On Sat, Mar 30, 2019 at 10:12:33PM +0900, Michael Paquier wrote:
> Okay, I'll use that then.
And finally committed. I have changed the debug1 message so as "to
prevent wraparound" is used instead of "anti-wraparound". I have
noticed something which was also missing from all the patches proposed
on this thread: a reset of the progress state was not done, and we
need to call pgstat_progress_end_command() in order to do that.
--
Michael