Hi
On 10/25/2018 09:47 AM, Michael Paquier wrote:
> And committed. I double-checked the code, and tweaked a bit the tests
> so as the test using wait_mode = false is removed as it did not seem
> worth the extra cycles. I also added a check on the return value of
> pg_promote when using the wait mode. Another thing was that the
> function needs to be parallel-restricted.
Documentation for this [*] says "Returns true if promotion is successful and false otherwise",
which is not correct if "wait" is false, as it will always return TRUE.
[*] https://www.postgresql.org/docs/devel/static/functions-admin.html#FUNCTIONS-RECOVERY-CONTROL
Attached patch contains a suggested rewording to clarify this.
Regards
Ian Barwick
--
Ian Barwick http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services