Managing Thresholds in Hardware Sentry KM for PATROL

How to manage thresholds in Hardware Sentry KM for PATROL.

Related Topics

Predetermined Thresholds and Dynamic Thresholds

There are two types of thresholds in Hardware Sentry: predetermined thresholds and dynamic thresholds.

Predetermined thresholds are thresholds that are known ahead of time because they are always the same. They are identical from one system to another and can therefore be statically configured in the Hardware Sentry engine. Typical examples of parameters with predetermined thresholds are: Status (0 = OK ; 1 = Degraded ; 2 = Failed), PredictedFailure (0 = No failure predicted ; 1 = Predicted Failure) or ByteRate (no alert threshold).

Dynamic thresholds are thresholds that vary from one system to another and cannot therefore be established in advanced. They will be determined by Hardware Sentry at run-time, depending on the server model, the device type and the information available. Typical examples of parameters with dynamic thresholds are Temperature (alert thresholds for the ambient temperature are different from those for the internal temperature of a processor), Fan Speed and Voltage.

Thresholds management in Hardware Sentry

Predetermined and dynamic thresholds are applied to the PATROL Agent configuration during the discovery of Hardware Sentry. Predetermined thresholds are set using the ANYINST/ALL_INSTANCES method, dynamic thresholds are instance-specific and therefore set per instance.

The ANYINST/ALL_INSTANCES method is a way in the Event Management KM or in the “Parameter Overrides” settings in the PATROL Agent’s configuration to set alert thresholds on one parameter for all instances of a class. ANYINST refers to the Event Management KM /AS/EVENTSPRING/PARAM_SETTINGS/THRESHOLDS/<class>/ANYINST/<parameter> configuration variables.

The Event Management KM also allows filtering which instances the threshold configuration will apply to with a regular expression. By extension, ANYINST will refer to both ANYINST and regular expression-based threshold settings. ALL_INSTANCES refers to the “Parameter Overrides” /___tuning___/<class>/ALL_INSTANCES/<parameter> configuration variables.

Thresholds are applied only if there are no existing thresholds. Typically thresholds are saved in the PATROL Agent configuration during the very first discovery of the product and are never modified afterwards. Of course, if new devices or components are discovered or if remote servers are added to the monitoring process, additional thresholds will be configured.

Thresholds can therefore be modified at any time. Hardware Sentry KM will not override the customizations. One can think of the thresholds initially set by Hardware Sentry KM as default thresholds that can be tuned up later if required.

In earlier versions of the product, Hardware Sentry KM used to reset the alert thresholds upon each discovery and the administrator had to specifically flag the instances for which the thresholds should not be overridden upon each discovery.

Modifying Hardware Sentry Thresholds

If thresholds need to be modified, standard tools provided by BMC such as pconfig, PCM or the Event Management KM can be used. Hardware Sentry KM does not provide any specific way to customize thresholds.

Selecting the Appropriate Thresholds Mechanism

Before making any modifications to the Hardware Sentry KM thresholds, it is strongly advised to make sure that the KM uses the appropriate thresholds mechanism.

To do so, right-click any Hardware icon and select [KM Commands] -> [KM settings] -> [Additional Settings] -> [Thresholds Mechanism Selection].

Select the appropriate mechanism and click [OK].

Hardware Sentry automatically determines the default mechanism. If the Event Management KM is pre-loaded, the Event Management mechanism will be used. If not, the Tuning mechanism will be used. This test is performed only once during the initial discovery. The thresholds mechanism will not be changed by the product afterwards.

It is possible to switch from one mode to another by using the dialog box above. When switching mechanism, specific thresholds are migrated to the new mechanism. However, because of PATROL limitations, predetermined thresholds (ANYINST/ALL_INSTANCES) are migrated: the existing ones are deleted and Hardware Sentry default predetermined thresholds are set using the new mechanism.

Note that a backup of the existing thresholds is performed before the migration. The backup is saved in the “$PATROL_HOME\config” directory, the filename is: MS_HW_<hostname>_<port>_backup_<date and time>.cfg

Path to the Hardware Sentry Thresholds

The path to the thresholds in the PATROL Agent configuration depends on the currently active thresholds mechanism.

Please refer to KB1108 and KB1109 for additional information about the content of these variables and to learn how to set thresholds in the PATROL Agent configuration.

“Tuning” Thresholds Mechanism

Type Path Set
Predetermined /___tuning___/<class>/ALL_INSTANCES/<parameter> Per class
Dynamic /___tuning___/<class name>/<object ID>/<parameter name> Per parameter

“Event Management” Thresholds Mechanism

Type Path Set
Predetermined /AS/EVENTSPRING/PARAM_SETTINGS/THRESHOLDS/<class>/<connector name with no ‘.’ ><host>_{re:*}/<parameter> Per connector, host and class
Predetermined (connector) /AS/EVENTSPRING/PARAM_SETTINGS/THRESHOLDS/MS_HW_CONNECTOR/__ANYINST__/Status Per class
Dynamic /AS/EVENTSPRING/PARAM_SETTINGS/THRESHOLDS/<class name>/<object ID>/<parameter name> Per parameter

Reinitializing the Hardware Sentry Thresholds

To reapply default thresholds, Hardware Sentry KM needs to be reinitialized. To do so, right-click any Hardware Sentry host instance and select [KM Commands -> Reinitialize].

Make sure that the “Reset alert thresholds and polling intervals” option is checked and click the [Reinitialize] button.

All predetermined and dynamic thresholds will be removed from the PATROL Agent configuration and replaced by the Hardware Sentry KM default thresholds.

Parameters List

Parameters with Predetermined Thresholds (ANYINST/ALL_INSTANCES)

The following table contains the list of parameters for which ALL_INSTANCES or ANYINST thresholds will be applied as well as the thresholds value.

Class Parameter Alarm 1 Alarm 2
Min Max Type Min Max Type
MS_HW_BATTERY Status 1 1 WARN 2 2 ALARM
MS_HW_BATTERY Charge 0 30 ALARM 30 50 WARN
MS_HW_BLADE Status 1 1 WARN 2 2 ALARM
MS_HW_CAPACITYREPORT DegreesBelowWarning 0 0 VOID 0 0 VOID
MS_HW_CONNECTOR Status 1 1 WARN 2 2 ALARM
MS_HW_CPU CurrentSpeed 0 0 VOID 0 0 VOID
MS_HW_CPU Status 1 1 WARN 2 2 ALARM
MS_HW_CPU PredictedFailure 1 1 WARN 0 0 VOID
MS_HW_CPUCORE Status 1 1 WARN 2 2 ALARM
MS_HW_CPUCORE UsedTimePercent 0 0 VOID 0 0 VOID
MS_HW_DISKCONTROLLER BatteryStatus 1 1 WARN 2 2 ALARM
MS_HW_DISKCONTROLLER ControllerStatus 1 1 WARN 2 2 ALARM
MS_HW_DISKENCLOSURE IntrusionStatus 1 1 WARN 2 2 ALARM
MS_HW_DISKENCLOSURE Status 1 1 WARN 2 2 ALARM
MS_HW_ENCLOSURE PowerConsumption 0 0 VOID 0 0 VOID
MS_HW_ENCLOSURE Status 1 1 WARN 2 2 ALARM
MS_HW_FAN Status 1 1 WARN 2 2 ALARM
MS_HW_LOGICALDISK Status 1 1 WARN 2 2 ALARM
MS_HW_LOGICALDISK UnallocatedSpace 0 0 VOID 0 0 VOID
MS_HW_LUN Status 1 1 WARN 2 2 ALARM
MS_HW_MEMORY ErrorStatus 1 1 WARN 2 2 ALARM
MS_HW_MEMORY Status 1 1 WARN 2 2 ALARM
MS_HW_MEMORY PredictedFailure 1 1 WARN 0 0 VOID
MS_HW_NETWORK Status 1 1 WARN 2 2 ALARM
MS_HW_NETWORK ReceivedBytesRate 0 0 VOID 0 0 VOID
MS_HW_NETWORK TransmittedBytesRate 0 0 VOID 0 0 VOID
MS_HW_NETWORK ReceivedPacketsRate 0 0 VOID 0 0 VOID
MS_HW_NETWORK TransmittedPacketsRate 0 0 VOID 0 0 VOID
MS_HW_NETWORK ZeroBufferCreditPercent 0 0 VOID 0 0 VOID
MS_HW_NETWORK DuplexMode 0 0 VOID 0 0 VOID
MS_HW_NETWORK LinkSpeed 0 0 VOID 0 0 VOID
MS_HW_NETWORK LinkStatus 0 0 VOID 0 0 VOID
MS_HW_OTHERDEVICE Status 1 1 WARN 2 2 ALARM
MS_HW_PHYSICALDISK PredictedFailure 1 1 WARN 0 0 VOID
MS_HW_PHYSICALDISK Status 1 1 WARN 2 2 ALARM
MS_HW_POWERSUPPLY Status 1 1 WARN 2 2 ALARM
MS_HW_ROBOTICS Status 1 1 WARN 2 2 ALARM
MS_HW_ROBOTICS MoveCount 0 0 VOID 0 0 VOID
MS_HW_TAPEDRIVE Status 1 1 WARN 2 2 ALARM
MS_HW_TAPEDRIVE NeedCleaning 1 1 WARN 2 2 ALARM
MS_HW_TAPEDRIVE MountCount 0 0 VOID 0 0 VOID
MS_HW_TAPEDRIVE UnmountCount 0 0 VOID 0 0 VOID
MS_HW_TEMPERATURE Status 1 1 WARN 2 2 ALARM
MS_HW_VOLTAGE Status 1 1 WARN 2 2 ALARM

Exceptions

In some cases, the parameter activation depends on a value obtained at run-time by Hardware Sentry KM. In such cases, specific thresholds may be set in addition to the ANYINST/ALL_INSTANCES thresholds for the following parameters:

Class Parameter
MS_HW_BLADE Status
MS_HW_DISKCONTROLLER BatteryStatus
MS_HW_DISKCONTROLLER ControllerStatus
MS_HW_ENCLOSURE Status
MS_HW_FAN SpeedPercent
MS_HW_FAN Status
MS_HW_LOGICALDISK UnallocatedSpace
MS_HW_MEMORY Status
MS_HW_NETWORK BandwidthUtilization
MS_HW_NETWORK ErrorPercent
MS_HW_NETWORK Status
MS_HW_NETWORK TransmittedPacketsRate
MS_HW_NETWORK ReceivedBytesRate
MS_HW_NETWORK ReceivedPacketsRate
MS_HW_NETWORK TransmittedBytesRate
MS_HW_NETWORK TransmittedPacketsRate
MS_HW_NETWORK ZeroBufferCreditPercent
MS_HW_OTHERDEVICE Status
MS_HW_PHYSICALDISK PredictedFailure
MS_HW_TAPEDRIVE MountCount
MS_HW_TAPEDRIVE NeedsCleaning
MS_HW_TEMPERATURE Status
MS_HW_VOLTAGE Status

Parameters with Instance-Specific (Dynamic) Thresholds

The following parameters have instance specific thresholds. They are determined and set the first time the instances are discovered in accordance to the manufacturer’s recommendations.

Class Parameter
MS_HW_CPU CorrectedErrorCount
MS_HW_CPUCORE UsedTimePercent
MS_HW_ENCLOSURE IntrusionStatus
MS_HW_FAN Speed
MS_HW_FAN SpeedPercent
MS_HW_LED Status
MS_HW_LOGICALDISK ErrorCount
MS_HW_LUN AvailablePathCount
MS_HW_MEMORY ErrorCount
MS_HW_NETWORK ErrorPercent
MS_HW_NETWORK LinkStatus
MS_HW_NETWORK BandwithUtilization
MS_HW_OTHERDEVICE Value
MS_HW_PHYSICALDISK ErrorCount
MS_HW_POWERSUPPLY UsedCapacity
MS_HW_ROBOTICS ErrorCount
MS_HW_TAPEDRIVE ErrorCount
MS_HW_TEMPERATURE Temperature
MS_HW_VOLTAGE Voltage

Exceptions: parameters which thresholds cannot be modified from the configuration

Some parameters’ thresholds cannot be customized with external tools. If required, modifications to these thresholds need to be made using the Hardware Sentry KM interface.

These parameters are:

Class Parameter KM Command to modify thresholds
All Missing KM Settings > Missing Device Detection
MS_HW_ENCLOSURE IntrusionStatus KM Settings > Intrusion Detection Alerts
MS_HW_LUN AvailablePathCount None, cannot be customized.
MS_HW_NETWORK LinkStatus This System’s Settings > Network Link Alerts
MS_HW_NETWORK DuxplexMode This System’s Settings > Network Link Alerts
MS_HW_NETWORK LinkSpeed This System’s Settings > Network Link Alerts