From: Tom Lane <tgl@sss.pgh.pa.us>
To: Davlet Panech <dpanech@gmail.com>
cc: pgsql-admin@lists.postgresql.org
Subject: Re: Abnormally high memory usage/OOM triggered
In-reply-to: <4d52fee7-73ba-16a9-9c27-8ae5107ee6ae@gmail.com>
References: <4d52fee7-73ba-16a9-9c27-8ae5107ee6ae@gmail.com>
Comments: In-reply-to Davlet Panech <dpanech@gmail.com>
	message dated "Wed, 17 Jan 2018 16:57:28 -0500"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <3137.1516234520.1@sss.pgh.pa.us>
Date: Wed, 17 Jan 2018 19:15:20 -0500
Message-ID: <3138.1516234520@sss.pgh.pa.us>
Precedence: bulk

Davlet Panech <dpanech@gmail.com> writes:
> I'm troubleshooting a problem with a Postgres installation (Linux): a 
> client process got killed by OOM while executing an update statement, 
> Is there a memory leak in there somewhere?
> I'm using Postgres 9.4.8 on x86_64-redhat-linux-gnu with 16GB of 
> physical RAM and 8GB of swap space.

I see a possibly relevant entry in the 9.4.10 release notes:

      Fix query-lifespan memory leak in a bulk UPDATE on a table
      with a PRIMARY KEY or REPLICA IDENTITY index

Looking at the relevant commit (ae4760d66), it seems the leak was
just a few bytes per row, but if the update touches enough rows ...

			regards, tom lane