On Fri, Jun 18, 2021 at 1:57 AM Michael Paquier <michael@paquier.xyz> wrote:
This doc addition is a bit confusing, as it could mean that each file has just one single checksum. We could be more precise, say: "When enabling checksums, each relation file block with a changed checksum is rewritten in place."
Agreed, I like that wording. New patch attached.
Should we also mention that the sync happens even if no blocks are rewritten based on the reasoning of upthread (aka we'd better do the final flush as an interrupted pg_checksums may let a portion of the files as not flushed)?
I don't know that we need to bother: the default is already to sync and one has to go out of one's way using the -N argument to NOT sync, so I think it's a pretty safe assumption to everyone (except those who read my first version of my patch!) that syncing always happens.
