Prevent blocking checkpoints on HDR servers

On an HDR secondary server, checkpoint processing must wait until the flushing of buffer pools is complete. You can configure non-blocking checkpoints on an HDR secondary server so that log data sent from the primary server is stored, or staged, in a directory until checkpoint processing is complete.

You configure non-blocking checkpoints on an HDR secondary server by setting the LOG_STAGING_DIR and LOG_INDEX_BUILDS configuration parameters. When non-blocking checkpoints is configured, log data sent from the primary server is staged in a directory specified by the LOG_STAGING_DIR configuration parameter. When the HDR secondary server finishes processing the checkpoint it reads and applies the log data stored in the staging area. When the staging directory is empty the HDR secondary server reads and applies log data as it is received from the primary server.

You enable non-blocking checkpoints by setting the LOG_STAGING_DIR configuration parameter on the HDR secondary server, and LOG_INDEX_BUILDS on both the primary server and the HDR secondary server. The value for LOG_INDEX_BUILDS must be the same on both the primary server and the HDR secondary server.

When the HDR secondary server encounters a checkpoint, it enters a buffering mode. While in buffering mode, the secondary server stages any log page data from the primary server into files in the staging directory.

When the HDR secondary server completes checkpoint processing, the server enters a drain mode. In this mode, the HDR secondary server reads data from the staging file and also receives new data from the primary server. After the staging area is empty, the HDR secondary server resumes normal operation.

Where log records are stored on the HDR server

The HDR secondary server creates additional directories named ifmxhdrstage_## in the directory specified by LOG_STAGING_DIR, where ## is the instance specified by SERVERNUM. The directories are used to store the logical logs sent from the primary server during checkpoint processing. The files within ifmxhdrstage_## are purged when no longer required.

Interaction of non-blocking checkpoints with secondary server updates

You must consider the interaction between secondary server updates and non-blocking checkpoints on HDR secondary servers. If the HDR secondary server receives an update request, the updates are not applied until the HDR secondary server processes the corresponding log records. When non-blocking checkpoints are enabled on the HDR secondary server, a delay in the application of data on the secondary server might occur because log data is staged at the secondary server due to checkpoint processing.