According to the current design if there is a network disconnection or hiccup between Primary and Secondary DMZ server, error msg slump terminated or slump died occurs and there is no reconnection or restart.
So, this communication depends on what is between the Primary and Secondary DMZ server : routers, switches, firewall, location and etc.
Bpvirtdb is the only process that tries to reconnect.
This DAR (Enhancement Request) requests that this design is changed in the future releases of Service Desk. So, all other Service Desk processes would try to reconnect as Bpvirtdb does.