Date: Fri, 8 Oct 2021 19:45:59 -0300
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Wells Oliver <wells.oliver@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Bruce Momjian <bruce@momjian.us>,
	pgsql-admin <pgsql-admin@postgresql.org>
Subject: Re: 13.4 on RDS, SSL SYSCALL EOF on restore
Message-ID: <202110082245.4wq5lie5euyx@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: 
 <CAOC+FBU=X5g129058p9=L2VvRNpsQ2ue95BLvn72JyKdwrHxJg@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/202110082245.4wq5lie5euyx%40alvherre.pgsql>
Precedence: bulk

On 2021-Oct-08, Wells Oliver wrote:

> Dug out some more logging:

> 2021-10-08 20:35:08 UTC::@:[12682]:LOG:  server process (PID 3970) was terminated by signal 9: Killed
> 2021-10-08 20:35:08 UTC::@:[12682]:DETAIL:  Failed process was running: CREATE INDEX ...

So what's happening here is that the instance is running out of RAM
while creating some index, and the kernel is killing the process.  I
would probably blame the combination of shared_buffers=4GB with
maintenance_work_mem=2GB, together with the instance's total RAM.

I would suggest to keep an eye on memory usage (if there is such a
thing) in the RDS console; with some experimentation, reducing
maintenace_work_mem as Bruce adviced and/or shared_buffers, you can
probably find some combination that doesn't take as long as with the
default params but is able to complete.

-- 
Álvaro Herrera         PostgreSQL Developer  —  https://www.EnterpriseDB.com/