[OpenAFS] read/write replication

[Craig_Everhart@transarc.com]

The easier thing to do is to put each office's read/write data in
separate volumes from the data from the other offices, to replicate
those volumes to all the other offices, then to periodically update
those replicas via various mechanisms.  Alternatively, you could simply
use R/W data housed in each office, then just use the cache managers
(the clients) in all the different offices to fetch and cache the
updated data from various offices.

		Craig