of:
A Replication Server configured for warm standby replication
Two Adaptive Server Enterprise servers and corresponding databases
Two OpenSwitch servers
Two RCM instances configured to coordinate failover through the OpenSwitch servers
The
RCM does not support concurrent coordination modules. When the RCM
establishes a connection to OpenSwitch, OpenSwitch sets the COORD_TIMEOUT to
zero (0), which turns off the coordinated CM functionality.
Figure 4-1 illustrates a redundant system before and after the failover of the active Adaptive Server. Before a failover, application end users connect to the active Adaptive Server through the primary OpenSwitch server, and decision-support-system users connect to the standby Adaptive Server through either the primary or the secondary OpenSwitch server.
After failover, the primary OpenSwitch server switches the application end users to the standby Adaptive Server. The application end users are still connected through the primary OpenSwitch server, but now are connected to the standby Adaptive Server. Decision-support-system users continue to connect to the standby Adaptive Server through either the primary or the secondary OpenSwitch server.
Figure 4-1: A redundant, high availability, warm standby environment
CMs coordinate end-user connections that pass through the OpenSwitch server to the Adaptive Servers. If the RCM determines that the active Adaptive Server has failed, it connects to the Replication Server to fail over to the warm standby server, and coordinates the switch of end users through an OpenSwitch server. Decision-support-system users stay connected through the secondary OpenSwitch server.
Figure 4-2 illustrates the relationship between OpenSwitch and the RCM.
Figure 4-2: How OpenSwitch and the RCM work together
The RCM, in conjunction with Replication Server and OpenSwitch, is designed to help meet the requirements of a high availability, warm standby environment. It can:
Remove all single points of failure
Achieve an environment with fault-tolerant, redundant servers
Perform automatic failover of end users when the active data server fails
Coordinate access to the active and standby Adaptive Servers
Coordinate a geographically dispersed system, where the active and standby Adaptive Servers are at separate locations
