Redundant heartbeat networks
Heartbeats in vSphere HA are sent between the master and slave hosts in order to determine a host’s failure. A host is deemed to have failed if these events occur:
- no network heartbeats are received.
- the host doesn’t respond to pings.
- no storage heartbeats are detected.
Heartbeat networks are implemented using a VMkernel Management port. It is recommended that you create redundant heartbeat networks on ESXi hosts in order to avoid a single point of failure. This way, if a master host’s connection fails, other connections can still be used to send heartbeats to other hosts.
There are two ways to create redundant heartbeat networks on ESXi hosts:
- NIC teaming – a team of two NICs is created.
- configure additional heartbeat networks – a second VMkernel port on a separate virtual switch with its own physical adapter is created.