Hi,
I have data nodes across 2 VLANs. Is my indexing performance will be affected ? I am indexing nearly 4.5 TB (primary+replica) ~5 billion messages per day. 44 data node instances (22 machines each with 128GB RAM - running 2 instances per machine). I am seeing indexing performance are affected. I have seen little improvement after setting
- index.store.throttle.type=none
- index.merge.scheduler.max_thread_count=1
But still indexing was not getting catch up .