VMware HA: monitoring VM

[Iwan_Rahabok]

Hi,

VMware HA now enables us to monitor individual VM (and not just the host).

The manual said that "Machine

Monitoring checks for a heartbeat every 20 seconds. If heartbeats have

not been received within a specified time (failure interval), Virtual

Machine Monitoring declares that virtual machine as failed."

What is the channel or mechanism that VMware Tools use to send a "heartbeat" to VMM? Is it via the Comm Channel (formely known as the back door)?

If yes, does it mean that VMware HA will not consider a VM as fail if it loses connection to the outside world, as the heart beat is not sent via TCP/IP?

Thanks

e1

PS: It seems to me, even in the physical world, a clustering software won't consider a node as fail if it loses connection to the production LAN.

[AndrewSt]

IIRC, this is heart beat from the VMTools. It is not dependent on network connections to the vCenter Server.

EX. A Win2k3 server has VMtools installed, and there is a 'heartbeat' sent. Everything is 'Good'.

Now, the server blue screens. Normal operation, shows that the system is still up - so ESX isn't going to reboot or fail the system. However, while in the blue screen state, the VMtools is offline, and the 'heartbeat' will not be heard. Now ESX will see the guest is down, and reset it.

-

---

-Andrew Stueve

[Iwan_Rahabok]

Thanks. Yes, it's VMware Tools that send the "heartbeat" to VMM.

I'm curious how exactly it does it.