SNMPD

[MRSG]

Hi,

I'd apreciate some urgent help.

The snmpd daemon on all our servers is taking up 100% CPU, we think it's becuase the TNG server we send messages to is unreachable at the moment so snmp mesages are not getting through.

Anyway, is it safe to stop and start the daemon during daytime (ie when the server is in use with db's running). I know it only taken 2 minutes with the following commands:

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

Any advice would be great, thanks.

[Michael Steele_2]

There's several subagents to snmp that could be using the snmp protocol, ems, nnm, for example.

Can you attach some more information first please?

sar -u 5 5 (* %idle near 0 *)
sar -v 5 5 (* numerator maxing out *)
sar -d 5 5 (* %busy > 50% *)
vmstat 5 5 (* po > 0 *)
swapinfo -tam (* tot % util > 85% *)

[eran maor]

Hi

the problem is in the side of the tng .

i solve it with upgrading the patch level of the snmpd .

the patch for 11.X is :

PHSS_27858 - s700_800 11.X OV EMANATE14.2 Agent Consolidated Patch

[Ravi_8]

Hi,

have you applied recent patch bundle (from support plus CD)