As suggested, the result I've posted used standard checkpoint_timeout.
With longer checkpoint timeout, overall WAL size tend to decrease. So
I've also run DBT-2. The measurement shows that the resultant WAL size
is reasonably small, while preserving full page writes to WAL,
maintaining the
> Our evaluation result is as follows:
> Database size: 2GB
> WAL size (after 10hours pgbench run): 48.3GB
> gzipped size: 8.8GB
> removal of the physical log: 2.36GB
> fullpage_writes=off log size: 2.42GB
> I'm planning to do archive log size evalutation with other benchmarks
> such as DBT-2 as
