In creating virtual web servers (VirtualBox for me), does web content usually go "inside" the virtual machine or is there usually a file server running on bare metal serving files (content) to all the virtual machines?
Is there a typical scenario for storing content for virtual servers?
It depends on your architecture and how the information is going to be updated. If the information is going to be updated frequently and replicated across multiple machines, it may make sense to have a file server on bare metal that the VMs connect to over the network. However, you should be aware that with VirtualBox in particular you're going to take a pretty significant hit on network performance even with the guest tools installed. I've noticed I get about 20% of the network speed of VMWare from VirtualBox.
It depends is this development or production? If production with multiple servers store everything on a file share with fast network access so you don't have to deal with the hassle of updating every server. If it's development just put everything on the virtual server since performance isn't critical and it's simpler.
What are you creating the virtual servers for?
I usually use vms as testing machines, in which case the content will go "inside" (where apache is running).
If you are trying to simulate a networked system (like a load balancer or something), you will set up the content on the simulated machine that represents the file server in your scenario.