My cluster configuration;
13 machines with 7 SSD disk
37 machines with 12 SSD disk
The Elasticsearch Version:
2.1.1
The elasticsearch.yml about disk is:
cluster.routing.allocation.disk.watermark.low: 0.9
cluster.routing.allocation.disk.watermark.high: 0.9
The cluster cotains about 1,000 indices and about 10,000 shards.
With the increase in the amount of data, the disk usage of some machines which have 7 SSD disk exceeded 90%, or even up to 100%. But the disk usage of the machines which have 12 SSD disk is less than 50%. How to deal with this situation?