Introduction to checkpointing

The ObjectServer supports data persistence by using checkpoints and logs to copy the data in memory to disk. This enables you to recover the data after a planned or unexpected shutdown occurs.

The following types of files are maintained on disk:

• Checkpoint files, containing the data for entire tables
• Replay logs, containing the changes made to tables since the last checkpoint

Every 60 seconds, a checkpoint occurs, and all persistent data is copied to checkpoint files. Between checkpoints, new and changed data is written to replay log files.

Checkpoints also occur each time there is a planned ObjectServer shutdown.