Re: [RFC] Should we fix postmaster to avoid slow shutdown?

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Wed, Oct 26, 2016 at 7:12 AM, Ashutosh Bapat
> <ashutosh(dot)bapat(at)enterprisedb(dot)com> wrote:
>> Also, many other SIGQUIT handlers like bgworker_quickdie() call on_exit_reset()
>> followed by exit(2) instead of just exit(1) in pgstat_quickdie(). Why is this
>> difference?

> Well, for that, you'd need to look at how postmaster.c treats those
> exit codes. exit(2) from a regular backend or background worker will
> cause a crash-and-restart cycle; I'm not sure whether the handling for
> the stats collector is similar or different.

I'm fairly sure it's different --- there are different policies for
child processes that aren't connected to shared memory (such as the
stats collector) because we shouldn't force a system-wide restart
when they crash.

regards, tom lane