Message-ID: <4ae13029956a4bbcb7f60aee6f97b62c6bf85b98.camel@cybertec.at>
Subject: Re: free(): invalid pointer , when to dumping data to a new server
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Lars Aksel Opsahl <Lars.Opsahl@nibio.no>,
	"pgsql-admin@lists.postgresql.org"
	 <pgsql-admin@lists.postgresql.org>
Date: Tue, 27 Oct 2020 15:10:50 +0100
In-Reply-To: 
 <AM7P189MB1028645B919276B0B6055FA49D190@AM7P189MB1028.EURP189.PROD.OUTLOOK.COM>
References: 
	<AM7P189MB10282FF31D45127630C767739D1D0@AM7P189MB1028.EURP189.PROD.OUTLOOK.COM>
	,<AM7P189MB1028756DDD017F9828C8FDE99D1D0@AM7P189MB1028.EURP189.PROD.OUTLOOK.COM>
	 <AM7P189MB1028645B919276B0B6055FA49D190@AM7P189MB1028.EURP189.PROD.OUTLOOK.COM>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.36.5 (3.36.5-1.fc32) 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Mon, 2020-10-26 at 10:22 +0000, Lars Aksel Opsahl wrote:
> I have done some more testing and postgres also crashes right after  'pg_restore  --section=pre-data -1'
>  is done also.  Here I also see the 'free(): invalid pointer' in the postgres log.
> 
> This happens on with Postgres 12 and Postgres 13. and any version of postgis I have tested with. 
> 
> The pre-data input is little bit more than 200.000 lines with more than 6000 CREATE statements,
>  if I test with the first 100 lines it does not crash. It seems like no data are lost here because
>  the server does recovery at startup.
> 
> I tested on another with CentOS and the same software and it worked ok
> 
> The server diagnostic does not indicate any memory errors.
> 
> Dnf search does not give mee any debuginfo file for postgres

You'll have to edit "/etc/yum.repos.d/pgdg-fedora-all.repo" and
set the "PostgreSQL 13 for Fedora $releasever - $basearch - Debuginfo"
repository to "enabled=1".

Then you can install the debuginfo packages and generate a stack trace.

Yours,
Laurenz Albe
-- 
Cybertec | https://www.cybertec-postgresql.com