Horton

Re: Fix lag columns in pg_stat_replication not advancing when replay LSN stalls

Fujii Masao <masao.fujii@gmail.com>

View thread View on postgr.es
From: Fujii Masao <masao.fujii@gmail.com>
To: Chao Li <li.evan.chao@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2025-10-20T01:20:41Z
Lists: pgsql-hackers 
On Mon, Oct 20, 2025 at 10:12 AM Fujii Masao <masao.fujii@gmail.com> wrote:
>
> On Sat, Oct 18, 2025 at 9:16 AM Chao Li <li.evan.chao@gmail.com> wrote:
> > I think I put all concentration on the big picture yesterday, so I ignored this implementation detail:
> >
> > ```
> > +               if (lag_tracker->overflowed[head].lsn > lsn)
> > +                       return now - lag_tracker->overflowed[head].time;
> > ```
> >
> > Should this “>” be “>=“?
>
> I think either ">" or ">=" would work. However, the following loop
> in LagTrackerRead() advances the read head when the LSN
> in the current buffer entry is equal to the given LSN

I forgot to mention which loop actually advances the read head in that case.
It's the following code in LagTrackerRead().

------------------------
/* Read all unread samples up to this LSN or end of buffer. */
while (lag_tracker->read_heads[head] != lag_tracker->write_head &&
   lag_tracker->buffer[lag_tracker->read_heads[head]].lsn <= lsn)
{
time = lag_tracker->buffer[lag_tracker->read_heads[head]].time;
lag_tracker->last_read[head] =
lag_tracker->buffer[lag_tracker->read_heads[head]];
lag_tracker->read_heads[head] =
(lag_tracker->read_heads[head] + 1) % LAG_TRACKER_BUFFER_SIZE;
}
------------------------

Regards,

-- 
Fujii Masao

Commits

  1. Add comments explaining overflow entries in the replication lag tracker.

  2. Fix stalled lag columns in pg_stat_replication when replay LSN stops advancing.