(a) Overview
This facility outputs the information described below and checks the various items.
- Execution content common to deferred processes and HiRDB server processes
- The following information is output to a module trace:
- Semaphore values
- History of internal flags (management table status flags)
- Semaphore issuance history
- Execution content of deferred processes
- The following two types of information are checked:
- Values of issued semaphores
- Number of times the internal flags (management table status flags) were set
- If the contents of these items do not match, the message KFPH23203-E is issued.
- Execution content of HiRDB server processes
- The following two types of information are checked:
- Number of times the internal flags (management table status flags) were cleared when transactions were terminated or semaphores were issued
- Number of times semaphores were issued
- If these items do not match, HiRDB performs the following processes:
- Issues the KFPH23204-E message
- Forcibly terminates a deferred write process, acquires a core, and stops the HiRDB server process
(b) Preparation
Before you use this facility, expand the size of the module trace area by perform the following procedure described:
- Terminate HiRDB normally.
- Add the following operand to the system common definition $PDCONFPATH/pdsys:
set pd_module_trace_max = 16383
Specify a module trace entry count (48 bytes per entry).
- Start HiRDB normally.
Note: When the size of the module trace area is expanded, the memory size is increased for each HiRDB server process by the amount of the expansion (the default is 126 entries).
(c) Action to take when a message is issued
When a message is issued, take one of the following actions as appropriate.
- When the message KFPH23203-E is issued:
- Indicates that an internal conflict occurred in a deferred write process. Collect system information by following the instructions from the customer support center.
- When the message KFPH23204-E is issued:
- Indicates that an internal conflict occurred in a HiRDB server process. Collect the troubleshooting information that is output to the directory $PDDIR/spool.
- When the facility for parallel writes in deferred write processing is not used, the memory requirement increases by the amount indicated below for each deferred write process.
For a 32-bit system:
320 bytes x number of global buffers defined in server
For a 64-bit system:
640 bytes x number of global buffers defined in server
- When the facility for parallel writes in deferred write processing is used, the shared memory requirement increases by the amount indicated below for each server that uses the facility for parallel writes.
For a 32-bit system:
320 bytes x number of global buffers defined in server
For a 64-bit system:
640 bytes x number of global buffers defined in server
All Rights Reserved. Copyright (C) 2011, 2015, Hitachi, Ltd.