On 4/6/22 12:59, Andres Freund wrote:
> Hi,
>
> On 2022-04-06 11:50:11 -0400, Andrew Dunstan wrote:
>> It does work, but Tom prefers not to have the test at all, so I'll just
>> rip it out.
> If I understand correctly the reason a large table is needed is to test
> parallelism, right? Wouldn't the better fix be to just tweak the parallelism
> settings for that table? See select_parallel.sql:
>
> -- encourage use of parallel plans
> set parallel_setup_cost=0;
> set parallel_tuple_cost=0;
> set min_parallel_table_scan_size=0;
> set max_parallel_workers_per_gather=4;
>
> might be worth also setting
> set parallel_leader_participation = off;
>
> to avoid the leader processing everything before workers have even started up.
>
OK, done that way, thanks. I also kept the table as unlogged and dropped
it at the end.
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com
