Service Group Workload Management

Service Group Workload Management is a load-balancing mechanism that determines which system hosts an application during startup, or after an application or server fault.

With Service Group Workload Management, you can statically associate system capacity and service group load within the main configuration file, main.cf. This is particularly useful when managing multiple service groups and systems. VCS also maintains the functionality of dynamic system load, which can be used when service groups deviate heavily from their static load. Dynamic load is specified at the system level, meaning you can specify the load regardless of which service group is loading the system. This helps account for system load contributed by an application outside VCS of control.

Deciding Startup and Failover Locations

Service Group Workload Management provides tools for making intelligent decisions about startup and failover locations, based on system capacity and resource availability. This feature is enabled when the service group attribute FailOverPolicy is set to Load. This attribute governs how VCS calculates the target system for failover. There are three possible values for FailOverPolicy: Priority, RoundRobin, and Load.

• Priority

The Priority failover policy is ideal for simple two-node clusters or small clusters with few service groups. With FailOverPolicy set to Priority, the system with the lowest priority is selected as the failover target. Priority is set in the SystemList attribute implicitly via ordering, such as SystemList = {SystemA, SystemB} or explicitly, such as SystemList = {SystemA=0, SystemB=1}. Priority is the default behavior.

• RoundRobin

The RoundRobin failover policy selects the system running the fewest service groups as the failover target. This is ideal for large clusters running many service groups with similar server load characteristics (for example, similar databases or applications).

• Load

The Load failover policy comprises the following components:

• System capacity and service group load, represented by the attributes Capacity and Load respectively.
• System limits and service group prerequisites, represented by the attributes Limits and Prerequisites, respectively.