In general, BCAM detects the loss of a BCAM connection. In particular situations, however, BCAM cannot determine the loss of connection itself (e.g. if the remote processor fails). For this reason, MSCF monitors the MSCF connections by periodically sending monitoring messages to the partner, which must be answered within the defined time span (FAIL-DETECTION-LIMIT / 11). If the response is not received or if BCAM reports a loss of connection, for MSCF there is an error in the MSCF connection and a connection recovery is initiated. In this case, the associated BCAM connections are first cleared down. An attempt is then made to reestablish these connections. During the recovery procedure, the command SHOW-MSCF-CONFIGURATION indicates the connection as CONNECTED, even though jobs to the partner can be rejected during this time.
If the connection can be reestablished (successful recovery), message MCS0018 is output on the console of both processors. Due to internal temporal dependencies during the recovery, it may happen that messages MCS0014 and (later) MCS0013 are output on one of the processors, while message MCS0018 is output on the other processor.
If the connection cannot be reestablished (connection failure), message MCS0014 is output on the console. The START-MSCF-CONNECTION command can be issued in an attempt to reestablish the connection manually.
A connection error impairs the partner monitoring mechanisms. If a connection error occurs in conjunction with a write error during the disk monitoring, this can lead to erroneous crash detection by the partner and, if operation continues, to inconsistencies in the data. This situation is dealt with in section “Input/output errors when accessing the watchdog file”.