Nonstop Database, HiRDB Version 9 System Operation Guide
In HiRDB version 07-02 and earlier, if file creation, deletion, or extension processing is canceled, some area might fall into a status that belongs neither to unmanageable files nor to unreferenceable areas. This is the area of corrupted management information and corrupted file caused by operations such as a duplicate allocation of area; this occurrence has adverse effects on system operation. The following illustrates the concept of duplicate allocation:
Because the area management information is corrupted, any attempt to recover the corrupted target file results in corruption of other files. If you upgrade the system in this status, the corrupted area will be inherited.
Re-create all HiRDB file system areas. Then take the action described below for each HiRDB file system area purpose (-k option value in the pdfmkfs command).
Recover the database from its backup. This backup must contain the RDAREAs that were stored in the HiRDB file system area before the area management information was corrupted.
Create the system log files, synchronization point dump files, and status files. Note that the system cannot be restarted because the system files required for restart cannot be provided. Perform a normal start.
If the security audit facility is used, create the audit trail file and then restart the HiRDB and security audit facility.
No action is needed because work tables are created automatically by HiRDB during SQL execution.
Because the backup file for the pdcopy command and the unload data files for the pdrorg command have been lost, re-execute the pdcopy or pdrorg command, if necessary.
All Rights Reserved. Copyright (C) 2011, 2015, Hitachi, Ltd.