Changing the default cluster tunables
You can use the SMIT interface to change the settings for dynamic LPARs and On/Off CoD.
To change the default cluster tunables, complete the following steps:
- From the command line, enter
smit sysmirror
- From the SMIT interface, select Cluster Applications and Resources > Resources > Configure User Applications (Scripts and Monitors) > Resource Optmized High Availability > Change/Show Default Cluster Tunables, and press Enter.
- Change the following fields:
- Always start resource groups
- If you select Yes, the resource might be started when the total requested
resource exceeds the maximum value for the LPAR profile or the combined available resources value.
Therefore, selecting Yes, ensures the best possible allocation of resources. The
default value for this field is Yes.
If you select No, the resource groups are not started.
- Resource allocation order
- Specifies the order in which resources are allocated to the node. You can select
the following options:
- Free pool before the Enterprise pool
- Select this option to first allocate resources from the free pool on the system. If
you choose this option, the resources are allocated in the following order:
- PowerHA® SystemMirror® allocates resources from the free pool on the system.
- If the free pool on the system has insufficient resources, PowerHA SystemMirror requests more resources from the Enterprise pool and the required resources are allocated.
- If the Enterprise pool has insufficient resources, PowerHA SystemMirror requests more resources from other systems in the same Enterprise pool first and then requests the Capacity on Demand (CoD) pool to allocate the remaining resources.
- Enterprise pool before free pool
- Select this option to allocate resources from the Enterprise pool on the system. If
you choose this option, the resources are allocated in the following order:
- PowerHA SystemMirror allocates resources from the Enterprise pool on the system.
- If the Enterprise pool on the system has insufficient resources, PowerHA SystemMirror considers resources from the free pool and allocates the required resources.
- If the Enterprise pool has insufficient resources, PowerHA SystemMirror requests more resources from other systems in the same Enterprise pool first and then requests the Capacity on Demand (CoD) pool to allocate the remaining resources.
- All Enterprise pool before free pool
- Select this option to allocate resources from the Enterprise pool on the system. If
you choose this option, the resources are allocated in the following order:
- PowerHA SystemMirror allocates resources from the Enterprise pool on the system.
- If the Enterprise pool has insufficient resources, PowerHA SystemMirror requests more resources from other systems in the same Enterprise pool first and then requests the Capacity on Demand (CoD) pool to allocate the remaining resources.
- If the Enterprise pool on the system has insufficient resources, PowerHA SystemMirror considers resources from the free pool and allocates the required resources.
Note: When more resources are requested from an Enterprise pool, the state of the Enterprise pool might change to Out of compliance. For more information about Enterprise pool states, see Power® Enterprise Pool compliance.The default option of this field is Free pool before the Enterprise pool. The resources are released in the reverse order in which they are allocated.
- Adjust Shared Processor Pool size if required
- Select Yes to authorize to dynamically change the maximum shared
processor pool value. If required, the allocation process increases the maximum limit of the shared
processor pool during the allocation process.
Select no to not have adjust the shared processor pool size.
- Force synchronous release of DLPAR resources
- Select yes to have release the CPU and memory resource
synchronously. By default, automatically detects if the resource is released by checking whether the
active node and backup node are on the same or different CECs.
Select no to not force synchronous release of DLPAR resources. Using asynchronous releases instead of synchronous releases does not cause a delay during the takeover process.
- I agree to use On/Off CoD and be billed for extra costs
- Select yes to have use the On/Off CoD function to obtain enough
resources to fulfill the optimal amount of resources that your application requires. You must enter
an activation code in the Hardware Management Console (HMC) to use the On/Off CoD function.
Select no if you do not want to use the On/Off CoD function.
- Number of activating days for On/Off CoD requests
- Specify the number of days that you want for activating On/Off CoD requests. If your environment has a multi-cluster deployment, you must use the same value for this field on all the different clusters. A multi-cluster deployment is an environment with multiple clusters deployed across two different systems.
- Verify and synchronize the cluster by using one of the
following methods.
- SMIT:
-
- From the command line enter
smit sysmirror
. - Select Cluster Applications and Resources- > Verify and Synchronize Cluster configuration and press Enter.
- From the command line enter
- clmgr command:
- To verify and synchronize the cluster, enter the following
command:
clmgr sync cluster