Date: Tue, 8 Mar 2022 00:44:55 -0600
From: Justin Pryzby <pryzby@telsasoft.com>
To: Luiz Felipph <luizfelipph@gmail.com>
Cc: pgsql-performance@lists.postgresql.org
Subject: Re: Optimal configuration for server
Message-ID: <20220308064455.GT27651@telsasoft.com>
References: 
 <CABCV0czrGy1XfyZ-tXKH=Xts+xo6vvvYnxN4Eb7Q136zYAk9sA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <CABCV0czrGy1XfyZ-tXKH=Xts+xo6vvvYnxN4Eb7Q136zYAk9sA@mail.gmail.com>
User-Agent: Mutt/1.9.4 (2018-02-28)
Archived-At: 
 <https://www.postgresql.org/message-id/20220308064455.GT27651%40telsasoft.com>
Precedence: bulk

On Mon, Mar 07, 2022 at 08:51:24AM -0300, Luiz Felipph wrote:
> My current problem:
> 
> under heavyload, i'm getting "connection closed" on the application
> level(java-jdbc, jboss ds)

Could you check whether the server is crashing ?

If you run "ps -fu postgres", you can compare the start time ("STIME") of the
postmaster parent process with that of the persistent, auxilliary, child
processes like the checkpointer.  If there was a crash, the checkpointer will
have restarted more recently than the parent process.

The SQL version of that is like:
SELECT date_trunc('second', pg_postmaster_start_time() - backend_start) FROM pg_stat_activity ORDER BY 1 DESC LIMIT 1;

-- 
Justin