EMS configuration - system/events/status

[Maxim Yakimenko]

During last two days I'm trying to up EMS monitor sysstat_em. I want to check every 1-2 minutes presense among alives of some HPUX boxes. Can anyone show me example config, because I failed to find out what and how I should configure?

[RAC_1]

Do mean to say that monitor is down and you tried to bring it up? How did you try to bring it up? Best thing you can do is check it's status and restart it if required using /etc/opt/resmon/lbin/monconfig

[Maxim Yakimenko]

Sorry for inconvenience - let me provide some details. In man on this monitor written - first it trys to communicate with diagmond on specified machine, if it fails it trys to ping. So I want to check using this monitor if machine is alive and receive via email all notifications - is reachible or not, responfing or not and so on every specified time interval.
In
/var/stm/config/tools/monitor/sysstat_em.cfg I have pointed which machines I want to check. Also I found files "default_sysstat...clcfg" and "rst_sysstat...clcfg". Text in this files seems to be config of event qulification. But I can not understand how to specify time interval of checking.

[RAC_1]

I think you can test a monitor using send_test_event command. man send_test_event for details.

send_test_event -v sysstat_em should give you what you want.

[Maxim Yakimenko]

Well it seems something doesnot work. send_test_event outputs this:
bash-3.00# ./lbin/send_test_event -va sysstat_em

Finding resource name associated with monitor sysstat_em.

Found resource name /system/events/system_status
associated with monitor sysstat_em.

Creating test file /var/stm/config/tools/monitor/sysstat_em.test
for monitor sysstat_em.

Making test file /var/stm/config/tools/monitor/sysstat_em.test
for monitor sysstat_em
indicate send test event for all resources.

Performing resls on resource name /system/events/system_status
for monitor sysstat_em to cause generation of test event.
Contacting Registrar on nop

NAME: /system/events/system_status
DESCRIPTION: System Status Monitor

This resource monitors the state of a system on the network. The monitor
can be configured to report events directly using the Monitoring Request
Manager, or as a change in device status using the Event Monitoring
Service (EMS).

For more information, see the monitor man page, sysstat_em.1m

TYPE: /system/events/system_status is a Resource Class.

There are 2 resources configured below /system/events/system_status:
Resource Class
/system/events/system_status/ui_host
/system/events/system_status/default


No email notif receive I

[RAC_1 · Accepted solution]

Check /var/opt/resmon/log/event.log file details on your test event.

[Maxim Yakimenko]

I made check in monconfig and it said:

>/system/events/system_status ... OK.
For /system/events/system_status/ui_host/nop.itdep:
There are no monitoring requests.

If I do "(S)how monitoring requests configured via monconfig" I see:
1) Send events generated by all monitors
with severity >= MAJOR WARNING to SYSLOG
2) Send events generated by all monitors
with severity >= INFORMATION to TEXTLOG /var/opt/resmon/log/event.log
3) Send events generated by all monitors
with severity >= MAJOR WARNING to EMAIL root
4) Send events generated by monitors
/system/events/system_status
with severity >= INFORMATION to EMAIL emsmail

[Maxim Yakimenko]

In event.log I just see my yesterday attempts :)

[Maxim Yakimenko]

Done. I've edited default_sysstat_em.clcfg in /var/stm/config/tools/monitor dir and set there correct values for suppresion, time window and so on. My trouble arose because comments in this file, that should bring light on config of this monitor are poor and not clear. Thanks anyway for your participation.

[Andrew Merritt_2]

Hi Maxim,
Can you give details about the problems you had? What was unclear in the config file?

There is some more information about what needs to be set up here - http://www.docs.hp.com/en/diag/ems/emd_sys.htm

Does this make things any clearer?

Andrew

[Maxim Yakimenko]

Thanks. Now everything is fine.
Trouble was in comments in file file /var/stm/config/tools/monitor/default_sysstat_em.clcfg in part that concerns config data of event checking - event suppression time, time window, threshold and so on - I did not understand it from comments and get to understanding by way of trial.

[Andrew Merritt_2]

Hi Maxim,
You shouldn't need to change the event configurations for what you are doing, just add the details of the hosts that you want to monitor to the sysstat_em.cfg file (as you already have, I think).

You can change the polling frequency by modifying the value for POLL_INTERVAL in the sysstat_em.cfg file. (Change it to '2' to poll every 2 minutes.)

The documentation of the event thresholds etc. seems fairly comprehensive to me, but then I've been using it for some time. If you have any specific comments about what is confusing let me know. I do agree that the information about the 'HOST_ID' lines is not clear, but that just controls what information is displayed in the events about the host in question.

In general, you shouldn't need to change the event settings as they are chosen to give an EMS event when there really is a problem, and avoid giving false warnings. What changes did you need to make to get things working?

Andrew

[Maxim Yakimenko]

Well, for final config it seem reasonable to change only POLL_INTERVAL.