Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com> writes:
> On Sat, Mar 26, 2022 at 10:57 PM Andres Freund <andres@anarazel.de> wrote:
>> -1. I think it's a perfectly reasonable function to have, it doesn't cause
>> architectural / maintenance issues to have it and there's several plausible
>> future uses for it (moving fsyncing of received WAL to different process,
>> optionally allowing logical decoding up to the written LSN, reporting function
>> for monitoring on the standby itself).
> Given the use-cases that it may have in future, I can use that
> function right now in pg_stat_get_wal_receiver instead of
> pg_atomic_read_u64(&WalRcv->writtenUpto);
I do not really see a reason to change anything at all here.
We have far better things to spend our (finite) time on.
regards, tom lane
