On Mon, Nov 7, 2016 at 1:22 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> jdnelson@dyn.com writes:
> > We encountered a problem where checkpoints appeared to queue up and never
> > finish.
> > ...
> > 01:08:27 the system clock was synchronized. It was behind by about one
> > hour.
>
> I wonder if this is a problem similar to the autovacuum issue we fixed
> in da1a9d0f5, ie perhaps moving the system clock setting confuses the
> checkpoint timing logic.
>
That is more or less what we were thinking as well.
> What checkpoint-related settings are you using? Can you be more specific
> about the system clock change?
>
checkpoint_completion_target = 0.85
checkpoint_segments = 2048
checkpoint_timeout = 60
We have noticed that the system clock has not been the most reliable. In
this case, though, there is only evidence of a single clock correction at
the noted time. No other corrections or drift during the actual event were
detected, but that doesn't mean that there weren't any.
--
Jon Nelson