Greg Stark <gsstark(at)mit(dot)edu> writes:
> Can someone explain exactly what the problem being defeated by writing whole
> pages to the WAL log?
Partial writes. Without the full-page image, we do not have enough
information in WAL to reconstruct the correct page contents.
>> A further optimization would be to write the backup pages to the same 8k
>> file over and over again rather than adding I/O to pg_xlog WAL files.
How does that work, and why is it a win compared to doing the same
amount of I/O to WAL?
regards, tom lane