Re: Peripatus: Can someone look?

On Wed, Oct 2, 2019 at 4:49 AM Larry Rosenman <ler(at)lerctr(dot)org> wrote:
> On 10/01/2019 10:46 am, Tom Lane wrote:
> > Larry Rosenman <ler(at)lerctr(dot)org> writes:
> >> My Buildfarm animal (peripatus) has been failing check since
> >> yesterday.
> >> Can someone look at it?
> >
> > It's been doing this in parallel queries, in v11 and up:
> >
> > 2019-09-29 19:00:15.534 CDT [49513:1] ERROR: could not open shared
> > memory segment "/PostgreSQL.1225945786": Permission denied
> > 2019-09-29 19:00:15.535 CDT [48596:491] pg_regress/partition_prune
> > ERROR: parallel worker failed to initialize
> > 2019-09-29 19:00:15.535 CDT [48596:492] pg_regress/partition_prune
> > HINT: More details may be available in the server log.
> > 2019-09-29 19:00:15.535 CDT [48596:493] pg_regress/partition_prune
> > CONTEXT: PL/pgSQL function explain_parallel_append(text) line 5 at
> > FOR over EXECUTE statement
> > 2019-09-29 19:00:15.535 CDT [48596:494] pg_regress/partition_prune
> > STATEMENT: select explain_parallel_append('select avg(ab.a) from ab
> > inner join lprt_a a on ab.a = a.a where a.a in(1, 0, 0)');
> > 2019-09-29 19:00:15.535 CDT [48596:495] pg_regress/partition_prune
> > WARNING: could not remove shared memory segment
> > "/PostgreSQL.1225945786": Permission denied
> >
> > which looks like an external problem to me --- certainly, nothing
> > we changed yesterday would explain it. Did you change anything
> > about the system configuration, or do a software update?
> >
> > regards, tom lane
>
> I did do an upgrade to a later SVN rev.
>
> Let me reboot and see if that fixes anything.
>
> (this is -CURRENT on FreeBSD, so it's always a moving target).

Hi Larry,

I'm seeing this on my FreeBSD 13 bleeding edge system too (built a
couple of days ago) and will see if I can find out what's up with
that. The most obvious culprit is the stuff that just landed in the
kernel to support Linux-style memfd_create() and thereby changed
around some shm_open() related things. Seems to be clearly not a
PostgreSQL problem.

Thanks,
Thomas