HOW IS A SWITCH 4500G CONFIGURED AS THE MANAGEMENT DEVICE FOR A CLUSTER?

Symptoms:

Facts:

Causes: Clustering uses two layer 2 protocols when identifying the candidate and potential members of a cluster. Those protocols are Network Discovery Protocol (NDP) and Network Topology Discovery Protocol (NTDP). Between the 2 protocols, candidates for membership in the cluster are identified using NDP. Once candidates are identified, the NTDP will build the cluster topology and identify the interfaces that the devices will communicate through.

If using the default configuration for the switch 4500G, NDP and NTDP are enabled on all interfaces as is the cluster?

feature.

However, interfaces that are not to be members of the cluster will require that NDP and NTDP be disabled (undo ndp enable and undo ntdp enable) on that specific interface.

Additionally, under the cluster configuration, the MAC address of the specific unit that is not to participate will have to be defined via the ?blacklist? option.

The MAC address of the blacklisted device can be attained via the ?display STP? command. It is the MAC address defined under the heading lt;st1:place w:st=“on”>CIST Bridge.

If there currently is a single vlan defined (default vlan1) then a second vlan will be required to be defined. It is this vlan that communication between the switch acting as the management device for the cluster and the members of the cluster are to communicate.

Thus each interface on the cluster management device that accesses the clustered device will be required to be configured as a trunk port and tagged with the vlan used for management.

Fixes: Steps to configure assuming that that the switch 4500G has default settings:

1. On all interfaces that there is to be a black-listed entry, disable NDP and NTDP

2. At the system level create a vlan that is to be used exclusively for the management cluster

3. At the system level define the management vlan as the vlan used to manage the cluster.

4. At the system level, type cluster? to enter the cluster? view.

5. Confirm that the expected candidates have been identified for the cluster using the command: Display Cluster Candidates.

6. Once candidates are identified under the Cluster view, execute the command “topology accept” of either “all” the candidates or a subset of those listed.


David Gonzalez 2021/04/14 14:54