The table below shows the system log file's standby status when the HiRDB Datareplicator linkage facility is used. You can use HiRDB's pdlogls and pdls commands to check the status of the system log file.

Table 6-6 System log file standby status when the HiRDB Datareplicator linkage facility is used

Status	Remarks
Standby	Swappable target	Overwrite enabled	System log file can be swapped in if all three statuses are true.
		Unload completed
		Extracting status#
	Unswappable target	Overwrite disabled	System log file cannot be swapped in if any of the three statuses is true.
		Unload wait
		Extracting status#

#

This status is added only while HiRDB Datareplicator linkage is executing.

Note

If the source Datareplicator's extraction processing is stopped for an extended period of time after HiRDB Datareplicator linkage started, HiRDB's system log file might have become full and HiRDB might be shut down. We recommend that you do not stop the source Datareplicator's extraction processing while HiRDB is running.

There are two extraction statuses for the system log files:

• Extraction completed status
  In this status, extraction processing has been completed at the source Datareplicator. When the source Datareplicator finishes reading all system log information from a system log file, that system log file is placed in extraction completed status and becomes eligible to be swapped in if it is also in the overwrite enabled and unload completed statuses.
  
• Extracting status
  This is the status in which extraction processing has not yet been completed at the source Datareplicator. A system log file is placed in extracting status when the system log information it contains has not been extracted or is being extracted by the source Datareplicator. While it is in extracting status, the system log file cannot be swapped in even if it is also in overwrite enabled and unload completed status. If a system log file in extracting status is initialized or placed forcibly in swappable status, conformity between the source and target databases is lost, because some of the update information for the source database cannot be imported into the target database. If this happens, you must re-create the target database.
  

The pdlogls command might output the current file group as being in the unextracting status for HiRDB internal control purposes even when there is no unextracted update information in the system log. To determine whether the system log actually contains unextracted update information, execute pdls -d rpl -j.

All update information in the system log has been extracted when the values of Run ID, Gen No, and Block No displayed as System Log Extract Point are all greater than the values of the same items that are displayed as System Log Sync info.

When HiRDB Datareplicator linkage starts, the extraction status for a system log file is specified by the following methods:

• pd_rpl_init_start=Y in the source HiRDB's system common definition
• pdrplstart command

The source HiRDB checks the extraction status of each system log file when a synchronization point dump is collected. When the source Datareplicator completes extraction from a system log file, the file's status changes from extracting to extraction completed at the time of the next synchronization point dump. This means that a system log file is not placed in extraction completed status until a synchronization point dump occurs.

A system log file's extraction status is released when it becomes eligible to be swapped in. Until then, the system log file is displayed as being in the extraction status when the pdlogls command executed. A system log file in this status is still subject to extraction processing by the source Datareplicator.

The extraction status is released forcibly in the following cases:

• Execution of the pdrplstop command.
• pd_log_rpl_no_standby_file_opr=continue is specified in the system common definition, but the system log file cannot be swapped in because it is not in extraction completed status.
• The pd_rpl_init_start operand is modified in the source HiRDB's system common definition (changing from Y to N or omitting the pd_rpl_init_start operand).
• Execution of the hdestart command with the -i option specified.
• System log file in extracting status is initialized or its extraction status is changed forcibly.
• Data linkage file is initialized.

When the extraction status is released, conformity between the source and target databases might be compromised, because the HiRDB Datareplicator linkage preexisting at the source Datareplicator cannot be restored the next time HiRDB Datareplicator linkage is started. In such a case, you must re-create the target database.

If none of the system log files can be swapped in, you can specify an action to be taken when any of the system log files is in extracting status. This specification is made in the source HiRDB's system common definition, as shown below:

pd_log_rpl_no_standby_file_opr=stop|continue

For details about the specification of pd_log_rpl_no_standby_file_opr in the system common definition, see 5.6 Source HiRDB definition. For details about the handling procedure when a system log file cannot be swapped in, see 6.5.6 Source HiRDB handling procedure.

Terminating the corresponding source HiRDB unit forcibly in order to continue data linkage (when stop is specified or pd_log_rpl_no_standby_file_opr is omitted)

To continue data linkage, specify forced termination of the HiRDB unit in the system common definition. The corresponding source HiRDB unit will be terminated forcibly when there is no file that can be swapped in due to the extracting status.

To restart a forcibly terminated source HiRDB unit, check that a log file that is being extracted has been transformed to another file by the source Datareplicator (that is, check that there is at least one file in extraction completed status), and then restart the source HiRDB unit. You can determine the extraction status of a system log file by first executing the pdlogls command to obtain the status of each system log file, and then executing the pdls -d rpl -j command to obtain the progress of extraction processing.

For details about how to cancel data linkage and continue processing only at the source HiRDB when the source HiRDB unit is terminated forcibly, see (6) below.

Cancelling data linkage and continuing processing only at the source HiRDB (when continue is specified)

To cancel data linkage and continue processing only at the source HiRDB, specify cancellation of HiRDB Datareplicator linkage in the system common definition. The source HiRDB cancels HiRDB Datareplicator linkage at all back-end servers when there are no more files that can be swapped in due to the extracting status, and then continues its processing. The source HiRDB releases the extraction status of all system log files when there are no more files that can be swapped in. As a result, the source HiRDB can resume processing because a system log file previously in extracting status can now be swapped in.

When the extraction status is released, you must re-create the target database before restarting HiRDB Datareplicator linkage because conformity between the source and target databases is lost.

If the source HiRDB unit is terminated forcibly in accordance with a specification in the system common definition because the system log file is in extracting status, you must take one of the following actions to continue processing at the source HiRDB:

• To continue data linkage
  Start the source Datareplicator if it has terminated. If both the extraction and transmission processes are active, check the target Datareplicator's status at the destination and start it if it has terminated.
  
• To cancel data linkage and continue processing only at the source HiRDB
  If the system log file cannot be placed in extraction completed status due to an error in the source or target Datareplicator, cancel HiRDB Datareplicator linkage and restart the source HiRDB in order to continue processing. In this case, you must take one of the following actions before restarting the source HiRDB:
  
  • Change to continue the specification of the pd_log_rpl_no_standby_file_opr operand in the system common definition.
  • Use the pdlogchg command to forcibly place all system log files in extraction completed status.

When you cancel HiRDB Datareplicator linkage, conformity between the source and target databases is lost, in which case you must re-create the target database before restarting HiRDB Datareplicator linkage. For details about how to cancel data linkage when the source HiRDB unit is terminated forcibly, see 6.5.6 Source HiRDB handling procedure.

The following topics concerning manipulation of system log files are discussed here, because they are especially important with respect to data linkage:

• Checking a system log file's extraction status (pdlogls command)
• Changing a system log file's extraction status forcibly (pdlogchg command)
• Deleting a system log file in extracting status (pdlogrm command)

For details about other manipulations, see the HiRDB Version 9 System Operation Guide.

Checking a system log file's extraction status (pdlogls command)

You can use the pdlogls HiRDB command to check a system log file's extraction status. If you are using the HiRDB Datareplicator linkage facility, the following statuses are added to the unload status of file groups, element files, and physical files:

a: Unload wait and extracting status

A system log file in unload wait status contains log information that needs to be unloaded. The system log file is in extracting status when extraction of the data linkage information by the HiRDB Datareplicator has not been completed.

u: Unload wait and extraction completed status

e: Unload completed and extracting status

-: Unload completed and extraction completed status

A system log file's extraction status does not change when the corresponding unit becomes inactive at the source HiRDB. Even if extraction has been completed by the source Datareplicator, the pdlogls command displays the corresponding system log file as being in extracting status. Therefore, to check a system log file's extraction status while the corresponding unit is inactive, you must use the pdls HiRDB command to check the extraction status of the source Datareplicator.

Changing a system log file's extraction status forcibly (pdlogchg command)

If a system log file cannot be placed in extraction completed status due to an error at the source or target Datareplicator, you can use the pdlogchg HiRDB command to change the system log file's extraction status forcibly in order to continue processing at the source HiRDB. If you specify the -R option in the pdlogchg command, the system log file changes from extracting status to extraction completed status.

When you change the system log file's extraction status forcibly, conformity between the source and target databases is lost. In such a case, cancel HiRDB Datareplicator linkage, re-create the target database, and then restart HiRDB Datareplicator linkage. For details about the handling procedure during execution of the pdlogchg command, see 6.5.6 Source HiRDB handling procedure. For details about how to specify the -R option in the pdlogchg command, see 7. Command Syntax.

Deleting a system log file in extracting status (pdlogrm command)

To forcibly delete a system log file in extracting status, execute the pdlogrm HiRDB command with the -u option specified. You cannot delete a system log file that is in extracting status without specifying the -u option. When you delete a system log file in extracting status, conformity between the source and target databases is lost, in which case you must re-create the target database before starting HiRDB Datareplicator linkage.