Re: signal 11 segfaults with parallel workers

Thanks! I've got the -dbg package installed and I've restarted the server
and the database this morning. We've continued to crash almost every
night, so if the restart doesn't do something strange, I should have a core
file within a day or two.

One thing that is bugging me is I think when the database crashes, it
doesn't clean up the temp_tablespace(s). I've noticed as I'm working
through this issue that the temp tablespace keeps creeping up in size and
there doesn't seem to be any obvious way to recover that space. I've been
keeping up with it for now by making the disk bigger, but obviously I can't
do that indefinitely.

I was debating making a new temp tablespace and then dropping the old one,
but there must be an easier, safe way to clear dangling temp tablespace
stuff? Google wasn't terribly helpful to uncover strategies for dealing
with temp tablespace bloat.

On Fri, Jul 28, 2017 at 3:02 PM, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
wrote:

> Rick Otten wrote:
> > I'm using the Ubuntu PostgresSQL 9.6.3 from this repo:
> > deb http://apt.postgresql.org/pub/repos/apt/ xenial-pgdg main
> >
> > It looks like there is a "-dbg" package available:
> > postgresql-9.6-dbg - debug symbols for postgresql-9.6
> >
> > I'll give that a try when I get the restart opportunity.
>
> You can install the -dbg package without waiting for a restart; it won't
> disrupt anything. Also, if you already got a core from the last crash,
> installing that package now would be enough to be able to extract info
> from the core file, assuming the -dbg package is the same version as the
> package version that was running when it crashed.
>
> --
> Álvaro Herrera https://www.2ndQuadrant.com/
> PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
>