Enabling Automatic Restart and Coordinator Failover

If the coordinator fails, local SCC IQ agents can automatically restart any server in the multiplex or start a new coordinator.

Prerequisites
Automatic restart and coordinator failover require a running SCC IQ Agent on the host where the multiplex server runs.
  • To enable automatic restart of multiplex servers, configure SCC agents for the servers.
  • To enable automatic coordinator failover, set the root logical server policy option ENABLE_AUTOMATIC_FAILOVER.
Task
Warning!   If you enable automatic coordinator failover, never restart the coordinator node or the designated failover node manually. Database corruption may result.

If manual intervention is required, shut down both the coordinator and the designated failover node. Start the coordinator. If unsuccessful, follow the steps in the related task Replacing the Coordinator (Manual Failover).

  1. Start an agent for each server. (If you installed SCC as a Windows service, the agents start automatically.)
    Option Description
    Unix or Linux $SYBASE/SCC-3_3/bin/scc.sh
    Windows %SYBASE%\SCC-3_3\scc.bat
  2. Run a CREATE AGENT statement for each multiplex server to associate a particular SCC IQ agent with each server in the multiplex.

    Coordinator failover now occurs automatically provided that the agents on the coordinator and designated failover node are running when the coordinator fails.

    SAP Sybase IQ cannot handle more than one point of failure. If the coordinator node fails at the same time as a secondary node, the secondary node does not restart automatically and you must start it manually.

Parent topic: Coordinator Failure
Related tasks
Replacing the Coordinator (Manual Failover)
Related reference
ALTER AGENT Statement
ALTER LS POLICY Statement
CREATE AGENT Statement
CREATE LS POLICY Statement
DROP AGENT Statement
SYSIQMPXSERVERAGENT System View
Server Startup Utility (start_iq)