Task 4: Recovering from damaged objects and unreadable sectors

If your system stops without warning or disk errors occur, some object description information might not be updated correctly. If this happens, the object is considered damaged.

When you perform an initial program load (IPL), the system attempts to locate damaged objects and write them to the object recovery list. It writes a message (CPI18xx) to the history (QHST) log for each damaged object that it finds. If any damaged objects are added to the object recovery list during the IPL, message CPI8197 is sent to the QSYSOPR message queue.

Note: Some damage, such as damage to the contents of a database file, might not be detected until the object is used. If you suspect that a large number of objects on your system have been damaged, contact your service representative for advice on how to recover.

To check and recover damaged objects, follow these steps:

  1. Display the QHST (history) log by typing DSPLOG and pressing F4 (Prompt).
  2. Display the QHST (history) log by typing DSPLOG and pressing F4 (Prompt).
  3. On the display, type *PRINT for the Output prompt and press the Enter key.
  4. Type: WRKSPLF.
    You are shown a list of spooled files for your job.
  5. Locate the spooled file for the DSPLOG command.
    Use option 3 to hold the spooled file.
  6. Use option 5 to display the spooled file.
  7. Look for entries for damaged objects that are not synchronized.
    You can use the Find function to search for lines that have these character strings: damage and sync. Following are some examples of messages you might see:
    CPF3113
    Member damaged
    CPF3175
    File is not synchronized
    CPF3176
    Data space is partially damaged
    CPF3171
    Journal is damaged
    CPF3172
    Objects are not synchronized with the journal
    CPF3173
    Journal receiver is damaged
    CPF3174
    Journal receiver is partially damaged
    CPF700C
    Object of type *object-type cannot be synchronized with journal.
    CPF81xx
    General messages about object damage
  8. Write down the names and types of the objects you find.
    Consult Table 1 for the correct recovery procedure, based on the type of object that is damaged.
    Table 1. Recovery for damaged objects by object type
    Type of object Recovery procedure
    Operating system object in QSYS library Contact software support for assistance. You might need to install the operating system again.
    IBM-supplied user profile Perform an abbreviated installation of the operating system.
    Job description that is specified on the workstation entry for the console in the controlling subsystem If no other workstation entries exist for the controlling subsystem, the system is not usable. Contact software support for assistance.
    Job queue Perform an IPL. Restore or re-create the damaged job queue. All entries are lost.
    Output queue Perform an IPL. If the output queue is the default output queue for a printer, it is re-created and its entries are rebuilt. Other output queues must be restored or re-created. Their entries are not recovered.
    Damaged file whose name starts with QAOSS Delete the file. Restore it from a backup copy. Run the RCLDLO DLO(*DOCDTL) command.
    Journal See Recovering a damaged journal.
    Journal receiver See Recovering a damaged journal receiver.
    Journaled object See Recovering a journaled object that is damaged or not synchronized.
    All others See Recovering other types of damaged objects.

  9. Watch for additional indications that objects have been damaged.
    Here are some indications:
    • You cannot start the system because auxiliary storage is full.
    • The system has ended abnormally several times since the last time you ran the Reclaim Storage (RCLSTG) procedure.
    • You see objects on the Work with Objects by Owner display that have no library associated with them.
    • The system status display shows an unexpectedly high percentage of auxiliary storage that is used.
    • You cannot access the data in a database file because a member is damaged. Message CPF8113 indicates this.
    • You cannot access objects because a damaged authorization list or authority holder secures them.

    If you see these indications on your system, the following actions can help you identify damaged objects:

    1. Choose from the following actions to identify damaged objects in critical files where you suspect damage:
      • Use the Copy File (CPYF) command
      • Use the Display Object Description (DSPOBJD) command
      • Perform a save of your critical data
    2. Choose from the following actions to identify damaged objects at the system level:
      • Use the Retrieve Disk Information (RTVDSKINF) command
      • Use the Display Object Description (DSPOBJD) command and specify DSPOBJD OBJ(*ALL) OBJTYPE(*ALL)
      • Perform a full system save using GO SAVE option 21
      • Run the RCLSTG procedure. Running the procedure is described in Reclaim Storage (RCLSTG) command topic.

    If you see these indications after a disk unit was replaced and the data was restored from a partial pump, you should recover the entire ASP that contained the failed disk unit. See the appropriate checklist.