Christopher Kings-Lynne <chriskl(at)familyhealth(dot)com(dot)au> writes:
> What's going on in initdb now that we are checkpointing too quickly?
Hm, haven't seen that myself, but I bet it's because (a) create database
now forces checkpoint, and (b) we now do two of 'em in a row.
This seems worth fixing, but it is just cosmetic, so I don't plan to
panic about it this week ...
regards, tom lane