On Tue, 2024-07-23 at 16:21 +0300, Heikki Linnakangas wrote:
> So, I got a feeling that adding this to the rmgr interface is not
> quite
> right. The rmgr callbacks are for things that run when WAL is
> *replayed*, while checkpoints are related to how WAL is generated.
> Let's
> design this as an independent hook, separate from rmgrs.
That's a good way to look at it, agreed.
Regards,
Jeff Davis