Checkpoints are essential to data integrity as well as performance. The further checkpoints are apart, the better performance usually is. In PostgreSQL, the default configuration is usually fairly conservative and checkpoints are therefore comparatively fast. If a lot of data is changed in the database core at the same time, it can be that PostgreSQL tells us that it considers checkpoints to be too frequent. The LOG file will show the following entries:
LOG: checkpoints are occurring too frequently (2 seconds apart)LOG: checkpoints are occurring too frequently (3 seconds apart)
During heavy writing due to dump/restore or due to some other large operation, PostgreSQL might notice that the configuration parameters ...
