Some parameters of the Provisioning Manager might need to be adapted to your environment. These parameters are part of the Provisioning Manager parameters data set, prefix.PARM, in the members ENV for the Provisioning Manager runtime environment data and PARM for the Provisioning Manager configuration information.
The ENV member contains information about the runtime processing environment for your Provisioning Manager. Modify the following paths to match your installation settings:
This entry must contain:
This entry must contain the Capacity Provisioning JAR file cpom.jar from the installation directory /usr/lpp/cpo/classes, and the SAF JAR file /usr/include/java_classes/IRRRacf.jar.
Optionally, if your Capacity Provisioning
installation directory or your CIM installation directory are not
at the default locations, you might need to add the CIM Client for Java
/usr/lpp/wbem/jclient/sblim-cim-client2.jar
Add the CIM Client for Java Version 2 before adding the CIM Client for Java Version 1 if your Provisioning Manager is defined to be restarted on a system that has a lower release.
For example:
LIBPATH=/usr/lib:/usr/lpp/java/J6.0/bin:usr/lpp/java/J6.0/bin/classic:...
/usr/lpp/cpo/lib
CLASSPATH=/usr/lpp/cpo/classes/cpom.jar:...
/usr/include/java_classes/IRRRacf.jar
The PARM member contains configuration information for the Provisioning Manager. It has the structure of a Java property-file with keyword-value pairs. Both keywords and values are case-sensitive. Comment lines must start with a hash character (#). In the PARM member, you can enter keywords for:
To use ARM to monitor availability of the Provisioning Manager, you must set the value of ARM.Register to Yes. You can do this simply by removing the comment symbol from this statement in the sample member provided. If this key is not specified, or is given any value other than Yes, the Provisioning Manager is not registered with ARM. The value is not case-sensitive.
If you use ARM to monitor the Provisioning Manager, you must define ARM policy. This policy specifies an ARM element type and an ARM element name. If you chose the default element type and name, SYSCPM and SYSCPO, no changes are needed. If you changed these values, you must replace the values of the keys ARM.ElementType and ARM.ElementName with the ones you have chosen. The defaults are:
# ARM settings
ARM.Register = No
ARM.ElementType = SYSCPM
ARM.ElementName = SYSCPO
# Command authorization definitions
CIM.ReadGroup=CPOQUERY
CIM.ModifyGroup=CPOCTRL
SystemObservation.UserName=CPOCIM
If you
do not specify a userid for this key, the Provisioning Manager user
is used for system observation.
# Service data location
Trace.Path = /tmp
Log.Path = /tmp
Key | Default value | Description |
---|---|---|
Planner.BlockingTime | 15 | Number of minutes the Provisioning Manager waits after detection of a CPC capacity change before any new capacity change action is initiated. This time allows the workload to be redistributed after a capacity change. If faster provisioning or deprovisioning is warranted, the value can be lowered. The recommended range is a value of 5 - 15 minutes. |
Planner. ProvisioningRejectTime |
120 | Number of minutes a CPC is not considered for Provisioning Manager actions after a provisioning request proposed by the Provisioning Manager in confirmation mode has been rejected. |
Planner. DeprovisioningRejectTime |
120 | Number of minutes a CPC is not considered for Provisioning Manager actions after a deprovisioning request proposed by the Provisioning Manager has been rejected in confirmation mode. |
Planner. MinimumActivationTime |
240 | Number of minutes temporary capacity must remain active before it is considered for deactivation. This time interval starts again with any new activation and applies to all active resources. When specified, the minimum activation time must be longer than the blocking time. |
Analyzer.Threshold. TotalSharedPhysicalUtilCp | 95 | Lower limit of physical utilization on shared CPs that must be exceeded before additional general purpose capacity is considered. In some environments, severe processor contention occurs at lower levels of processor utilization. In such environments you can specify a lower percentage value, such as 90. |
Analyzer.Threshold. TotalSharedPhysicalUtilZaap | 95 | Lower limit of physical utilization on shared zAAPs that must be exceeded before additional zAAP capacity is considered. In some environments, severe processor contention occurs at lower levels of processor utilization. In such environments you can specify a lower percentage value, such as 90. |
Analyzer.Threshold. TotalSharedPhysicalUtilZiip | 95 | Lower limit of physical utilization on shared zIIPs that must be exceeded before additional zIIP capacity is considered. In some environments, severe processor contention occurs at lower levels of processor utilization. In such environments you can specify a lower percentage value, such as 90. |
Analyzer.Threshold. MvsUtilCp |
95 | Lower limit of MVS utilization on CPs that must be exceeded before additional general purpose capacity is considered. In some environments, severe processor contention occurs at lower levels of MVS utilization. In such environments you can specify a lower percentage value, such as 90. |
Analyzer.Threshold. MvsUtilZaap |
95 | Lower limit of MVS utilization on zAAPs that must be exceeded before additional zAAP capacity is considered. In some environments, severe processor contention occurs at lower levels of MVS utilization. In such environments you can specify a lower percentage value, such as 90. |
Analyzer.Threshold. MvsUtilZiip |
95 | Lower limit of MVS utilization on zIIPs that must be exceeded before additional zIIP capacity is considered. In some environments, severe processor contention occurs at lower levels of MVS utilization. In such environments you can specify a lower percentage value, such as 90. |
Analyzer.CycleTime | 200 | The Capacity Provisioning management cycle time in seconds. This time interval is used to analyze capacity metrics whether the capacity requirements have changed. |
SystemObservation. MovingAveragePiWeight |
100 | Percentage from 1 to 100 that expresses the
weight applied to the calculation of the moving average PI. The weight
applies both to calculation of system and sysplex PI calculation.
Smaller values increase smoothing by giving more weight to older PI
data, while larger values discount older data in benefit of the latest
PI data and produce a less smoothened moving average PI. Values below
100 enable the calculation of moving average PI for all observed workloads.
The average is also the basis for the management process.
The default value of 100 leaves moving average management disabled. |
SystemObservation. MovingAveragePiCapping |
5.5 | Maximum PI limit that is considered for calculating the moving average PI. Prevents extremely high PI values from distorting computed moving average PI in the long term. Zero indicates that capping is disabled. Other allowed values are 1.3 - 55. |
SystemObservation. PollingSpreadLimit |
0 | The default value of zero optimally times system
observation requests. The value can be changed if it is important
to distribute monitoring work more evenly over the observation cycle.
If multiple systems are observed, the value helps limit the overflow
of work eligible for specialty processor to general purpose processors.
Allowed values are 0 - 100. The default is zero (polling spread disabled). |
DefinedCapacity. BlockingTime |
15 | Number of minutes the Provisioning Manager waits after detection of a Defined Capacity or Group Capacity change for the system or capacity group before any new Defined Capacity or Group Capacity change action is initiated. This time allows the workload to be redistributed after a capacity change. The recommended range is a value of 5 - 15 minutes. |
DefinedCapacity. ProvisioningRejectTime |
120 | Number of minutes a system or capacity group is not considered for Provisioning Manager defined capacity actions after a provisioning request proposed by the Provisioning Manager has been rejected in confirmation mode. |
DefinedCapacity. DeprovisioningRejectTime |
120 | Number of minutes a system or capacity group is not considered for Provisioning Manager defined capacity actions after a deprovisioning request proposed by the Provisioning Manager has been rejected in confirmation mode. |
DefinedCapacity.LeadTime | 5 | Lead time in minutes before Defined or Group Capacity is exhausted. The projected time until capping reported by the monitoring product must fall below this value before the Provisioning Manager will increase Defined Capacity or Group Capacity on behalf of suffering workload. |
DefinedCapacity. MinimumActivationTime |
15 | Number of minutes a Defined Capacity or Group Capacity increase must remain active before it is considered for decrease by the Provisioning Manager. This time interval starts again with any new Defined Capacity or Group Capacity change. When specified, the minimum activation time must be longer than the blocking time. |
DefinedCapacity. GlobalMinimumDefinedCapacity | 5 | Minimum allowed Defined Capacity for the Provisioning Manager. The Provisioning Manager will not decrease any Defined Capacity below this value. |
DefinedCapacity. GlobalMinimumGroupCapacity |
5 | Minimum allowed Group Capacity for the Provisioning Manager. The Provisioning Manager will not decrease any Group Capacity below this value. |