The config below is basically what I use for all my ESX hosts. You can leave out the port speed/duplex settings if you are confident autonegotiation works reliably for you. Some network admins are wary of the 'spanning-tree portfast trunk' command too but it is almost a necessity to keep hosts in an ESX HA cluster from going haywire if you have a minor network hiccup.
You can go one step further if you have all your uplink ports plugged into the same physical switch and configure them to be in an etherchannel or LACP bond if you need more than a single uplink's worth of bandwidth available to a single VM (useful for file servers with lots of clients) but you need to be careful about the vswitch config on each host or you will end up with random connectivity issues (some clients can connect fine but others have issues).
<snip>
<snip>interface <int name="">
description <port description><port name="" here="">
switchport trunk encapsulation dot1q switchport trunk allowed vlan <list of permitted vlans><list of="" allowed="" vlans="">
switchport mode trunk switchport nonegotiate speed 1000 duplex full spanning-tree portfast trunk!<snip></list></port></
int></snip
>
