The following figure provides an overview of the recovery procedure using the system log file.

Figure 9-6 Overview of the recovery procedure using the system log file

The following table provides the details of the recovery procedure using the system log file.

Table 9-9 Details of the recovery procedure using the system log file

Item	Target	Description and example of command execution	Check item
Step 1: Obtain events	Source or target	Determine the cause of the error.	--
Step 2: Terminate extraction processing	Source	Stop the source Datareplicator. Example of command execution: hdestop	--
Step 3: Terminate import processing	Target	Terminate the target Datareplicator. Example of command execution: hdsstop -t immediate	--
Step 4: Transfer the recovery information file	Target	Transfer the recovery information file to the source in binary mode. If there are multiple recovery information files for the source system, transfer all of them to $HDEPATH in the corresponding back-end servers (including back-end servers where no error occurred). If a recovery information file is a character special file for UNIX, use the dd command to copy it as a regular file, and then transfer the resulting recovery information file. Example of command execution: Transferring a recovery information file: $ cd $HDSPATH $ ftp source-server-name ftp> cd pathname-of-corresponding-HDEPATH ftp> binary ftp> put recovery-information-file-name ftp> bye Copying a recovery information file: In AIX $ dd if=$HDSPATH/recovery-information-file-name \ $ of=any-path/recovery-information-file-name \ $ bs=1024 skip=1 count=recovery-information-file-size (KB) In UNIX $ dd if=$HDSPATH/recovery-information-file-name \ $ of=any-path/recovery-information-file-name \ $ bs=1024 count=recovery-information-file-size (KB)	--
Step 5: Eliminate the cause of the error	Source or target	Eliminate the cause of the error.	--
Step 6: Initialize for import	Target	Initialize the target Datareplicator. Example of command execution: hdsstart -i -f	--
Step 7: Initialize for extraction	Source	Initialize the source Datareplicator. Example of command execution: pdrplstop -f hdestart - i hdeprep -f extraction-definition-file-name pdrplstart	--
Step 8: Perform recovery processing	Source	Start the source Datareplicator during data linkage recovery. Example of command execution: hdestart - v	Check that the KFRB05041-I message that reports completion of recovery has been output.#

Legend:

--: Not applicable

Note 1:

If an error occurs during recovery processing, Datareplicator handles the error in the same manner as when an error occurs during normal data linkage. For details about the error handling procedure, see 9.1 Error handling procedures for the source Datareplicator and 9.2 Error handling procedures for the target Datareplicator.

Note 2:

If the source Datareplicator is restarted during recovery processing and a recovery start message had already been output before the previous session started, the message is not output again during the restart.

If no recovery information file is specified or the contents of the recovery information file differ from the last time Datareplicator started, Datareplicator uses the contents of the recovery information file that had been specified when the recovery start message was output.

Note 3:

If either of the errors listed below occurs on the recovery information file and Datareplicator is started with the hdestart - v command, the KFRB05007-E message is output and Datareplicator terminates with an error. In such a case, use the unload log files to recovery the error.

1. I/O error on the recovery information file
2. The target Datareplicator has not terminated normally, but the recovery information file was allocated to the source Datareplicator during recovery processing

If error 1 occurred and the error message is output, processing resumes. However, recovery information will not be updated thereafter.

As with error 2, if the target Datareplicator has not terminated normally (for a reason such as abnormal termination or a power-loss event), the recovery information might not have been updated correctly. If an attempt is made to recover an error by using such an incomplete recovery information file, the update information that had already been imported might be retransmitted and imported again during normal data linkage processing.

#

After outputting the message, the source Datareplicator continues normal data linkage processing and sends update information.