Re: potential stuck lock in SaveSlotToPath()

On 2020-Mar-25, Peter Eisentraut wrote:

> On 2020-03-20 16:38, Robert Haas wrote:
> > On Fri, Mar 20, 2020 at 11:32 AM Peter Eisentraut
> > <peter(dot)eisentraut(at)2ndquadrant(dot)com> wrote:
> > > On 2020-03-19 16:38, Robert Haas wrote:
> > > > Incidentally, the wait-event handling in SaveSlotToPath() doesn't look
> > > > right for the early-exit cases either.
> > >
> > > There appear to be appropriate pgstat_report_wait_end() calls. What are
> > > you seeing?
> >
> > Oh, you're right. I think I got confused because the rename() and
> > close() don't have that, but those don't have a wait event set either.
> > Sorry for the noise.
>
> Any concerns about applying and backpatching the patch I posted?

It looks a straight bug fix to me, I agree it should be back-patched.

> The talk about reorganizing this code doesn't seem very concrete at the
> moment and would probably not be backpatch material anyway.

Agreed on both counts.

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services