Is it possible to stop/restart EMS on live platform?

[Vincent Abad]

Hi All,

We are monitoring LAN ressources through EMS,
but when we stop/restart the snmp master agent (/sbin/init.d/SnmpMaster start/stop) to take into account a modification into the snmp.cnf file, following error appears into the syslog
file for all monitored LAN every polling interval:

Aug 7 21:22:55 saylor EMS [883]: ------ EMS Event Notification ------ Value: "error" for Resource: "/net/interfaces/lan/status/lan900" (Threshold: != " 1") Execute the following command to obtain event details: /opt/resmon/bin/resdata -R 57868409 -r /net/interfaces/lan/status/lan900 -a

I think that the lanmond (mibmond), was not able to re-connect to snmp master agent.
The only way to correct this behavior is to reboot the host, as the /sbin/init/ems script
doesn't stopt/start.

Does someone has a idea on how respawning ems without rebooting the box.

THanks in advance,

Best Regards,


--Vincent

[Jeff_Traigle]

/sbin/init.d/emsa starts/stops the EMS agent. Not sure why HP left the /sbin/init.d/ems script there since it doesn't do anything.

[Geoff Wild]

Interesting - I'm going through something simliar - with OEM - Oracle Event Monitor.

In a Clustered system - our Oracle binaries are on cluster aware vg's - so in a failover - we have to stop/start OEM.

On our CSS level system - this caused issue with pkgmond:

Jul 10 19:38:08 svr03 pkgmond[5949]: SNMPGet of /cluster/package/package_status/prdbcipkg failed: Timed Out While Waiting for Response Pdu

One of the things OEM does is try to restart snmpdm...

Not too sure the relationship - but in: /sbin/init.d/ems

'stop')
print "This script does not stop the EMS Services"
if [ -d /var/stm/data/tools/monitor ]
then
print "Do not run this script manually from the command line prompt."
print "EMS monitors may not behave as expected."
print "If you have performed this action already, execute the command"
print "/usr/bin/rm -f /var/stm/data/tools/monitor/.system_shutdown_ems"
print "You can ignore this message in /etc/rc.log and /etc/rc.log.old"


So, that tells me that you shouldn't be messing with EMS "live".

Rgds...Geoff

[Sridhar Bhaskarla]

Hi Vencent,

If you would need to restart EMS, do

/etc/opt/resmon/lbin/monconfig

You can disable/enable monitoring there.

-Sri

[Vincent Abad]

Hi All,

It seems that we do not apply the good procedure to stop/start snmp master agent.
Using the following one, find in EMS user's guide solve our problem

# /sbin/init.d/SnmpMaster stop
# /usr/sbin/snmpd

If we use "SnmpMaster start", EMS is not able to reconnect well.
On the other way, we've got the error message only one time, and finally the process could again monitor ressource without error.

Thanks all for your help!

Best Regards,

--Vincent