How does ece limit disk iops for es node?

(Rockybean) #1

As we know, docker provides capability to limit disk iops by using 'blkio-weight' parameter.

I notice ece set different 'blkio-weight' for different size es node.

What is the strategy behind this?


How to configure a deployment use more disk resource?
(Alex Piggott) #2

The relative weight of the different clusters is proportional to their capacity (which is how CPU shares are also calculated)

Specifically, the calculation is something like min(1000, 1024*(cluster_mem/total_allocator_memory)*120%), but that's an internal implementation detail subject to change.


(system) closed #3

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.