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

From: pgsql-hackers-owner(at)postgresql(dot)org
> [mailto:pgsql-hackers-owner(at)postgresql(dot)org] On Behalf Of Ashutosh Bapat
> In pgstat_quickdie(), I think a call to sigaddset(&BlockSig, SIGQUIT) is
> missing before PG_SETMASK(). Although there are some SIGQUIT handlers which
> do not have that call. But I guess, it will be safer to have it.

I didn't add it because pgstat_quickdie() just exits, like some other postmaster children. I thought those processes which are concerned about their termination processing call sigaddset(SIGQUIT), so I went after the processes who aren't. Is this really necessary?

> 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?

As Robert and Tom said, either exit(1) or exit(2) is OK because reaper() handles non-zero exit code the same. Regarding on_proc_reset(), stats collector is not attached to the shared memory and does not register on_proc_exit() callbacks. These situations are the same as the archiver process, so I followed it.

Regards
Takayuki Tsunakawa