Guardian Procedure Calls Reference Manual

Considerations
Takeovers and selectable data segments in use
If the stack has never been checkpointed, then at a takeover, the selectable segment in use at
the time of the call to the CHECKMONITOR or CHECKSWITCH procedure is put into use. No
segment is put into use if the segment is not available; that is, the
SEGMENT_ALLOCATE_CHKPT_ or CHECKALLOCATESEGMENT procedure was not called
to allocate the segment to the backup process, or the SEGMENT_DEALLOCATE_CHKPT_ or
CHECKDEALLOCATESEGMENT procedure was called to deallocate the segment from the
backup process.
Messages
If CHECKMONITOR (or another checkpointing procedure) returns a value indicating that a takeover
has occurred due to a processor failure, the system subsequently delivers a system message -2
(processor down) to the $RECEIVE file of the new primary process. This might not be the first
message delivered to the new primary process if other system messages have arrived since the
last checkpoint operation of the old primary.
If CHECKMONITOR returns a value indicating that a takeover has occurred due to the primary
process stopping, the new primary process receives a system message -101 (Process Deletion) on
its $RECEIVE file.
For the format of system message -2 (processor down) or -101 (Process Deletion), see the Guardian
Procedure Errors and Messages Manual.
CHECKMONITOR Procedure 139