The import transaction synchronization facility does not support creation of an import information editing UOC.

Partial initialization is not supported for just some of the data linkage identifiers that constitute a synchronous import group. If partial initialization is executed on a data linkage identifier constituting a synchronous import group, the KFRB04344-E message is output and initialization processing is cancelled. However, if either of the following conditions is satisfied, the entire synchronous import group is initialized:

• The synchronous import group consists of only the one data linkage identifier.
• Other than the data linkage identifier for which partial initialization is specified, no data linkage identifier has been defined or all the other data linkage identifiers are missing numbers.

In the example shown below, the KFRB04344-E message is not output. The situation is treated as initialization of the target Datareplicator.

Command specification

This example starts the target Datareplicator with initialization of dsid003 specified.

 
hdsstart -i -D 003

Import system definition

Other than dsid003, all data linkage identifiers are missing numbers:

 
dsid001 = **
dsid002 = **
dsid003 = c1
dsid004 = **
dsid005 = **
syncgroup001 = Grp001,c1

If the synchronous event for each back-end server is not synchronized during the import transaction synchronization facility processing, the transaction becomes large and a resource shortage might occur at the target HiRDB. If a large transaction cannot be completed, resulting in rollback, a large amount of rollback log information is output.

To avoid such a problem, specify the maximum number of transactions to wait for until synchronization. Make this specification in the syncwait_limit_tran_count operand of the import system definition in such a manner that the target HiRDB's lock pool size and system log file size are not exceeded.

When you execute PURGE TABLE, data linkage is supported if the source HiRDB is a parallel server and the table is not row-partitioned among multiple servers.

Notes:

If PURGE TABLE is executed, an import transaction is committed automatically. Therefore, the following actions must be taken while no update transaction has been executed on the source database:

1. Execute a synchronous event to achieve synchronization and obtain the status where no other transaction has been executed.
2. Execute PURGE TABLE. The PURGE TABLE execution conditions are as follows:
   There is only one PURGE TABLE in the transaction to be executed.
   There is neither INSERT, nor UPDATE, nor DELETE before PURGE TABLE.
   
3. Re-execute the synchronous event.

If any other operation is performed, only the import transaction with the data linkage identifier on which PURGE TABLE was executed is committed, and the import transaction's synchronization with other data linkage identifiers can no longer be guaranteed. If this occurs, you must execute a synchronous event immediately to synchronize the import transaction.

To change the extraction definition, you must initialize the target Datareplicator. For details about how to initialize the target Datareplicator, see 6.8.4 Changing the configuration when the import transaction synchronization facility is used.

If you change the extraction definition without initializing the target Datareplicator, the KFRB03317-E message is output during execution of the target Datareplicator, resulting in an error. To avoid this error, you must disable the synchronous import group, and then restart the system.

If you have specified USER in the HDSCLTWAITTIME environment variable for the target Datareplicator, specify a value of 0 for PDSWAITTIME and PDSWATCHTIME in the HiRDB client environment definition. If any other value is specified, the import SQL process might result in a connection timeout.