Amazon RDS (PostgreSQL): optimizes memory usage

Question

In Amazon RDS (PostrgreSQL version) I did not quite understand. Some queries take a long time to show their results. I set all the relevant indexes (as shown using EXPLAIN). Therefore, I think this is not due to my circuit design.

I do not use a large machine (m3.xlarge) as the larger ones are too expensive. My database is about 300 GB in size.

Postgres does not seem to use all available memory (only ~ 5 GB, the "Free memory" report on the console shows that there is always ~ 10 GB for free ...). I am trying to tune my “parameter group” suggested by tune-your-postgres-rds-instance , especially set EFFECTIVE_CACHE_SIZE to 70%. But it does not change anything.

I’m probably mistaken somewhere ... Any idea?

Answer 1

To make more memory available for your requests, you configure work_mem . There are consequences for this, since with this memory on the backend.

effective_cache_size . .

" " - - , ( ) postgres .

shared_buffers, postgres , , , 25% .