I saw a lot of posts and papers that set the following values:
max_parallel_workers - set the number as the number of cores.
max_parallel_workers_per_gather - default set to 2 or set as max_parallel_workers / 2
when using parallel query the benfit and speed improvment is based on max_parallel_workers_per_gather
- why I can't find posts that recommend to set
max_parallel_workers_per_gather = max_parallel_workers? - if we set
max_parallel_workers_per_gather = max_parallel_workerswhat traps can be ?