On 7/24/17 3:28 PM, David Steele wrote:
>
> Yes, and this is another behavioral change to consider -- one that is
> more likely to impact users than the change to pg_stop_backup(). If
> pg_basebackup is not currently waiting for WAL on a standby (but does on
> a primary) then that's pretty serious.
My bad, before PG10 pg_basebackup did not check that WAL was archived on
a primary *or* a standby unless the -x option was specified, as documented.
--
-David
david(at)pgmasters(dot)net