Clustering (not indexes)

Has anyone clustered a Postgres database server - e.g. made it run on two
seperate machines?  (The ideal solution would be a high-availability
cluster, where one machine could take the load of both of the other went
down.)

Any ideas?

Thanks,
Neil