There is no issue with database because all nodes would use the same database.
The issues are with the file system, if each node in the farm uses the same file system it should work fine but if they each have different copies there are many issues to solve and they will not be easy to solve. Especially difficult will be solving the search index which is file system based, other issues include user uploads and cache dependency files where we clear the cache by touching files.
I hope to complete web farm support solving these issues sometime in 2010 but it is not coming any time real soon as it is not my top priority.
I would go with as much scaling up with processors and memory on the web server as possible before trying to scale out to more servers. Typically first step is have the db on a separate server cluster as the db is usually a bottle neck long before the web code becomes a bottle neck, and you can scale out the db separately with no issues that I know of. All the challenges t be solved are for scaling out the web server.