We recently swapped out our Cisco 6500 core switch, replacing it with a much cheaper pair of Dell PowerConnect 6248 stacks.  One of the first issues that came up after the swap over was that VMware High Availability stopped working.  The error message was that each VMware host could not communicate with its isolation address.

By default VMware hosts ping their default gateway to determine whether or not they are isolated from the cluster.  However, if you are running a pair of Dell PowerConnect switches using VRRP for redundancy, the VRRP IP address is not pingable (I’m not sure why Dell set this up – this isn’t the case with Cisco’s HSRP equivalent).  So the solution to the problem was to nominate another device in the same subnet as the Service Console as an isolation address and force VMware to use this.  We nominated our monitoring server, as we attempt to keep this running 24/7.  You need to go to the Advanced Settings for VMware HA, and set das.isolationaddress to the IP of the device you will be pinging, and then create a new variable das.usedefaultisolationaddress and set that to false.  After doing this, disable HA, then once that has completed, reenable.

HA should work fine after that.