Re: BUG #15021: Postgres crashes unexpectedly

On 01/19/2018 10:47 PM, PG Bug reporting form wrote:
> The following bug has been logged on the website:
> 
> Bug reference:      15021
> Logged by:          Wallace Baggaley
> Email address:      wally.baggaley@gmail.com
> PostgreSQL version: 9.6.5
> Operating system:   Debian GNU Linux 8 container on CentOS 7.2.1511
> Description:        
> 
> Postgres crashes unexpectedly and indicates that it received a signal to end
> based on the following logs. But checking user history and journald show
> nothing has terminated the application. Is it possible the log "LOG: 
> received smart shutdown request" could be caused by something other than a
> signal. What signals could be associated with this log or is this possibly
> related to a crash?
> 

Apparently something happened between 10:34:07.436 and 11:26:04.678, as
indicated by these log messages:

    January 19th 2018, 11:26:04.692 LOG:  database system was not
        properly shut down; automatic recovery in progress
    January 19th 2018, 11:26:04.678 LOG:  database system was
        interrupted; last known up at 2018-01-19 16:41:01 UTC
    January 19th 2018, 10:34:07.436 DETAIL:  Key ("code")=(ABC123)
        already exists.

but it's damn impossible to say why that happened. You have to inspect
some additional logs from that machine - it might have been rebooted,
the process might have been killed by OOM killer (although that usually
writes additional stuff to the log) etc.

My recommendation - look at dmesg and logs for other services on the
same host.


regards

-- 
Tomas Vondra                  http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services