Problem allocating node for VNC in VMWare vCenter

I’ve detected a problem using VNC (noVNC) with VMWare vCenter.

I’ve a vCenter with 4 nodes: node-1, node-2, node-3 and node-4.

When I create a VM from Sunstone, the VM is asigned (for example) to node-1. It’s ok.

But if vMotion starts, that machine can get into node-4 (for example). If this occurrs, when I try to connect with VNC with the VM, the Sunstone tries to connect with node-1 via VNC, but the machine has been moved to node-4 (or other).

Does the ONE query in what ESX node is the machine before to launch the VNC?

Is any way to solve quickly?



Sunstone uses the ESX_HOST variable in the VM template to establish a VNC connection. This variable points to the ESX where the VM is running, and gets updated on each monitoring cycle (by default each 60 seconds), so after the migration it may be inconsistent.

Let me know if this is not the behaviour you are experiencing.