Tuning the Sun Cluster HA for SAP DB Fault Monitors

Fault monitoring for the Sun Cluster HA for SAP DB data service is provided by the following fault monitors:

• The SAP DB fault monitor

• The SAP xserver fault monitor

Each fault monitor is contained in a resource whose resource type is shown in the following table.

Table 1–3 Resource Types for Sun Cluster HA for SAP DB Fault Monitors

Fault Monitor	Resource Type
SAP DB	SUNW.sapdb
SAP xserver	SUNW.sap_xserver

System properties and extension properties of these resources control the behavior of the fault monitors. The default values of these properties determine the preset behavior of the fault monitors. The preset behavior should be suitable for most Sun Cluster installations. Therefore, you should tune the Sun Cluster HA for SAP DB fault monitors only if you need to modify this preset behavior.

Tuning the Sun Cluster HA for SAP DB fault monitors involves the following tasks:

• Setting the interval between fault monitor probes

• Setting the timeout for fault monitor probes

• Specifying the response to a fault

Perform these tasks when you register and configure Sun Cluster HA for SAP DB. For more information, see Registering and Configuring Sun Cluster HA for SAP DB.

Operation of the Sun Cluster HA for SAP DB Fault Monitors

To determine whether SAP xserver and the SAP DB database instance are operating correctly, the Sun Cluster HA for SAP DB fault monitors probe these resources periodically.

During a probe, the SAP DB fault monitor performs the following operations:

1. The SAP DB fault monitor determines whether the SAP DB database instance is online.

2. If the SAP DB database instance is online, the SAP DB fault monitor determines whether the parent kernel process of the SAP DB database instance is running. You can control how the fault monitor responds if the parent kernel process is not running. For more information, see Forcing the SAP DB Database Instance to Be Restarted if the Parent Kernel Process Is Terminated.

3. The SAP DB fault monitor determines whether SAP xserver is available. This fault monitoring supplements the fault monitoring that the SAP xserver fault monitor provides.

During a probe, the SAP xserver fault monitor determines whether SAP xserver is available.

The SAP DB fault monitor responds to the following faults in SAP DB:

• A status of the SAP DB database instance that is not ONLINE, for example, OFFLINE or ADMIN

• Unexpected termination of the parent kernel process of the SAP DB database instance

The SAP DB fault monitor also responds to the unavailability of SAP xserver. This fault monitoring supplements the fault monitoring that the SAP xserver fault monitor provides.

The SAP xserver fault monitor responds to following faults:

• Unavailability of SAP xserver

• Persistent system errors

Setting the Interval Between Fault Monitor Probes

To determine whether SAP xserver and the SAP DB database instance are operating correctly, the Sun Cluster HA for SAP DB fault monitors probe these resources periodically. The interval between fault monitor probes affects the availability of each resource and the performance of your system as follows:

• The interval between fault monitor probes affects the length of time that is required to detect a fault and respond to the fault. Therefore, if you decrease the interval between fault monitor probes, the time that is required to detect a fault and respond to the fault is also decreased. This decrease enhances the availability of the resource.

• Each fault monitor probe consumes system resources such as processor cycles and memory. Therefore, if you decrease the interval between fault monitor probes, the performance of the system is degraded.

The optimum interval between fault monitor probes also depends on the time that is required to respond to a fault in the resource. This time depends on how the complexity of the resource affects the time that is required for operations such as restarting the resource.

For example, SAP xserver is a much simpler resource and can be restarted much quicker than SAP DB. Therefore, the optimum interval between fault monitor probes of SAP xserver is shorter than the optimum interval between probes of SAP DB.

To set the interval between fault monitor probes, set the Thorough_probe_interval system property to the interval in seconds that you require. Set this property for each resource that contains a Sun Cluster HA for SAP DB fault monitor that you need to tune. The resource types of these resources are shown in Table 1–3.

Setting the Timeout for Fault Monitor Probes

The timeout for fault monitor probes specifies the length of time that a fault monitor waits for a response from a resource to a probe. If the fault monitor does not receive a response within this timeout, the fault monitor treats the resource as faulty. The time that a resource requires to respond to a fault monitor probe depends on the operations that the fault monitor performs to probe the resource.

During a probe, the SAP DB fault monitor performs the following operations:

1. The SAP DB fault monitor determines whether the SAP DB database instance is online.

2. If the SAP DB database instance is online, the SAP DB fault monitor determines whether the parent kernel process of the SAP DB database instance is running.

3. The SAP DB fault monitor determines whether SAP xserver is available. This fault monitoring supplements the fault monitoring that the SAP xserver fault monitor provides.

During a probe, the SAP xserver fault monitor determines whether SAP xserver is available.

The time that is required for a resource to respond also depends on factors that are unrelated to the fault monitor or the application, for example:

• System configuration

• Cluster configuration

• System load

• Amount of network traffic

To set the timeout for fault monitor probes, set the Probe_timeout extension property to the timeout in seconds that you require. Set this property for each resource that contains a Sun Cluster HA for SAP DB fault monitor that you need to tune. The resource types of these resources are shown in Table 1–3.

Specifying the Response to a Fault

The SAP DB fault monitor responds to the following faults in SAP DB:

• A status of the SAP DB database instance that is not ONLINE, for example, OFFLINE or ADMIN

• Unexpected termination of the parent kernel process of the SAP DB database instance

The SAP DB fault monitor also responds to the unavailability of SAP xserver. This fault monitoring supplements the fault monitoring that the SAP xserver fault monitor provides.

The SAP xserver fault monitor responds to following faults:

• Unavailability of SAP xserver

• Persistent system errors

The Sun Cluster HA for SAP DB fault monitors enable you to specify the response to these faults. Specifying the response to a fault involves the following tasks:

• Defining the criteria for persistent faults

• Disabling the failover of the SAP DB resource

• Forcing the SAP DB database instance to be restarted if the parent kernel process is terminated

Defining the Criteria for Persistent Faults

To minimize the disruption that transient faults in a resource cause, a fault monitor restarts the resource in response to such faults. For persistent faults, more disruptive action than restarting the resource is required:

• For the SAP DB resource, the fault monitor fails over the resource to another node. The SAP DB resource is a failover resource.

• For the SAP xserver resource, the fault monitor takes the resource offline. The SAP xserver is a scalable resource.

The fault monitors treat a fault as persistent if the number of attempts to restart a resource exceeds a specified threshold within a specified retry interval. Defining the criteria for persistent faults enables you to set the threshold and the retry interval to accommodate the performance characteristics of your cluster and your availability requirements.

Dependencies of the Threshold and the Retry Interval on Other Properties

The maximum length of time that is required for a single restart of a faulty resource is the sum of the values of the following properties:

• Thorough_probe_interval system property

• Probe_timeout extension property

To ensure that you allow enough time for the threshold to be reached within the retry interval, use the following expression to calculate values for the retry interval and the threshold:

retry-interval ≥ threshold × (thorough-probe-interval + probe-timeout)

System Properties for Setting the Threshold and the Retry Interval

To set the threshold and the retry interval, set the following system properties:

• To set the threshold, set the Retry_count system property to the maximum allowed number of restarts.

• To set the retry interval, set the Retry_interval system property to the interval in seconds that you require.

Set these properties for each resource that contains a Sun Cluster HA for SAP DB fault monitor that you need to tune. The resource types of these resources are shown in Table 1–3.

Other Effects of the Retry Interval

Besides defining a criterion for persistent faults, the retry interval affects the response of a fault monitor to the following faults:

• Unavailability of SAP xserver that the SAP DB fault monitor detects. If the SAP DB fault monitor detects that SAP xserver is unavailable twice within the retry interval, the SAP DB fault monitor restarts SAP xserver.

• Persistent system errors. A persistent system error is a system error that occurs four times within the retry interval. If a persistent system error occurs, the fault monitor restarts SAP xserver.

Disabling the Failover of the SAP DB Resource

By default, failover of the SAP DB resource is enabled. The fault monitor fails over the SAP DB resource in response to a persistent fault. A persistent fault occurs if the number of attempts to restart exceeds Retry_count within the time that Retry_interval specifies.

Disable the failover of the SAP DB resource if you require the fault monitor to restart the SAP DB even in response to a persistent fault. When failover is disabled, the fault monitor reports a persistent fault in a resource by setting the status of the resource to faulted.

To disable the failover of the SAP DB resource, set the Failover_enabled extension property of the SUNW.sapdb resource to False.

Forcing the SAP DB Database Instance to Be Restarted if the Parent Kernel Process Is Terminated

By default, unexpected termination of the parent kernel process does not cause the SAP DB fault monitor to restart the SAP DB database instance. The SAP DB database instance can continue to function without the parent kernel process. Restarting the SAP DB database instance in this situation might cause unnecessary unavailability of the SAP DB database instance. Therefore, you should force the SAP DB database instance to be restarted only if you require a feature that the parent kernel process provides. An example of such a feature is maintaining the integrity of the log history.

To force the SAP DB database instance to be restarted if the parent kernel process is terminated, set the Restart_if_Parent_Terminated extension property of the SUNW.sapdb resource to True.