Switching

Hi everyone,

We have a switch stack which failed to report the temperature to our Cacti. We tried to restart the snmp process by using the command "restart snmp gracefully all-members". But some fpcs failed to restart.

Output:

root@CORESW1> restart snmp gracefully all-members
fpc0:
--------------------------------------------------------------------------
error: Simple Network Management Protocol process is not running
error: Simple Network Management Protocol process was not restarted

fpc1:
--------------------------------------------------------------------------
error: Simple Network Management Protocol process is not running
error: Simple Network Management Protocol process was not restarted

fpc2:
--------------------------------------------------------------------------
Simple Network Management Protocol process started, pid 51520

fpc3:
--------------------------------------------------------------------------
Simple Network Management Protocol process started, pid 43356

fpc4:
--------------------------------------------------------------------------
error: Simple Network Management Protocol process is not running
error: Simple Network Management Protocol process was not restarted

fpc5:
--------------------------------------------------------------------------
error: Simple Network Management Protocol process is not running
error: Simple Network Management Protocol process was not restarted

fpc6:
--------------------------------------------------------------------------
error: Simple Network Management Protocol process is not running
error: Simple Network Management Protocol process was not restarted

fpc7:
--------------------------------------------------------------------------
error: Simple Network Management Protocol process is not running
error: Simple Network Management Protocol process was not restarted

We then tried to restart the snmp using the same command on each member of the stack that had an error but the snmp still failed to restart.

I got an idea from this link, does anyone know would killing the snmpd from the shell of the master work? If snmpd is killed from the master, would the effects cascade to its members? Or do we need to kill it for each member? That's the only thing that I found useful so far; I searched the error message in Google but found no relevant results.

The switch is EX4500 running JUNOS11.4R2

Hope someone can share their insights. Any help or input is greatly appreciated.

Thank You!

Killing the snmpd process on the master will not propagate to the other FPCs.

Can you remove the snmp configuration, commit and then add back the snmp configuration?(perhaps using 'commit full').

=====

If this worked for you please flag my post as an "Accepted Solution" so others can benefit. A kudo would be cool if you think I earned it.

Hello,

The snmp process is only running on the "routing-engine" and "backup" members, the linecards doesn't run the process. The routing-engine one runs the system for all the switches, and the backup is running a backup system, for the failover in case the routing-engine crashes.

So in a VC well configured, only two members have the processes running, so only those 2 restart. It's the regular behavior.

Have a nice day.