WLMTK Overview: Using HP-UX WLM Effectively With Your Most Critical Applications

8
Workload: Instance1
Priority: 1
Goal: “Select” response time less than 10 seconds
Min CPU: 20%
Max CPU: 80%
Increase CPU resources when a particular user connects
The next example shows two SLOs for a workload containing one instance. These SLOs have different
priorities, allowing for a higher-priority, “must-meet” goal and an optional stretch goal. With these
SLOs, the Instance1 workload gets 10% of the CPU resources at priority 1. If all other priority 1 SLOs
are met, WLM then tries to satisfy the priority 2 SLO. If the user BATCHUSER is connected to the
instance, WLM increases the workload’s CPU resource allocation to 30%.
Workload: Instance1
Priority: 1
Min CPU: 10%
Max CPU: 10%
Workload: Instance1
Priority: 2
Min CPU: 30%
Max CPU: 30%
Condition: User BATCHUSER connects to the instance
Increase CPU resources when more than n users connect
The following example also has multiple SLOs for an instance’s workload. For the priority 1 SLO,
WLM attempts to grant 40% of the CPU resources to the workload, but only if more than 10 users are
connected to the instance. The second SLO, which is priority 2, grants 20% of the CPU resources to
the workload, but only if CPU resources are available after all priority 1 SLOs have their requirements
met. The second SLO ensures the workload gets some CPU resources when there are fewer than 10
users connected.
Workload: Instance1
Priority: 1
Min CPU: 40%
Max CPU: 40%
Condition: More than 10 users connected to the instance
Workload: Instance1
Priority: 2
Min CPU: 20%
Max CPU: 20%
Give an instance n CPU shares for each process in the instance
In this example, the instance’s workload gets an additional 3% of the CPU resources for each process,
up to a maximum of 80%. So, if the instance has 15 processes, the workload gets 45% of the CPU
resources. (Oracle background processes for an instance can affect the process count, depending on
how this number is measured.)
Workload: Instance2
Priority: 1
Min CPU: 5%
Max CPU: 80%
Goal: 3% of the CPU resources for each process in the instance