HiRDB Datareplicator Version 8 Description, User's Guide and Operator's Guide

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

9.7.3 Recovery procedure using the facility for recovering the extraction information queue file

The following figure shows the recovery procedure using the facility for recovering the extraction information queue file.

Figure 9-11 Recovery procedure using the facility for recovering the extraction information queue file

[Figure]

#1
If an initial start is executed, the facility for recovering the extraction information queue file can no longer be used because other files, including status files, are initialized. If you have executed an initial start, synchronize the source and target databases, and then execute an initial start on the source and target Datareplicators.

#2
Restore only the extraction server status file from its backup. Do not restore the data linkage file because this file is being used by HiRDB.

#3
If the source system is a HiRDB/Parallel Server, perform recovery processing for each back-end server. If multiple back-end servers are subject to recovery processing, repeat steps 3 through 5 for each back-end server that is subject to recovery processing.

The following table provides the details of the recovery procedure.

Table 9-15 Details of the recovery procedure using the facility for recovering the extraction information queue file

Item Target Description and example of command execution Check item
Step 1:
Obtain events.#1
Server subject to error recovery Determine the cause of the error. If the error occurred only in the extraction information queue file, use the procedure described here to recover the error. --
Step 2:
Terminate the source Datareplicator.
Manager node Stop the source Datareplicator.

Example of command execution:
hdestop
--
Step 3:
Make a backup of the extraction server status file.
Server subject to error recovery Make a backup of the extraction server status file.

Example of command execution:
See 6.4.2(7)(b) Backing up the extraction server status file.
--
Branch step inclusion if Step 5 fails:
If the initial recovery failed and a second recovery is attempted.
Server subject to error recovery Restore the extraction server status file from its backup.

Example of command execution:
See 6.4.2(7)(b) Backing up the extraction server status file.
--
Step 4:
Eliminate the cause of the error.
Server subject to error recovery Eliminate the cause of the error in the extraction information queue file. --
Step 5:
Perform recovery processing using the facility for recovering the extraction information queue file.
Manager node Recover the extraction information queue file.

Example of command execution:
hdestart -R -k queue -b bes1
Make sure that the range of files subject to recovery processing that is displayed in the KFRB05034-I message when recovery starts is the same in the KFRB05035-I message when recovery is completed.

Use the hdestate command to monitor the source Datareplicator until it terminates.#2 Alternatively, verify that the KFRB00510-I message indicating termination of the source Datareplicator has been output to the manager node's extraction master error information file. After the source Datareplicator has terminated, verify that no errors have been output to the error information file.
Step 6:
Restart the source Datareplicator.
Manager node Restart the source Datareplicator.

Example of command execution:
hdestart
--

Legend:
--: Not applicable

#1
If HiRDB's system log file has wrapped around and information needed for recovery has been overwritten, the facility for recovering the extraction information queue file cannot be used (in which case the KFRB05011-E message is output). In such a case, use the data linkage recovery facility.

#2
You cannot execute the hdestart or the hdestop command until the source Datareplicator has terminated.
Organization of this subsection
(1) Recovery processing using the facility for recovering the extraction information queue file
(2) Restarting the source Datareplicator
(3) If an error occurs during restoration of the extraction information queue file

(1) Recovery processing using the facility for recovering the extraction information queue file

The facility for recovering the extraction information queue file initializes the extraction information queue file, and then restores it.

(a) Initializing the extraction information queue file

To restore an extraction information queue file in which an error has occurred, you must first initialize the affected extraction information queue file.

The facility for recovering the extraction information queue file begins by initializing all extraction information queue files.

(b) Recovering the extraction information queue file

Restore the extraction information queue file if all extraction processing had been completed before the error occurred and there is update information that has not been sent to the target Datareplicator.

Restoration of the extraction information queue file involves extracting the update information that had been extracted before the error occurred from the system log file based on the extraction server status file, and then storing that update information in the extraction information queue file.

(2) Restarting the source Datareplicator

After you have restored the extraction information queue file, restart the source Datareplicator. When the source Datareplicator restarts, the update information re-extracted by the facility for recovering the extraction information queue file is sent to the target Datareplicator. Extraction processing is resumed for update information that was generated after the error occurred.

For an extraction information queue file whose update information had all been sent to the target Datareplicator before the error occurred, extraction processing is restarted from the initialized status.

(3) If an error occurs during restoration of the extraction information queue file

If an error occurs during restoration of the extraction information queue file, restore the extraction server status file from its backup, and then eliminate the cause of the error by referencing 9.1.2 Error handling methods. After that, re-execute the facility for recovering the extraction information queue file.

Note that if an error occurs in any of the files listed below, the facility for recovering the extraction information queue file can no longer be used. In such a case, use the data linkage recovery facility.