EMS Hardware Monitors (logo)

CPU Monitor (formerly known as LPMC Monitor or lpmc_em)

Monitor Description | Configuration Files


Monitor Description

The CPU Monitor monitors cache parity errors in the processors of a system. For each occurrence of the error, CPU Monitor generates an event with details of the problem, its cause and the recommended solution.

The monitor also takes Dynamic Processor Resilience (DPR) action when a predetermined number of events have occurred on a processor in a 24-hour period. The monitor deactivates the processor for the current boot session and/or marks the processor for deconfiguration for the next boot session. If any spare processors such as Instant Capacity On Demand (iCOD) or Pay Per Use (PPU) are available, one of them is activated to replace the processor that the monitor deactivates.

At each POLL_INTERVAL, the CPU Monitor checks whether the problem is corrected and generates another event as a reminder if the problem is not resoved. In addition, the monitor runs Floating-Point tests on each processor on the system and takes DPR action for the processor on which the tests fail. The monitor launches automatically when the system starts.

For detailed information on this process, see Dynamic Processor Deallocation and Dynamic Processor Resilience.

Note: When using HP's Virtual Partitions (vPars), HP recommends that you configure EMS Hardware Monitors on all virtual partitions. If a system uses processor modules, which contain multiple processor cores, and the cores are split between multiple vPars, CPU Monitor may report an error affecting the entire processor to only one of the vPars.

Release History

June 2000 release (Support Plus Media) and HP-UX 11i: (Multiple-view ("Predictive-enabled") support)
June 1999 release: Independent Product Release (IPR)

Systems supporting the monitor

All systems running the following processors support the monitor:

Pre-requisites

Any HP-UX release starting from HP-UX 10.20.

Resource Path

Event monitoring:
/system/events/cpu/lpmc
Status monitoring:
/system/status/cpu/lpmc

Peripheral Status Monitor (PSM) State Control

PSM does not support automatic state control. Use the set_fixed utility to return a hardware resource to the "up" state following a failure.

Monitor Name

/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em


Configuration Files

Startup Configuration File (.sapcfg)

File name:
/var/stm/config/tools/monitor/lpmc_em.sapcfg

Default values are standard.

Monitor Configuration File (.cfg)

File name:
/var/stm/config/tools/monitor/lpmc_em.cfg

Default settings:

POLL_INTERVAL 15 minutes
FP_TEST_ITERATIONS Replaced by "Suppression Time" in the .clcfg file
Severity/Action See the .clcfg file for specific event information

PSM Configuration File (.psmcfg)

File name:
/var/stm/config/tools/monitor/lpmc_em.psmcfg

Default settings:

PSM Resource Name /system/cpu/lpmc
State Handling Requires use of set_fixed to set UP state.
Down State Handling SERIOUS will map to DOWN state.

Default Client Configuration File (.clcfg)

File name:
/var/stm/config/tools/monitor/default_lpmc_em.clcfg

Monitors that contain the client configuration files (.clcfg) are Multiple-View-enabled monitors. You can use the client configuration files to control the contents of messages to specific targets (for example, to Predictive Support). The default_lpmc_em.clcfg controls the messages sent to all targets. Specific .clcfg files (for example, predictive_lpmc_em.clcfg) control the text sent to specific targets (in this example, to Predictive Support).

HOST_ID tags in the .clcfg files control the system information, which is reported under the "Additional System Data" label in the event messages. If a HOST_ID tag is enabled in the *.clcfg file, then the corresponding information will appear in the event message.

HOST_IDs available for all monitors are described in the Generic Monitor (default settings for all monitors) data sheet.

DEV_ID tags in the .clcfg files control the monitor-specific information, which is reported under the "Component Data" label in the event messages. If a DEV_ID tag is enabled in the *.clcfg file, then the corresponding information will appear in the event message.

The following table lists the DEV_IDs available for this monitor:

DEV_ID tag Default Text label
dev_hpa on Hardware Physical Address (HPA)
dev_cpu_no on Processor Number
dev_pdev on Physical Device Path
dev_serial_num on Serial Number

Top of Page / Diagnostics HOME
URL: http://docs.hp.com/hpux/onlinedocs/diag/ems/emd_lpmc.htm
Last updated: Thursday Oct 25 10:08:22 PST 2007