An unplanned switch is initiated by clustering when the operating system detects a condition that appears to be the loss of the *PRIMARY node. Possible causes include loss of power to the primary node, an abnormal ending of TCP/IP on the primary node, or the abnormal end of one or more IBM i clustering jobs.
When an unplanned switch occurs, one or more messages that require a response are sent to a predetermined message queue on the first *BACKUP node. The default message queue for this purpose is QSYSOPR. However, the application group may specify a different message queue or may specify *NONE when automation is used.
Do the following:
Message CPABB01 appears in the message queue on the first backup system and requires a response. Possible responses are.
C - (Cancel). The switch request is cancelled and the operating system will attempt to reestablish communications with the *PRIMARY node. If communications can not be established again, clustering will be suspended. IBM or Precisely may need to be consulted in order to restore full clustering capabilities. For this reason, exercise caution before canceling an unplanned switch.
G - (Go). Clustering will continue with the switch as if a planned switch had been requested.
Type the letter for the appropriate response and press Enter.