assertion at postmaster start - Mailing list pgsql-hackers

From Alvaro Herrera
Subject assertion at postmaster start
Date
Msg-id 20190615160950.GA31378@alvherre.pgsql
Whole thread Raw
Responses Re: assertion at postmaster start  (Andres Freund <andres@anarazel.de>)
List pgsql-hackers
Once in a blue moon I get this assertion failure on server start:

2019-06-15 12:00:29.650 -04 [30080] LOG:  iniciando PostgreSQL 12beta1 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu
7.4.0-1ubuntu1~18.04)7.4.0, 64-bit
 
2019-06-15 12:00:29.650 -04 [30080] LOG:  escuchando en la dirección IPv4 «127.0.0.1», port 55432
2019-06-15 12:00:29.650 -04 [30080] LOG:  escuchando en el socket Unix «/tmp/.s.PGSQL.55432»
2019-06-15 12:00:29.658 -04 [30956] LOG:  el sistema de bases de datos fue apagado en 2019-06-15 12:00:24 -04
2019-06-15 12:00:29.659 -04 [30080] LOG:  proceso de servidor (PID 30107) terminó con código de salida 15
2019-06-15 12:00:29.659 -04 [30080] LOG:  terminando todos los otros procesos de servidor activos
TRAP: FailedAssertion(«!(AbortStartTime == 0)», Archivo: «/pgsql/source/master/src/backend/postmaster/postmaster.c»,
Línea:2957)
 
Aborted (core dumped)

Apologies for the Spanish -- I cannot readily reproduce this.  In
essence, this shows a normal startup, until suddenly process 30107
terminates with exit code 15, and then while shutting everything down,
postmaster hits the aforementioned assertion and terminates.

One problem with debugging this is that I don't know what process 30107
is, since the logs don't mention it.

No idea what is going on.  But I'm going to set my script to start the
server with log_min_messages=debug1, in case I hit it again ...

Has anybody else seen this?

-- 
Álvaro Herrera



pgsql-hackers by date:

Previous
From: Alexander Lakhin
Date:
Subject: Fix typos and inconsistencies for v11+
Next
From: Tom Lane
Date:
Subject: Re: pgsql: Avoid spurious deadlocks when upgrading a tuple lock