From: | Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com> |
---|---|
To: | y(dot)kikuchi0714(at)gmail(dot)com |
Cc: | pgsql-bugs(at)lists(dot)postgresql(dot)org |
Subject: | Re: BUG #17269: Why is virtual memory usage of PostgreSQL growing constantly? |
Date: | 2021-11-08 01:00:13 |
Message-ID: | 20211108.100013.1114023989331664882.horikyota.ntt@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-bugs |
At Fri, 5 Nov 2021 18:15:01 +0900, 菊池祐 <y(dot)kikuchi0714(at)gmail(dot)com> wrote in
> I attached postgresql.conf file.
> max_connections = 10000 # (change requires restart)
I think the moderate setting of max_connection is 100 to 200 or 300 on
a 128GB box. Depending on workload but it seems that you need to
reduce it to the same range. 1000 might be possible for super-light
weight workload but it is almost definite that 10000 don't fit 128GB
memory, or on CPUs with 16-32 cores.
I'm sure the clients are executing queries with quite low
frequency. Maybe you want to use connection pooler such like Pgpool-II
or pgbouncer, yandex/odyssey if the clients need to maintain their
connections to the server for a long time.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center
From | Date | Subject | |
---|---|---|---|
Next Message | Michael Paquier | 2021-11-08 03:53:58 | Re: pg_upgrade test for binary compatibility of core data types |
Previous Message | Peter Geoghegan | 2021-11-08 00:54:45 | Re: BUG #17245: Index corruption involving deduplicated entries |