disable EMS alert

[CGEYROTH]

EMS is reporting that one of our dual power supplies in a SC10 enclosure has failed. Is there a way to tell it to ignore this particular error? we still want it to alert for other problems in the SC10 enclosure, just not for this particular power supply which we are aware of. Alternatively - it seems to spit out the error out every day at about 01:00 in the morning, is there a way to get EMS to spit out the error at a more reasonable time?

[Fabio Ettore]

Hi,

I think you can easly do that by file /var/stm/data/tools/monitor/disabled_instances. Insert the entry as described in the comments of the same file and restart the monitoring by /etc/opt/resmon/lbin/monconfig and click on E)nable Monitoring.
The most interesting part for you should be the following:

# For example:
# To not have monitor requests created for a disk at hardware
# path 52/8.5.0, currently monitored by the disk_em monitor the following
# entry would be added to this file:
# /storage/events/disks/default/52_8.5.0
/storage/events/tapes/SCSI_tape/4_0_1_0_0.9.20.255.0.0.0

Good luck...

Best regards,
Fabio

[Robert-Jan Goossens]

Hi,

Title: EMS: How to prevent (disable) EMS monitor requests for a device
Document ID: UEMSKBRC00011291
Last Modified Date: 8/4/05

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000079997128

Regards,
Robert-Jan

[CGEYROTH]

I saw that file but I'm not sure that it will work as we have defined montioring to be /storage/events/disk_arrays/FC60/. I can't disable this as it means no alerts to from the SC10 (which is attached to an FC60 controller), i need to disable a component of the sc10 (power supply) and not the entire f60/sc10. is it possible?

[Andrew Merritt_2 · Accepted solution]

Yes, using disabled_instances would hide all events for the device, and you don't want to do that.

You can edit the default_fc60mon.clcfg file in /var/stm/config/tools/monitor to suppress the particular event. You can either hide it completely, or increase the interval for which it is suppressed once it has been reported. The default interval for suppression is 24 hours, which is why you see the event reported at the same time every day. (The suppression means that once an event has been reported for a particular device, that same event won't be reported again for that device until the suppression time has expired (or the monitor is restarted).)

I think you're referring to event 21, so change from this:

#POWER_SUPPLY_FAILED
EQ:21:SERIOUS:TRUE:1440:ANY:1:NONE:NO_OP:NO_OP:NONE

to this:
#POWER_SUPPLY_FAILED
EQ:21:SERIOUS:FALSE:1440:ANY:1:NONE:NO_OP:NO_OP:NONE

to completely suppress event 21, or to this:

#POWER_SUPPLY_FAILED
EQ:21:SERIOUS:TRUE:2880:ANY:1:NONE:NO_OP:NO_OP:NONE

for example, to have 48 hours between events being reported (adjust as you prefer).

There is no way to suppress a particular event for a particular device path, so if you have more than one FC60 being monitored, I would suggest you increase the suppression time rather than disable the event, or you won't be informed of power failures in other units.

Andrew

[Yarek]

Hi,

Try to run /etc/opt/resmon/lbin/monconfig


rgds

[CGEYROTH]

Andrew was almost spot on, I think because we are running a older version of OS/EMS there were slight differences.

Basically there was no default_fc60mon.clcfg file, but there was a fc60mon.cfg file. the format was also different in that the poll interval is defined for the entire file rather than on individual components.

So what we did was set the poll interval to 6mins, then wait for it to alert and then reset it back to 1440 (24 hours). so hopefully even though it will keep alerting it will be 24hours from the last alert which means that it will be during our office hours and not in the middle of the night. we will find out 2nite.

[Andrew Merritt_2]

I'm glad you're making progress with this. 'An older version'? I think if there's no default_fc60mon.clcfg, we're talking practically prehistoric ;-)

If that is the case, I would recommend installing a current version ASAP, plus the current patch (and then my advice about modifying the .clcfg file would apply). What version do you have currently?

If it's 'POLL_INTERVAL' that you're changing in fc60mon.cfg, that may also give slightly different results from what you want; you might miss some events if the monitor is only polling the hardware once a day; the normal value is 15 (minutes). Changing 'REPEAT_FREQUENCY' in either fc60mon.cfg or Global.cfg might be closer to what you need, but you'd have to experiment as I've not played with a version as old as the one you appear to have.

Andrew

[CGEYROTH]

once again andrew is right, I meant to say that it was the "REPEAT_FREQUENCY" which we'd changed to 6mins to force an alert, then changed it back to 1440 (24 hours). we are running 11.00.

[Andrew Merritt_2]

I think you should have things working now, but I would still urge you to update to the latest version on 11.00, which is A.44.00 (HWE0403); you would then have a supported version installed. You can see the current version you have by running 'cstm'.

http://wtec.cup.hp.com/~hpux/diag/documentation/diags_ems_patches.html shows the supported versions. The current patch (likely to be the last one, too) for A.44.00 is PHSS_34834, which should be applied after the upgrade. With that version in place, you definitely would have a default_fc60mon.clcfg file.

(http://docs.hp.com/en/diag/stm/stm_ptch.htm shows the patches.)

Andrew