Re: Terminating connection because of crash of another server process

On Mon, Oct 2, 2023 at 11:34 AM Deo Felix <deofelix9(at)gmail(dot)com> wrote:

> Please note that the server is in production and debug is not enabled,
> that is why there is not much information in the log file.
>

We need the expanded error. Please enable (debug, detail) in production.
You can turn it off once the detailed error is found out.

> I'm starting working with your suggestion including reducing the work_mem
>
>
> On Fri, 29 Sept 2023 at 12:27, Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at>
> wrote:
>
>> On Fri, 2023-09-29 at 11:21 +0300, Deo Felix wrote:
>> > Please find the screenshot of error from the log
>> file postgresqlerror.png
>>
>> It is really amazing how much effort you make *not* to share
>> any relevant information with us. But there is finally some
>> interesting message hidden in your image:
>>
>> server process (PID 14257) was terminated by signal 9: Killed
>>
>> Since I don't expect that random users on your server machine
>> go around "kill -9"-ing PostgreSQL processes, it is very likely
>> that PostgreSQL is crashed by the Linux out-of-memory killer.
>>
>> You should disable memory overcommit, see
>>
>> https://www.postgresql.org/docs/current/kernel-resources.html#LINUX-MEMORY-OVERCOMMIT
>>
>> Then you will start getting "out of memory" errors instead of
>> crashes. To combat these, reduce the memory usage parameters,
>> mostly "work_mem", or reduce the number of active database sessions
>> by using a connection pool.
>>
>> Yours,
>> Laurenz Albe
>>
>