EMS, resource STATUS

[Mauro Garlaschelli]

Does EMS (default configuration) change the peripheral status every time a monitor signal a MINORWARNING, MAJORWARNING, a CRITICAL or I have to set PSM in order to do that ???

I'm asking it cause I want to see my peripheral status using "set_fixed -l" command.

Thanks.

[Andrew Merritt_2 · Accepted solution]

Hi Mauro,
The peripheral status change is determined by the .psmcfg file for each monitor, together with the default value. This is documented in the EMS Hardware Monitors Users Guide, http://www.docs.hp.com/hpux/onlinedocs/2512/B6191-90029.pdf

See page 121, headed "Peripheral Status Monitor (PSM) Configuration File".

The default value is SERIOUS and CRITICAL events will cause the state to change to DOWN.
This is over-ridden by the .psmcfg file for each monitor. For some monitors, the default is set (by use of the .psmcfg file) to be CRITICAL events only.

For example, for the disk monitor (disk_em), the entry in /var/stm/config/tools/monitor/disk_em.psmcfg
is:
MONITOR_RESOURCE_NAME: /storage/events/disks/default
DOWN_SEVERITY_THRESHOLD: CRITICAL
DOWN_SEVERITY_OPERATOR: =

So for disk_em, a CRITICAL severity event will cause the disk to be shown as DOWN.

I would not recommend changing the defaults without very good reason; an event for a device does not mean that it is now DOWN. If you just want to be notified of any events for a device, you can either watch the event.log file, or set up an email notification in monconfig.

Andrew

[Mauro Garlaschelli]

Thanks Andrew,

in the doc I saw that PSM is able to set a resource status to DOWN but sometimes it is not able to set it back to UP, correct ??

This happens when an event monitor cannot detect when a hardware failure has been repaired and the resource has been returned to normal operation. Consequently, these
monitor cannot alert the PSM to change the status of its hardware resources from DOWN to UP.

Can you tell me which monitor is able to detect when an hardware failure has been repaired ???

These are the monitor configured:

/StorageAreaNetwork/events/SAN_Monitor
/storage/events/disk_arrays/AutoRAID
/storage/events/disks/default
/adapters/events/FC_adapter
/adapters/events/TL_adapter
/system/events/chassis
/system/events/core_hw
/connectivity/events/hubs/FC_hub
/connectivity/events/multiplexors/FC_SCSI_mux
/connectivity/events/switches/FC_switch
/adapters/events/iscsi_adapter
/system/events/memory
/adapters/events/ql_adapter
/adapters/events/raid_adapter
/storage/events/enclosures/ses_enclosure
/storage/events/tapes/SCSI_tape
/system/events/ups
/storage/events/disk_arrays/FC60
/system/events/ipmi_fpl
/storage/events/disk_arrays/FW_SCSI
/system/events/ia64_corehw
/system/events/cpu/lpmc
/adapters/events/scsi123_em
/system/events/system_status

thanks in advance.

[Andrew Merritt_2]

Mauro,
Yes, not all the monitors can determine whether the hardware being monitored is no longer down.

The datasheet for each monitor lists the state control for each monitor.

See http://www.docs.hp.com/hpux/onlinedocs/diag/ems/emd_summ.htm for the links to each data sheet. Look for the header 'PSM State Control' on each one for the answer to your question.

Andrew

[Andrew Merritt_2]

Mauro,
Did you find the information you required in the replies above? If so, please allocate points to indicate how helpful the replies were, so others with the same question will be able to find the answer in future.

There is another, quicker way to check which monitors can set their state back to UP.

Look in the .psmcfg file for each monitor in /var/stm/config/tools/monitor for an uncommented line beginning with:
MONITOR_STATE_HANDLING:

if the value is either ALL_STATE_CONTROL or UP_STATE_CONTROL, then that monitor does not need the use of set_fixed to reset it to UP.

On my system, this applies only to dm_ses_enclosure, sysstat_em and dm_fc_scsi_mux.

Some of the .psmcfg files have a complete description of all the different values.

Andrew