The extraction transaction information is used to uniquely identify a transaction that has occurred on the source HiRDB system. The extraction transmission information that is output when import processing is completed indicates that all information up to that transaction executed at the source system has been imported.

To apply data linkage recovery using unload log files, you must specify the transaction in an unload log file with which recovery is to be started. By specifying the extraction transaction information existing when import processing is completed, you can recover all the transactions that were completed after that transaction (the transactions that have not been imported into the target system).

(a) How to obtain the extraction transaction information when the target is Datareplicator

The following shows how to obtain the extraction transaction information when the target is Datareplicator:

1. After importing all the update information stored in the import information queue file, terminate the target Datareplicator.
2. Search the end of the error information file for the KFRB03009-I message and obtain the extraction transaction information displayed in Additional Transaction Info = in the message.
   If the error information file has been overwritten and there is no KFRB03009-I message, execute the command shown below to obtain the extraction transaction information from the execution results. Make sure that this command is executed before the target Datareplicator is initialized.
   

   hdsrefinfm -f target-status-file-name -l 9 -p analysis-results-output-target-file-name

Note: If a relative path is specified for the target status file name, Datareplicator assumes the path from $HDSPATH.

Example of command output and how to acquire the recovery start position

When the import mode is trn:

The recovery start position is Additional Transaction Info in <Reflect Entry Information[STAT]> HeadInformation:GroupNo=000.

When the import mode is tbl:

Import status	How to acquire the recovery start position
All Read Information is at the same location	The recovery start position is Additional Transaction Info in <Reflect Entry Information[STAT]> HeadInformation:GroupNo=001
Read Information is at different locations	The recovery start position is Additional Transaction Info for the import group whose Read Information is delayed most#

#

For other import groups, SKIP_TYPE_UNTIL is used as the import suppression control code. For details about the specification method, see 3.3.10 (1) Using import suppression to skip errors.

(b) How to obtain the extraction transaction information when the target is XDM/DS

The following table describes how to obtain the extraction transaction information when the target is XDM/DS:

Status of target XDM/DS	Acquisition method	Output target	Extraction transaction information
Normal or planned termination	See detailed information for the JWD396I console message that is displayed when the target XDM/DS is terminated	Console	kk...k part in EXTTRN INF: The kk...k that is displayed in JWD396I (24 hexadecimal characters).
	Acquire from the XDM/DS support utility's processing status output results	ABCPRINT	Information following LAST-EXT-TRN-INF= in the output results (24 hexadecimal characters)

This subsection describes the Perform recovery processing using the data linkage recovery facility step in the recovery procedure shown in Figure 9-8 Overview of the recovery procedure using unload log files.

(a) Setting up for data linkage recovery

Set up the data linkage recovery facility. When the data linkage recovery facility is set up, Datareplicator's normal extraction facility is replaced with the data linkage recovery facility. Therefore, note the following points:

• Set up the data linkage recovery facility after the source Datareplicator has terminated.
• Datareplicator's extraction facility is not available while the data linkage recovery facility is set up (Datareplicator cannot extract update information on processing executed on the HiRDB system while data linkage recovery is executing). While the data linkage recovery facility is executing, do not update a table subject to data extraction.

The data linkage recovery facility can process up to the point of extraction of update information. After execution of the data linkage recovery facility, you must re-transmit the update information to the target Datareplicator and execute import processing.

In the setup, execute the following two steps to extract the update information required for data linkage recovery into the extraction information queue file:

• Data linkage recovery facility 1 (transaction retrieval phase)
  Extracts transaction information subject to recovery from the unload log segment files created by the logmrg command and outputs it to the target transactions list file.
  
• Data linkage recovery facility 2 (extraction queue creation phase)
  Extracts update information for the transactions subject to recovery, which was extracted from the unload log segment files created by the logmrg command using the data linkage recovery facility 1, and stores it in the extraction information queue file.
  

If update information is extracted from multiple back-end servers, you must execute the recovery processing for each back-end server.