On 2019-Aug-26, Tom Lane wrote:
> I wrote:
> > I propose the attached. I'm inclined to think that the risk/benefit
> > of back-patching this is not very good, so I just want to stick it in
> > HEAD, unless somebody can explain why dead_end children are likely to
> > crash in the field.
>
> Pushed at ee3278239.
>
> I'm still curious as to the explanation for a dead_end child exiting
> with code 15, but I have no way to pursue the point.
Many thanks for all the investigation and fix!
Sadly, I have *no* idea what could have happened that would have caused
a connection at that point (my start scripts don't do it). It is
possible that I had a terminal running some shell loop on psql ("watch
psql -c something" perhaps). But I'm sure I didn't notice that when I
reported this, or I would have mentioned it. However, I have no idea
why would it have died with code 15. From my notes of what I was doing
that day, I can't find any evidence that I would have had anything in
shared_preload_libraries. (I don't have Frost's complete timestamped
shell history, however.)
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services