We have implemented 2x DSS Servers with replicated Fibre Channel volumes in a fully meshed fibre layout i.e. two fibre switches each dss server has a fibre card going to each switch and each vmware server has a fibre card going to each switch, this gives us redundent paths.
We then provide volumes to 4 x ESX4 vmware servers via Fibre channel. I have a couple of questions.
1) Should the path choice be set to round robin? I seem to recall somewhere that it mentions this.
2) If the primary DSS server crashes, when I change the source to the secondary server the volumes in ESX do not come back with the same name, which I thought it did with ESX3.5. Instead I have to add them back with new volume name, unassign the vm's running and add them back to the inventory from the renamed volume. This does make it a lenghty process.