Monitor checkpoint information

Monitor checkpoint activity to view information that includes the number of times that threads were required to wait for the checkpoint to complete. This information is useful for determining if the checkpoint interval is appropriate.

To monitor checkpoints, use the following commands.

Utility Command Additional Information
The onstat utility onstat -m View the last 20 lines in the message log.

If a checkpoint message is not in the last 20 lines, read the message log directly with a text editor. The database server writes individual checkpoint messages to the log when the checkpoint ends.

If a checkpoint occurs, but the database server has no pages to write to disk, the database server does not write any messages to the message log.
The onstat utility onstat -p Obtains these checkpoint statistics:
  • numckpts: Number of checkpoints that occurred since the database server was brought online.
  • ckptwaits: Number of times that a user thread waits for a checkpoint to finish. The database server prevents a user thread from entering a critical section during a checkpoint.
ON-Monitor (UNIX™) Status > Profile The Checkpoints and Check Waits fields display the same information as the numckpts and ckpwaits fields in onstat -p.

For information about tuning the checkpoint interval, see your HCL OneDB™ Performance Guide.