agree with a.p.
basically you create data luns or arrays, and present that to ESX. Then this raw storage is turned into a datastore - this is the storage that is used to hold virtual machines that you create. So when a new VM is created, you say how big the disk is going to be, and which datastore it will reside on. Then a file on that datastore is created that will be presented to the VM operating system as it's disk. You can shut down the vm and edit it's settings, and "grow" the disk to a larger size - then start the VM again and "expand" it's partitions to encompass the newly larger disk.
But in terms of having a common shared disk between vms, that is only possible with certain guest operating systems that offer clustering services - and the requirements for how to set it up are variable.
HTH