Re: Fsync IO issue

Merlin Moncure <mmoncure@gmail.com>

From: Merlin Moncure <mmoncure@gmail.com>
To: Thomas Munro <thomas.munro@gmail.com>
Cc: ProfiVPS Support <support@profivps.hu>, Andres Freund <andres@anarazel.de>, pgsql-performance@lists.postgresql.org
Date: 2023-05-30T20:56:36Z
Lists: pgsql-performance
On Thu, May 4, 2023 at 4:23 PM Thomas Munro <thomas.munro@gmail.com> wrote:

> On Fri, May 5, 2023 at 8:37 AM ProfiVPS Support <support@profivps.hu>
> wrote:
> >  I feel like ANYTHING would be better than this. Even risking loosing
> _some_ of the latest data in case of a server crash (if it crashes we lose
> data anyways until restart, ofc we could have HA I know and we will when
> there'll be a need) .
>
> Try synchronous_commit=off:
>
> https://www.postgresql.org/docs/current/wal-async-commit.html


Yeah, or batch multiple inserts into a transaction somehow.   In the worst
case, each insert can cause multiple things to happen, write to WAL, flush
to heap (8k write), commit bit set (another 8k write), etc.  In most
workloads these steps can aggregate together in various ways but not
always.

merlin