Re: "pg_ctl: the PID file ... is empty" at end of make check

On Fri, Oct 18, 2019 at 1:21 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Thomas Munro <thomas(dot)munro(at)gmail(dot)com> writes:
> > On Tue, Oct 15, 2019 at 1:55 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >> and now prairiedog has shown it too:
> >> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=prairiedog&dt=2019-10-14%2021%3A45%3A47
> >> which is positively fascinating, because prairiedog is running a
> >> bronze-age version of macOS that surely never heard of APFS.
> >> So this makes it look like this is a basic macOS bug that's not
> >> as filesystem-dependent as one might think.
>
> > Does https://github.com/macdice/unlinktest show the problem on that system?
>
> It does, though with a very low frequency:
>
> $ ./unlinktest
> $ ./unlinktest 10000
> read 0 bytes, unexpected
> $ ./unlinktest 10000
> read 0 bytes, unexpected
> read 0 bytes, unexpected
> $ ./unlinktest 10000
> $
>
> The failure rate on my recent-vintage laptop is more like one
> failure every five loops.

Wow. Ok, I'll add a note to the bug report to say it's reproducible
on "Darwin Kernel Version 8.11.0: Wed Oct 10 18:26:00 PDT 2007;
root:xnu-792.24.17~1/RELEASE_PPC" next time I'm near an Apple device
that will let me log into the bug thing. On the off-chance that
someone from Apple stumbles on this and is interested, the Radar
number is rdar://46318934 and the title is "unlink(2) is not atomic
(kernel/file system bug)".