Re: Postgresql15 crash with :FATAL: could not open shared memory segment "/PostgreSQL.0000000": No such file or directory - Mailing list pgsql-bugs

From Emile Amewoto
Subject Re: Postgresql15 crash with :FATAL: could not open shared memory segment "/PostgreSQL.0000000": No such file or directory
Date
Msg-id 537973380.2573525.1692634839959@mail.yahoo.com
Whole thread Raw
In response to Re: Postgresql15 crash with :FATAL: could not open shared memory segment "/PostgreSQL.0000000": No such file or directory  ("Euler Taveira" <euler@eulerto.com>)
Responses Re: Postgresql15 crash with :FATAL: could not open shared memory segment "/PostgreSQL.0000000": No such file or directory
List pgsql-bugs
Thanks Euler, that's exactly my problem. 
I will set RemoveIPC to no.
Your update is very helpful, thanks again. 

My environment is Rhel 7.9

Regards, 
Émile 


On Mon, Aug 21, 2023 at 17:24, Euler Taveira
<euler@eulerto.com> wrote:
On Thu, Aug 17, 2023, at 5:27 PM, Emile Amewoto (ZA) wrote:

Postgres version 15 randomly crash with the following error : FATAL: could not open shared memory segment "/PostgreSQL.2771626404": No such file or directory



You didn't provide enough information about your environment (operating system,
exact Postgres version, Postgres settings, etc).

I don't know if it is your exact problem but systemd contains a setting that
removes IPC during the logout. It means that if you are using the Postgres
service account (generally the 'postgres' user) to execute administrative tasks
and you log out while you're running a query (that requires dynamic shared
memory) then you get this error.

See 'man 5 logind.conf' for details. If you want to avoid it, disable RemoveIPC
in /etc/systemd/logind.conf. (IIRC only Debian/Ubuntu defaults to yes; RHEL
doesn't.)

If that's not your problem then you have to provide additional information as I
said before.


--
Euler Taveira

pgsql-bugs by date:

Previous
From: Andrew Dunstan
Date:
Subject: Re: BUG #17994: Invalidating relcache corrupts tupDesc inside ExecEvalFieldStoreDeForm()
Next
From: Tom Lane
Date:
Subject: Re: BUG #17994: Invalidating relcache corrupts tupDesc inside ExecEvalFieldStoreDeForm()