On 22 Nov 2004, at 19:52, J C Lawrence wrote: > Am I correct in assuming that Spread's membership model won't be happy > under fork()? Seems to be more stable to connect to the spread daemon in each child.