well, in this case trying a different hardware seems a logical step to me :) On 5/23/07, Diogo Biazus <diogob at gmail.com> wrote: > Yes, it's on the cluster that looses connection with the replicator. > -- Vlad