On Sat, Nov 17, 2018 at 10:15:08PM +0100, Daniel Gustafsson wrote:
> > On 15 Nov 2018, at 22:42, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> > Further point about that: pg_regress's method of creating a temp
> > directory under /tmp is secure only on machines with the stickybit
> > set on /tmp; otherwise it's possible for an attacker to rename the
> > temp dir out of the way and inject his own socket. We agreed that
> > that was an okay risk to take for testing purposes, but I'm much
> > less willing to assume that it's okay for production use with
> > pg_upgrade.
>
> That’s a good point, it’s not an assumption I’d be comfortable with when it
> deals with system upgrades.
As in https://postgr.es/m/flat/20140329222934.GC170273@tornado.leadboat.com, I
maintain that insecure /tmp is not worth worrying about in any part of
PostgreSQL.