Make server bulk RPC timeouts shorter than the client timeouts, so we don't

have cascading failures.  Server bulk timeout is 1/4 of the client timeout.
Also fix /proc variables to be int, as that is what the functions expect.
b=1845