We have a Cisco 2960S series switch which is a stacked switch with multiple physical switches managed as a single unit. A single management IP is assigned to the entire stacked switch.
Therefore, if one member of a switch stack fails, the IP address is still pingable. So we need to check whether any one of the individual switch in the stack has failed.
Can anyone highlight how we can check the individual switch failure using SNMP ?
You could try the following nagios check: http://exchange.nagios.org/directory/Plugins/Hardware/Network-Gear/Cisco/Check-cisco-3750-stack-status/details
As always, it is best to test that the check does what you want by causing a failure when possible. This check is based on the following oids (Taken from the script itself):
my $cisco_stack_table = "1.3.6.1.4.1.9.9.500.1.2.1.1.1";
my $cisco_stack_state = "1.3.6.1.4.1.9.9.500.1.2.1.1.6";
my $cisco_stack_ring = "1.3.6.1.4.1.9.9.500.1.1.3.0";
In particular, the description of cisco_stack_state (cswSwitchState according to the mib and of which is part of a table, so one for each switch) is:
The current state of a switch:
waiting - Waiting for a limited time on other switches in the stack to come online.
progressing - Master election or mismatch checks in progress.
added - The switch is added to the stack.
ready - The switch is operational.
sdmMismatch - The SDM template configured on the master is not supported by the new member.
verMismatch - The operating system version running on the master is different from the operating system version running on this member.
featureMismatch - Some of the features configured on the master are not supported on this member.
newMasterInit - Waiting for the new master to finish initialization after master switchover (Master Re-Init).
provisioned - The switch is not an active member of the stack.
invalid - The switch's state machine is in an invalid state.
removed - The switch is removed from the stack.
