Spill processing scenarios

Spill processing can be useful in a number of different replication scenarios. Here are some possible reasons you might use spill processing to maintain your replication environment.

Reorganizing data

If you need to make the target VSAM data set available for temporary processing that does not change the logical contents of the VSAM cluster, you can use spill processing to stop updating the target cluster. One example action is using IDCAMS REPRO to reorganize the VSAM cluster to reclaim space. Follow these steps:

  1. Use CECBTCCP to request SPILL DRAIN processing for the replication mapping.
  2. When the CECBTCCP command completes, take the target file for the replication mapping offline for your processing.

    Changes for the replication mapping are written to zFS while your processing is done.

  3. Make the VSAM cluster available again to the replication CICS region without changing the logical contents of the file. You can change the physical organization of the data.
  4. Use CECBTCCP to request ACTIVATE processing for the replication mapping to apply the spilled data and return the replication mapping to normal replication.

Resynchronizing data

If you need to repopulate the target VSAM data set by using a copy of the source VSAM cluster, you can follow these steps:

  1. Use CECBTCCP to request SPILL DISCARD processing for the replication mapping because you are preparing to replace the target data and do not need changes in the apply cache to be applied to the target data store.
  2. When the CECBTCCP command completes, take the target file for the replication mapping offline.
  3. Delete and define the target file.
  4. Take a copy of the source VSAM cluster and restore at the target.
  5. Activate the replication mapping using CECBTCCP to request ACTIVATE processing for the replication mapping with a timestamp that is prior to the copy. Data before the activation time is discarded when it is read from the spill queue; data after that time is applied and the replication mapping is eventually returned to normal replication.

Non-logged batch

Using non-logged batch processing is very similar to resynchronizing data, but this procedure might be needed nightly. Some applications are based on a sequential (QSAM) master file and use VSAM clusters during the day to make the data available to CICS TS processing. In these situations, it is not possible to log the data changes that are made in batch, so the target file must be resynchronized nightly after the source file is reloaded.

Take the following steps in these situations (some steps might overlap):

  1. Take the source file offline to reload with output from nightly batch processing.
  2. Use CECBTCCP to request SPILL DISCARD processing for the replication mapping because you are preparing to replace the target data and do not need changes in the apply cache to be applied to the target data store.
  3. Reload the source file and make available to the source application.
  4. When the CECBTCCP command completes, you can take the target file for the replication mapping offline.
  5. Delete and define the target file.
  6. Copy the output of the source batch processing to the target and restore into the target VSAM cluster.
  7. Activate the replication mapping by using CECBTCCP to request ACTIVATE processing for the replication mapping with a timestamp just prior to making the source VSAM cluster available to the application. Data before the activation time is discarded (if it was not previously flushed) when read from the spill queue; data after that time is applied and the replication mapping is eventually returned to normal replication.

Mirrored batch

You might decide to run the batch process on both the source and target LPARs. In that situation, you take the following steps (some steps might overlap):

  1. Take the source file offline for batch processing.
  2. Use CECBTCCP to request SPILL DRAIN processing for the replication mapping.
  3. Start the source batch processing.
  4. When the CECBTCCP command completes, take the target file for the replication mapping offline.
  5. Start the target batch processing.
  6. When the target batch processing is complete, make the target VSAM cluster available to the replication CICS region.
  7. Activate the replication mapping by using CECBTCCP to request ACTIVATE processing for the mapping with a timestamp just prior to making the source VSAM cluster available to the application. Data before the activate time is discarded (if it was not previously flushed) when read from the spill queue; data after that time is applied and the replication mapping is eventually returned to normal replication.

Temporarily entering spill processing to diagnose data inconsistencies

This scenario is similar to Resynchronizing data , but you do not change the target VSAM cluster in any way. You use spill processing to better understand the validity of the target VSAM cluster's contents before you decide either to reload the target VSAM cluster, park the replication mapping, or continue with some data inconsistencies (assumes adaptive apply processing for the replication mapping).

Temporarily parking a replication mapping because of data inconsistencies.

If you determine that a target VSAM cluster is no longer synchronized with the source VSAM cluster, you might choose to park the mapping to avoid adaptive apply messages and the overhead of processing data if you decide to resynchronize the data later. This approach allows you to decide that the replication mapping's data is not critical to the subscription and can be parked to stop replication.

To temporarily park a mapping, use CECBTCCP to request PARK WHILE ACTIVE processing for the mapping.