I'm benchmarking Elasticsearch and observes high buffer_pools for mapped, it keeps on growing when performing heavy indexing and search operations at the same time.

"buffer_pools":
  {"mapped":{"count":1264,"used_in_bytes":114462548303,"total_capacity_in_bytes":114462548303},"direct":{"count":141,"used_in_bytes":539763989,"total_capacity_in_bytes":539763988}},"classes":{"current_loaded_count":11601,"total_loaded_count":11820,"total_unloaded_count":219}}

That's around 114GB on buffer_pools. Elasticsearch is running on machine with 32GB RAM (16GB for ES heap), memory lock enabled.

Please advice on what might be the cause of this high memory usage on buffer_pools. And why it exceeds RAM by significant amount? does it mean swapping is somehow happening?

0 Answers