Re: PostgreSQL XID exceeded crash.

Mario Lopez wrote:
> Hi Alvaro,
>
> Ok, maybe is not a crash but this issue makes postgresql reap all of
> it's child processes and restart operations, after this error I get in
> pg_restore this message:
>
> pg_restore: [archivador (bd)] Error durante PROCESAMIENTO DE TABLA DE
> CONTENIDOS:
> pg_restore: [archivador (bd)] Error en entrada de la tabla de contenidos
> 1893; 0 1615776 TABLE DATA xxxxx postgres
> pg_restore: [archivador (bd)] could not execute query: server closed the
> connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> Command was: INSERT INTO xxxxx VALUES (1755298, 2, 117, 28, 1155,
> 52457, 1, 1, '', '', '', '', NULL);
> pg_restore: [archivador (bd)] could not execute query: sin resultado
> desde el servidor
>
> I have a attached a log from the XID exhausted resources till it reaps
> all of it's child processes, which ends in a crash from pg_restore or
> any application using this table.

The actual problem is shown in these lines:

DEBUG: 00000: reaping dead processes
UBICACION: reaper, postmaster.c:2021
DEBUG: 00000: proceso de servidor (PID 31622) termino con codigo de salida 255
UBICACION: LogChildExit, postmaster.c:2419
LOG: 00000: proceso de servidor (PID 31622) termino con codigo de salida 255

What you need to find out is why is this process is terminating with
code 255. What is it anyway? If you added %p to log_line_prefix we
could have a better idea of what is it and what is it doing.

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support