I would like to know how to raise the performance for massive data loading.
As online documentation, Memsql support two types of parallel bulk loading.
I have tried LOAD DATA INFILE and FS PIPELINE. However, I am not sure those worked well.
Please check my HW information in the below and recommend how to ingest data much faster.
Server : (28 CPUs, 192GB RAM, 1TB DISK) per node
Cluster : 1 master, 1 child, 2 leaf node
Test data : 100GB