We currently have an ESX5.1 host which has 4 * NICs and is hosting 3 VMs including the company's main file server.
The 4 NICs form a vSphere "Standard Switch" and the "Load Balancing" on this switch is set to "Route based on the originating virtual port ID"
The physical switch is a Linksys SRW2048 which allows Link Aggregation, but currently this is not set up on the 4 ports connected to the VM server
Q1 - is there any benefit in setting up a LAG on the switch for these 4 ports? Am I likely to see any performance increase across the network of approx 100 users?
Q2 - is there any good reason NOT to set up the LAG?
Q3 - I've read the following http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1001938 - The way I read this is that I need to change the Load Balancing setting on the VM switch to "Route Based on IP Hash" first - at this point, I will lose network connectivity on the ESX server until I set up the LAG on the switch, at which point connectivity will be re-established. The question is, what if something doesn't work on the physical switch end - how do I then reverse the change on the VM switch, considering that I've lost network connectivity and therefore my vSphere client connection?
Thanks
Mike