Monitoring hardware alerts with VMware ESX

With VMware ESX you can install the IBM Director according to the compatibility matrix (add link).The inband monitored events comprise everything that is surfaced by the service processor (either the Baseboard Management Controller through the openIPMI layer, or the Remote Supervisor Adapter through its device driver). These events can also be monitored out of band.

The service processor however is not able to monitor devices such as ethernet or fibre channel adapters. These objects are owned by the vmkernel, and IBM Director 5.20 (the most current version at the time of writing) only monitors the adapters seen by the Service Console.In a real life scenario, most of the times ethernet and fibre channel cards are redundant, i.e. teams of adapters are created so that if one card or path fails, the remaining adapter or path fails over without any service disruption. Even in such a high availability scenario, you want to receive a notification should a failover occur in order to take the appropriate actions.

IBM Director agent is not able to surface these events. What you can do though is to create a script that queries the vmkernel to see if any adapter is dead and creates an IBM Director event that is sent to the IBM Director Server. As the script polls the vmkernel, you will need to execute it with the frequency that you believe most appropriate to your environment, trading off the additional workload such a script is putting on your servers and the maximum time allowed between the failure (not failover!) and the notification of such a failure.

for Step by step guide click the following link below
http://www-941.haw.ibm.com/collaboration/ibmwiki/display/redwiki/Monitoring+hardware+alerts+with+VMware+ESX