Configuring the SFMIndicationProvider

This section describes how to configure the SFMIndicationProvider. It also describes how to switch between the SFMIndicationProvider and the EMS Hardware Monitors.

The SFMIndicationProvider operates independent of the EMS framework, and generates WBEM indications if it detects any abnormal activity on the resources it manages.

To switch from the cpe_em, cmc_em, and memory_ia64 monitors to the SFMIndicationProvider, enter the following command at the HP-UX prompt:

# /opt/sfm/bin/sfmconfig -w -s

The following output on the terminal that was used to start the CIMOM indicates that the SFMIndicationProvider has successfully replaced the three EMS Hardware Monitors:

Disabling EMS hardware monitors and enabling SysFaultMgmt. This may take a few minutes. SysFaultMgmt will now monitor the devices & EMS hardware monitors will be shutdown.

To check whether the SFMIndicationProvider is working properly, send a test event by entering the following command at the HP-UX prompt:

# /opt/sfm/bin/sfmconfig -t -m

The following output indicates that the SFMIndicationProvider is working properly:

Sending test event for memory monitor.

	NOTE: You can also send test events for other devices that the SFMIndicationProvider monitors.

For information on the devices monitored by the SFMIndicationProvider, see Table 1-2.

To view the list of events, enter the following command at the HP-UX prompt:

# evweb eventviewer -L

A list of events along with the details such as event number, severity, and event category are displayed.

To view the details of an event, enter the following command at the HP-UX prompt:

# evweb eventviewer -E -n <event number>

The details of the specified event such as the summary, description, probable cause, and recommended actions for the error are displayed.

To switch back from the SFMIndicationProvider to the three EMS Hardware Monitors, enter the following command at the HP-UX prompt:

# /opt/sfm/bin/sfmconfig -w -e

The following output on the terminal that was used to start the CIMOM indicates that the three EMS Hardware Monitors have successfully replaced the SFMIndicationProvider:

Enabling EMS hardware monitors and disabling SysFaultMgmt. This may take a few minutes. SysFaultMgmt will no longer monitor the devices & EMS hardware monitors will be enabled.

Creating SFMIndicationProvider Subscriptions

The SFMIndicationProvider uses the default Event Qualification (EQ) entries for generating events. To view the default EQ entries, enter the following command at the HP-UX prompt:

# evweb subscribe -T -g default -l

You cannot modify or delete the default EQ entries. However, you can create custom subscriptions to receive required indications.

Table 3-1 describes the EQ entries and their values.

Table 3-1 Event Qualification for the SFMIndicationProvider

Entry	Description	Available Values
event_number	A unique number assigned to each event of a monitor.	You can use any number. If no entry matches the event number, use OTHER.
device_category	The monitored resource.	On Itanium-based systems, the SFMIndicationProvider generates indications related to the following: Temperature problems Fan failures Power supply problems Power unit problems Corrected Machine Checks (CMCs) Corrected Platform Errors (CPEs) System Memory On HP 9000 systems, the SFMIndicationProvider generates indications related to the following: Temperature problems Fan failures Power supply problems Power unit problems
severity	The severity of the event.	SEVERITY_FATAL_NONRECOVERABLE SEVERITY_CRITICAL SEVERITY_MINOR SEVERITY_DEGRADED_WARNING SEVERITY_INFORMATION
enable flag	Switch to enable or disable an event.	TRUE - enables an event. FALSE - disables an event.
suppression time	Time to suppress the generation of an event, in minutes.	NOT_USED - never suppress an event. 1 - maxint - duration of suppression in minutes. 9999999 - Generates an event only once.
time window	Threshold value of events, in minutes. If an error is detected for a duration longer than that specified in this field, an event is generated.	NOT_USED - threshold value not used. ANY - time window threshold used, but no time window is specified. 1 - maxint -
threshold	Number of times an error must be detected in the time window to qualify as an event.	1 - maxint NOTE: To generate an event every time an error is detected, you must set the threshold value to 1, and the time window to “ANY”.
value threshold X, operator X	Value threshold is a case-sensitive condition. It can denote percentage in some cases. For example, percentage of the Page Dellocation Table (PDT) full. Operator denotes conditions to qualify value threshold. Value threshold and operator collectively represent an equation.	Value threshold categories: NONE - value threshold is not used. Any positive value. Operator values: NO_OP - operator is not used. >, <, >=, <=, ==, != For example, if the percentage of PDT required to qualify an event is between 60 and 70, enter the following: 60:<=:<=70

 

To create a new subscription with custom EQ criteria, make a copy of the default event configuration file and edit the contents of the file using a text editor. To enter a comment, begin the line with the # character.

Following are the conditions for creating custom subscriptions:

1. Every custom setting must begin with EQ.

2. Any two EQ entries must be separated by a colon.

3. Only the available values for EQ entries must be used.

4. All EQ entries for a custom setting must be in the same line. Carriage returns are not allowed.

5. EQ entries must be case sensitive.

Example:

In the following default configuration file for memory, event 3000 is suppressed:

# EV_SBE_DETECTED (Single bit error detected)EQ:memory:3000:SEVERITY_INFORMATION:FALSE:NOT_USED:ANY:1:NONE:NO_OP:NO_OP_NONE

To receive all indications related to single bit errors occurring on the DIMMs, make a copy of the default event configuration file and edit the appropriate line using a text editor, as follows:

# EV_SBE_DETECTED (Single bit error detected)EQ:memory:3000:SEVERITY_INFORMATION:TRUE:NOT_USED:ANY:1:NONE:NO_OP:NO_OP:NONE

To create a throttling configuration, enter the following command at the HP-UX prompt:

# evweb subscribe -T -c -g <throttling configuration name> -p <full path of the file containing the throttling configuration policy>

To associate the newly created throttling configuration with the subscription, enter the following command at the HP-UX prompt:

# evweb subscribe -C -n <subscription name> -e <severity> -v <device> -t <destination> -g <throttling configuration name>

To modify a throttling configuration with a subscription, enter the following command at the HP-UX prompt:

# evweb subscribe -M -n <subscription name> -g <throttling configuration name>

To create, modify, and delete custom subscriptions using the GUI, see “Administering Throttling Configuration Policy Using EVWEB”“Administering Throttling Configuration Policy” on page 110.

For information on the SFMIndicationProvider, see “Indication Providers”.