pgsql: Stabilize postgres_fdw tests on 32-bit machines

From: David Rowley <drowley(at)postgresql(dot)org>
To: pgsql-committers(at)lists(dot)postgresql(dot)org
Subject: pgsql: Stabilize postgres_fdw tests on 32-bit machines
Date: 2023-11-02 23:35:56
Message-ID: E1qyhEB-003ekg-RA@gemulon.postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-committers

Stabilize postgres_fdw tests on 32-bit machines

cac169d68 adjusted DEFAULT_FDW_TUPLE_COST and that seems to have caused
a test to become unstable on 32-bit machines.

4b14e1871 tried to fix this as originally the plan was flipping between
a Nested Loop and Hash Join. That commit forced the Nested Loop, but
there's still flexibility to push or not push the sort to the remote
server and 32-bit seems to prefer to push and on 64-bit, the costs
prefer not to.

Here let's just turn off enable_sort to significantly encourage the sort
to take place on the remote server.

Reported-by: Michael Paquier, Richard Guo
Discussion: https://postgr.es/m/ZUM2IhA8X2lrG50K@paquier.xyz

Branch
------
master

Details
-------
https://git.postgresql.org/pg/commitdiff/b690e5facb4fbd2600dc477cfe64415538d306cc

Modified Files
--------------
contrib/postgres_fdw/expected/postgres_fdw.out | 29 +++++++++++++-------------
contrib/postgres_fdw/sql/postgres_fdw.sql | 2 ++
2 files changed, 16 insertions(+), 15 deletions(-)

Browse pgsql-committers by date

  From Date Subject
Next Message David Rowley 2023-11-03 07:13:08 pgsql: Add missing unicode_category.c to MSVC build scripts
Previous Message Tom Lane 2023-11-02 15:53:42 pgsql: Be more wary about NULL values for GUC string variables.