If access privileges for the HiRDB file system area for an RDAREA are incorrectly set, HiRDB detects a file access error for the RDAREA and causes an error shutdown for that RDAREA. If you use the facility for taking a unit down when a file access error is detected, you can take the unit down without shutting down the RDAREA when a file access error for the RDAREA is detected (in the case of a master directory RDAREA, however, the unit can go down without shutting down the RDAREA when a file access error is detected even if the facility for taking a unit down when a file access error is detected is not being used). This means that you can remove the cause of the file access error before restarting HiRDB, thereby avoiding the recovery task that would have been required had an error shutdown occurred in the RDAREA.

If a file access error is detected and the unit does go down while a UAP or utility is being executed in one of the operating modes described below, the processing target RDAREA might be shut down:

• Pre-update log acquisition mode or no-log mode
• A UAP or utility is executed on a user LOB RDAREA for which NO (no-log mode) is specified for the RECOVERY operand of CREATE TABLE

When using the facility for taking a unit down when a file access error is detected, avoid these operating modes if at all possible. If one of these operating modes is necessary, acquire a backup before executing a UAP or utility so that you can restore the system to the most recent status even if an error shutdown occurs for the RDAREA.

To use this facility, specify the system common definition (pdsys) as described below.

1. Specify unitdown for the pd_db_access_error_action operand.
   When this value is specified, the message KFPH23040-I is output when a file access error is detected, and the unit goes down. In this case, since no error shutdown of RDAREAs occurs, restarting HiRDB is all that is needed for recovery.
   
2. Specify MANUAL2 for the pd_mode_conf operand.
   Specifying this value prevents HiRDB from automatically starting after the unit goes down.
   If AUTO or MANUAL1 is specified, after the unit goes down following the detection of a file access error, HiRDB might restart before the cause of the error can be removed. In this case, the file access error is detected again, and the unit goes down and HiRDB restarts repeatedly.
   

If a unit goes down when the facility for taking a unit down when a file access error is detected is being used, the HiRDB administration must take action according to the procedure described in the following figure.

Figure 10-1 Action to be taken when a file access error occurs