We have a basic requirement where we need to implement the parallel index scan instead of parallel seq scan at table level. Please suggest the best way to do this.
Postgresql DB Version : 12.6
That's far behind the latest patch level.
Table size : 2-5 GB
OS : Centos-7
RAM : 96 GB
CPU : 32
shared_buffer=20GB
Plural: shared_buffers
Effective_cache_size=60GB
Max_worker_process = 32
Max_parallel_worker=32
Max_parallel_worker_per_gather=8
Please let me know if any further details required .
Is there proper index support on the relevant columns? Have you analyzed the tables? Are the values you're filtering on uncommon enough for PG to think that indexes are relevant?