Server failure may be due to:
CIMPLICITY software shutdown on either server
Network failure between the primary and secondary servers
Loss of server due to loss of power or equipment failure
Server failure is detected by the IPC Router, which is the communications process that runs on each server.
The Router sets up links to each server in the system and sends messages to each node at a set interval.
The probe interval is defined in REDUND_PROBE_DELAY that is set to 3000 millisecond by default.
If no reply is received from the server for a set number of tries (defined in REDUND_PROBE_COUNT) the server is then declared to have failed.
The Router sends a Partner Dead message to any processes that have outstanding messages to processes on that server.
Server failure may be detected on a primary server, secondary server or Viewer.
When a secondary server fails, functionality is not lost, because all functions are also running on the primary server.
When a primary server fails, the secondary server initiates procedures to take over redundant CIMPLICITY functions.
Primary server failure.