Re: Improve pg_sync_replication_slots() to wait for primary to advance

[shveta malik]

On Wed, Nov 12, 2025 at 1:54 PM Ajin Cherian <itsa...@gmail.com> wrote:
>
>
> Attaching patch v23 addressing these comments.
>

Thanks for the patch.

I observed that if the API is taking a nap in between slot sync cycles
and a promotion is triggered during that time, the promotion has to
wait for the entire nap period to finish before slot-sync stops and
the process can continue. There should be a mechanism to wake up the
backend so the API can exit early once stopSignaled is set. How about
doing SetLatch for the process doing synchronization in
ShutDownSlotSync()?

thanks
Shveta