3750 SNMP stack monitoring

[W_E_S_T_Y]

I am building a few 3750 stacks, I want to be able to poll/monitor each individual switch in the stack but as it only has 1 ip associated with the stack how can I do this. I am polling the uplink interfaces but as I only have uplinks on the top and bottom switch I am blind to a switch going down in the middle of the stack.

[henniae]

Stacked 3750 switches (i.e. StackWise) for management purposes are a single device, thus you do not manage each individual switch as a separate entity. Each switch has a copy of the entire config for the stack and one is elected a master.

If you have the stack setup with all the switches connected in a ring connectivity will not be lost to an individual switch in the stack except the switch that goes down. When a switch goes down an SNMP trap will be sent by the stack to the configured management hosts(s). I think a syslog message is sent too. To determine if a single switch in the stack fails you need to watch for the trap or syslog message.

What do you mean by "I am polling the uplink interfaces"? Do you have some sort of layer 2 monitoring?

Alan

[W_E_S_T_Y]

These are client edge switches so I can't leave snmp linkup linkdown traps on as we don't want to know when the clients connect/disconnect. I am polling the gi1/0/1 and gi5/0/1 fibre uplinks to check they stay up.

I still have the problem that if I have a stack of say 5 switches I can't find a mib to poll the individual switches. If I power off a switch in the middle of the stack I do get a bunch of snmp alerts but they are not very meaningfull and I don't know which one (if any)is related to the switch going offline.

[SDaniels1]

I do this on our stacked 3750s by SNMP polling the 'operational status' for 'UP' on the 'StackPortX' interfaces.

For example, in a stack of two switches you will see two interfaces, if you do an SNMP query for them, names 'StackPort1' and 'StackPort2'.

The SNMP index for these is 5179 and 5182 respectively.

Use you SNMP monitoring tool to poll the switches IP checking the variable(s).

RFC1213-MIB|ifOperStatus.5179 = UP

RFC1213-MIB|ifOperStatus.5182 = UP

My largest stack is of four, the numbers increment.

Technically of course your actually checking the status of the stack module rather than the switch, but if it failed I would like to know!

hope thats of use.

Simon.

[David Stanford]

Have a look at this link for some good info:

Please have a look at the following URL:

http://www.cisco.com/en/US/products/hw/switches/ps5023/products_configuratio

n_guide_chapter09186a00801cdee1.html#1023898

Basically to access one particular member of the cluster, the community string to be used must be appended with @esN where N is the stack member number.

[wstockdale]

"snmp-server enable traps fru-ctrl"  will alert when an individual switch is removed or inserted into a stack.