1. Microservices Architecture Documentation
  2. Prepare
  3. Prepare Databases
  4. Db2 z/OS
  5. Prepare Db2 z/OS Transaction Logs for Oracle GoldenGate

Prepare Db2 z/OS Transaction Logs for Oracle GoldenGate

Learn to configure the Db2 transaction logging to support data capture by Oracle GoldenGate Extract.

Oracle GoldenGate can capture Db2 transaction data from the active and archived logs. Follow these guidelines to configure the logs so that Extract can capture data.

To enable change capture for Oracle GoldenGate for Db2 z/OS, see Enable TRANDATA for Non-Oracle Databases

Enable Access to Log Records

Activate Db2 Monitor Trace Class 1 ("TRACE(MONITOR) CLASS(1) ") so that Db2 allows Extract to read the active log. The default destination of OPX is sufficient, because Oracle GoldenGate does not use a destination.

To Start the Trace Manually

  1. Log on to Db2 as a Db2 user with the TRACE privilege or at least SYSOPR authority.

  2. Issue the following command:

    start trace(monitor) class(1) scope(group)

To Start the Trace Automatically When Db2 is Started

Do either of the following:

  • Set MONITOR TRACE to "YES" on the DSNTIPN installation tracing panel.

  • Set 'DSN6SYSP MON=YES ' in the DSNTIJUZ installation job, as described in the Db2 UDB Installation Guide.

Note:

The primary authorization ID, or one of the secondary authorization IDs, of the ODBC plan executor must also have the MONITOR2 privilege.

Size and Retain Logs

More data gets logged when tables are defined with DATA CAPTURE CHANGES than when they are defined with DATA CAPTURE NONE. If any of the following is true, you might need to increase the number and size of the active and archived logs.

  • Your applications generate large amounts of Db2 data.

  • Your applications have infrequent commits.

  • You expect to stop the Extract for long periods.

  • Your network is unreliable or slow.

Use the DSN6LOGP MAXARCH system parameter in the DSNTIJUZ installation job to control log retention.

Retain enough log data so that Extract can start again from its checkpoints after you stop it or after an unplanned outage. Extract must have access to the log that contains the start of the oldest uncommitted unit of work and all logs after that.

If data that Extract needs during processing was not retained, either in online or archived logs, one of the following corrective actions might be required:

  • Alter Extract to capture from a later point in time for which log data is available (and accept possible data loss on the target).

  • Resynchronize the source and target tables, and then start the Oracle GoldenGate environment again.

Note:

The IBM documentation makes recommendations to improve the performance of log reads. In particular, you can use large log output buffers, large active logs, and make archives to disk.

Use Archive Logs on Tape

Oracle GoldenGate can read Db2 archive logs on tape, which will degrade performance. For example, Db2 reserves taped archives for a single recovery task. Therefore, Extract would not be able to read an archive tape that is being used to recover a table until the recovery completes. You could use DFHSM or an equivalent tool to move the archive logs seamlessly between online DASD storage and tape, but Extract will have to wait until the transfer completes. Delays in Extract processing increase the latency between source and target data.

Control Log Flushes

When reading the transaction log, Extract does not process a transaction until it captures the commit record. If the commit record is on a data block that is not full, it cannot be captured until more log activity is generated to complete the block. The API that Extract uses to read the logs only retrieves full physical data blocks.

A delay in receiving blocks that contain commits can cause latency between the source and target data. If the applications are not generating enough log records to fill a block, Extract generates log records by issuing SAVEPOINT and COMMIT statements until the block fills up one way or the other and is released.

In a data sharing group, each API call causes Db2 to flush the data blocks of all active members, eliminating the need for Extract to perform flushes.

To prevent Extract from performing flushes, use the Extract parameter TRANLOGOPTIONS with the NOFLUSH option.