Scaling a service
Scale a service when you want to increase or decrease the number of service instances to adjust load during peak and low periods.
Before you begin
- You must be a cluster administrator or have the Services Manage permission to scale a service.
- The service must be in the DEFINED, STARTED, or ALLOCATING state.
About this task
When you scale a service, you change the maximum number of instances of the service that is defined in the service profile's MaxInstances parameter. To scale a service on each host, you must change the MaxInstancesPerHost parameter. (The MinInstances parameter specifies the minimum number of service instances that must run for the service to be considered started.)
If the allocation is higher than the new number of maximum instances, the service controller scales down the service. If the allocation is lower than the new number of maximum instances, the service controller attempts to start additional service instances.
For example, say you want 12 instances of a service to run on 4 hosts; you must then set MaxInstances to 12 and MaxInstancesPerHost to 3. With this configuration, the service controller starts 3 services on each available host, up to a total of 12. If only 3 hosts are available, only 9 instances are started. If 5 hosts are available, the service controller on one host allocates slots of all 5 hosts up to a maximum of 3 slots on any one host.
Procedure
While you can usually scale a service from the cluster management console, you might need to scale a service manually from the command line if you stopped the WEBGUI service or you cannot access the console.