The following table shows the actions to take when an error occurs while the automatic log unloading facility is being used.

Table 3-9 Handling of errors when the automatic log unloading facility is used

Error detail	Corrective action
The KFPS01157-W (reason=1662) message is output.	The directory path length specified in the pd_log_auto_unload_path operand, or the number of directories specified, is invalid. Correct the value specified for the pd_log_auto_unload_path operand, and then restart HiRDB.
The KFPS01155-W (reason=1667) message is output.	An error occurred on the disk containing the unload log file creation directory. Because the unload log files might become unavailable, stop operations and make backups. Make backups for the RDAREAs under all servers that are using the directory for unload log files in which the error occurred. Once the disk on which the error occurred has been restored, the automatic log unloading facility can be used again.
The KFPS01151-I (reason=1665) message is output.	The disk containing the unload log file creation directory is full. If there is only a single directory, stop the automatic log unloading facility. Transfer the unload log files in the directory to another medium, or delete unneeded files. Then use the pdlogatul command to start the automatic log unloading facility. If multiple directories for unload log files are being used, switch directories. If the disks for all directories become full, stop the automatic log unloading facility, transfer the unload log files in the directories to another medium, then use the pdlogatul command to start the automatic log unloading facility.
The KFPS01150-E message is output.	The automatic log unloading facility stopped. If Y is specified for the pd_log_auto_unload_restart operand, or if the specification is omitted Correct the error. HiRDB will automatically retry unloading when system log files are swapped. If N is specified for the pd_log_auto_unload_restart operand First correct the error, and then use the pdlogatul -b command to restart the automatic log unloading facility. It is possible that the error occurred because the unload log files cannot be used due to the fact that all directories for unload log files have reached one of the following statuses: Disk full# Disk error Wrong permission Nonexistent directory Unload log files created during the previous operation of HiRDB are still present. If none of the above is applicable, notify maintenance personnel after collecting the following information: Syslogfile All files under $PDDIR/spool pdlogls command's execution results Backup of the HiRDB system definition file
The KFPS01280-E message is output.	Automatic log unloading failed. Check whether the KFPS01271-I message was output immediately following the KFPS01280-E message. If the KFPS01271-I message was output No action is required because unload log files have been created normally. If the KFPS01271-I message was not output Verify the two points described below. No action is required if these conditions are satisfied. Unload log files of the applicable generation exist in the unload log file creation directory. The execution result of the pdlogucat command shows that the unload status of the applicable unload log files is completed. If the above conditions are not satisfied, some of the unload log files have disappeared and consequently, the database cannot be restored using unload log files. In this case, stop operation and acquire a backup of HiRDB.
The KFPS01158-W message is output.	Automatic log unloading failed to restart. Check the reason code of the KFPS01150-E message that is output to the syslogfile of the system manager unit to identify the cause of the error, and then take the necessary action. Furthermore, because the automatic log unloading facility has stopped, a system log file group is still waiting to be unloaded. Execute either of the following commands to unload it: pdlogunld command pdstart command (unloading occurs via the automatic log unloading facility the next time HiRDB starts)

#: If the unloading process fails because of a disk space shortage, HiRDB deletes an incomplete unload log file. Because this file is deleted after the disk became full during unload log file creation, the disk might now appear to have available space.