Re: pg_terminate_backend() issues

Tom Lane wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
> >> I think if we want pg_terminate_backend, we have to do it the way that
> >> was originally discussed: have it issue SIGTERM and fix whatever needs
> >> to be fixed in the SIGTERM code path.
>
> > Well, with no movement on this TODO item since it was removed in 8.0, I
> > am willing to give users something that works most of the time.
>
> If the users want it so bad, why has no one stepped up to do the
> testing?

Good question. Tom and I talked about this on the phone today.

I think the problem is testing to try to prove the lack of a bug. How
long does someone test to know they have reasonably proven a bug doesn't
exist?

I think the other problem is what to test. One SIGTERM problem that was
reported was related to schema changes. Certainly that has to be
tested, but what else: prepared statements, queries, HOT updates,
connect/disconnect, notify? There are lots of place to test and it is
hard to know which ones, and for how long.

I am starting to think we need to analyze the source code rather than
testing, because of what we are testing for.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +