Nonstop Database, HiRDB Version 9 System Operation Guide

[Contents][Index][Back][Next]

24.11 Actions to take when an error occurs while the facility for automatic loading is being used

Organization of this section
(1) When an error occurs during automatic data loading
(2) If an error occurs in HiRDB when the facility for automatically loading data is being used

(1) When an error occurs during automatic data loading

The following table shows the behavior of HiRDB when an error occurs during automatic data loading, and the corrective actions that the HiRDB administrator or the auditor must take.

Table 24-38 Description of errors that might occur during automatic data loading and the corrective actions to take

Error description Cause of error HiRDB behavior Can the facility continue to be used? Corrective action by the HiRDB administrator or the auditor
No audit trail table exists. Audit trail file swapping occurred before an audit trail table was created. Outputs the message KFPL15029-E and stops data loading. N Administrator
Execute the pdmod command to create an audit trail table.
The maximum number of pdload commands that can be concurrently executed has been exceeded. If 0 is specified for pd_utl_exec_mode, the number of pdload commands being executed outside automatic data loading exceeded 32. Starts data loading after the pdload command that is executing has terminated. Y Administrator
Reassess the maximum number of pdload commands that can be concurrently executed, and change the value of pd_utl_exec_mode to 1 if necessary.
The maximum number of pdload commands that can be concurrently connected has been exceeded. The number of pdload commands concurrently connected has exceeded the value specified in pd_max_users. Outputs the message KFPA11932-E. Halts data loading but re-executes it after a while. Y Administrator
Reassess the maximum number of pdload commands that can be concurrently connected, and increase the value of pd_max_users if necessary.
A locking error occurred. A UAP exists that is referencing the audit trail table using a method other than a NOWAIT search. Outputs the message KFPA11770-I and stops data loading. Y Auditor
If referencing the audit trail table requires a long time, use the pdaudatld -t command to temporarily stop the facility for automatically loading data, or consider using referencing based on a NOWAIT search.
The RDAREA cannot be accessed. The RDAREA storing the audit trail table is in shutdown or closed status. Outputs the message KFPL23500-E and stops data loading. N Administrator
Open the RDAREA storing the audit trail table.
An audit trail file reading error occurred. The content of the audit trail file (HiRDB file) cannot be read. Outputs the message KFPS05704-E and stops processing by rolling back data loading. N Administrator
Based on the message content, remove the cause of the error.
An audit trail file writing error occurred.#3 An attempt to change the status of the audit trail file (to the data-loaded status) failed. Outputs the message KFPS05704-E and stops processing by rolling back data loading. N Auditor
Perform the following procedure:
1. See syslogfile and identify the audit trail file that terminated abnormally during automatic data loading.#1
2. Specify the file name identified in step 1 and manually execute the pdload command.#2
3. Execute the pdaudatld -b command to restart the facility for automatically loading data.
There is no audit trail file waiting for data loading. The auditor manually loaded data into the audit trail table. Outputs the message KFPL23201-W. Y No action is required.
An RDAREA capacity shortage has occurred. An RDAREA capacity shortage has occurred because the data was not saved or deleted from the audit trail table. Outputs the message KFPA11756-E and stops processing by rolling back data loading. N Auditor
If the audit trail table contains unnecessary data, delete it. Alternatively, ask the HiRDB administrator to expand the RDAREA.
Administrator
If requested by the auditor, expand the RDAREA.
The audit trail table cannot be updated. Data cannot be loaded into the audit trail table. Outputs the cause of the error to an SQL message and stops processing by rolling back data loading. N Administrator
Based on the content of the SQL message, remove the cause of the error.
A communication error (server startup error) occurred. This server cannot be started using the pdload command. The server or unit might be down. Outputs the message KFPL20000-E and stops data loading. N Administrator
Based on the content of the message and remove the cause of the error.
A communication error (data receiving error) occurred. Data cannot be sent from the unit in which the audit trail file is located to the unit in which the audit trail table is located. Outputs the message KFPL20000-E and aborts the operation, and as a result stops processing by rolling back data loading. N Administrator
Based on the content of the message, remove the cause of the error.
The process went down.#3 The pdload command for executing automatic data loading was canceled or the server went down. Outputs the message KFPS01820-E and aborts the operation. N Administrator
Do not execute the pdcancel command on the pdload process for automatic data loading.
If the above cause is not applicable, remove the cause of the error based on the message content. If the problem cannot be solved, contact the customer support center.

Legend:
Y: The facility for automatically loading data does not stop.
N: The facility for automatically loading data stops. After removing the cause of the error, you must use the pdaudatld -b command to restart the facility.
Administrator: Corrective action to be taken by the HiRDB administrator
Auditor: Corrective action to be taken by the auditor

#1
If you cannot see syslogfile because the unit is down, have a user who has the DB area privilege execute the database analysis utility (pddbst) on the audit trail table, and identify the audit trail file based on the execution result.

#2
The pdload command executed here does not load data, but merely changes the status of the audit trail file to the data-loaded status.

#3
This error occurs when the audit trail file could not be changed to the data-loaded status.

(2) If an error occurs in HiRDB when the facility for automatically loading data is being used

The following table shows the behavior of HiRDB if an error occurs in HiRDB while the facility for automatically loading data is being used, and the corrective actions to take. The HiRDB administrator must address the cause of the error by taking the corrective actions described.

Table 24-39 HiRDB errors that occur while the facility for automatically loading data is being used and the corrective actions to take

Error description Cause of error HiRDB behavior Can the facility continue to be used? Corrective action by the HiRDB administrator
When audit trail file swapping occurred, communication with the manager unit failed.
  • Cause 1: The manager unit is stopped.
  • Cause 2: A communication error occurred between the manager unit and the non-manager unit where the swapped audit trail file is located.

  • Displays the message KFPS05193-E.
    However, no message is displayed when the following communication errors occur:
    [Figure] An error in communication with the manager unit that occurs simultaneously with transmission of a message.
    [Figure] The message destination process has no room to accept a thread.
  • Automatic data loading is executed when audit trail file swapping occurs after the cause of the error has been eliminated.
Y
  • Cause 1: Use the pdstart -q command to restart the manager unit.
  • Cause 2: Investigate the cause and resolve the communication error between the manager unit and the non-manager unit.
Even though the audit trail file was swapped, it does not become a target for automatic data loading. A communication error occurred between the manager unit and the non-manager unit where the swapped audit trail file is located. Investigate the cause and resolve the communication error between the manager unit and the non-manager unit.
The pdload command failed to start during automatic data loading.
  • Cause 1: The number of processes within the system has reached the limit.
  • Cause 2: Because of a memory shortage, new processes cannot be generated.

  • Displays the message KFPS05717-W.
  • Automatic data loading is re-executed after a while.
Y
  • Cause 1: Re-assess the maximum process count on the server machine on which the manager unit is running.
  • Cause 2: Reduce memory usage on the server machine on which the manager unit is running.
Automatic data loading is not executed even though there is a file targeted for automatic data loading.
  • Cause 1: The unit containing the audit trail file targeted for automatic data loading is stopped.
  • Cause 2: A communication error occurred between the manager unit and the non-manager unit targeted for automatic data loading.
  • Cause 3: The facility for automatically loading data is stopped.
  • Cause 4: HiRDB stop processing is taking place.

  • Displays the message KFPS05718-W.
  • If an audit trail file targeted for automatic data loading is located in another unit, data is loaded from that audit trail file first.
  • If error causes 1 through 3 are eliminated, automatic data loading from the applicable audit trail file starts after the earlier automatic data loading is completed.
Y
  • Cause 1: Use the pdstart -q command to start the unit containing the audit trail file targeted for automatic data loading.
  • Cause 2: Investigate the cause and resolve the communication error between the manager unit and the non-manager unit.
  • Cause 3: If the auditor has not executed the pdaudatld -t command to explicitly stop the facility for automatically loading data, investigate the cause of the stop and take an appropriate action.
  • Cause 4: No action is required because the audit trail file will become a target for automatic data loading again the next time HiRDB starts.
Automatic data loading ended in an error because the unit went down. The non-manager unit that is the location of the audit trail file being automatically loaded or the audit trail table was forcibly terminated, abnormally terminated, or underwent system switching. Displays the message KFPS05713-E. N Use the pdstart -q command to restart the non-manager unit.
The manager unit was forcibly terminated, abnormally terminated, or underwent system switching.
  • Automatic data loading stops simultaneously when the manager unit stops.
  • Automatic data loading is executed after the manager unit restarts.
Y Use the pdstart -q command to restart the manager unit.

Legend:
Y: The facility for automatically loading data does not stop.
N: The facility for automatically loading data stops. After removing the cause of the error, you must use the pdaudatld -b command to restart the facility.
Manager unit: Unit in which the system manager is located.
Non-manager unit: Unit in which the system manager is not located.