We recently upgraded our Virtual Center appliance from 6.0U3 to 6.5.
We have not yet upgraded the host which are still 6.0U3 with current patches.
Immediately after vCenter was upgraded we started getting alarms for HostConnectionState. A few seconds later it would go back to green.
Of course, the hosts where not actually going down. Pings to and from host/vcenter never dropped. We followed the KB to change config.vpxd.heartbeat.notRespondingTimeout to 120 seconds which did quite down the alarms.
However, looking in the vcenter vpxd.logs we can still see its losing the heartbeat with these messages: Connection not alive due to missing heartbeats;
Steps we have tried:
1. Verified the vCenter Managed IP.
2. Verified the VPXD cfg file on the ESXi servers
3. Disabled and re-enabled HA on the clusters.
KB1005757 suggest the network is congested but that seems odd concerning it was not an issue prior to 6.5 upgrade. We don't plan on upgraded the host until our next maint window so we want to get to the root.