The VIS6600 resets a single node, but parts of the ESX hosts can’t power on the virtual machines.
There has printing information “nmpDeviceAttemptFailover” in the ESX.
Restart the ESX physical host.
While the VIS is restarting, there has closed the virtual machine os in the host layer, but there hasn’t power off the ESX physical machine; this process takes about 10 minutes, there have ESX hosts entered in to the APD (All-Paths-Down) mode, the ESX hosts in this mode sometimes can’t be repaired without restarting.
While the VIS is upgrading and restarting, the virtual machine of the VMware needs to be powered off, meanwhile the ESX physical machine should be turned off too, or the ESX host will enter in the APD mode, and some ESX hosts can’t revive to normal after the VIS has restarted.