I apologize if this topic has been covered for a long time.
But I never found it in the open Elastic Сommunity.
We have 11 cluster servers in the infrastructure ES:
Servers HOT - 3
Servers WARM - 3
Servers COULD - 2
Servers Master - 3
Some time ago I noticed that on HOT servers there is a discrepancy in disk space utilization.
The HOT 1 server began to fill up disk space faster than on the HOT 2 and HOT 3 servers.
I don’t understand why there is an uneven distribution of indexes (shards) between HOT servers in the cluster.
Sending to these three servers HOT-1 HOT-2 HOT-3 is configured on all logstash servers in the output section.
Can anyone suggest what are the special parameters for properly balancing the distribution of indexes ?
How to understand this and, most importantly, how to set up the correct distribution of indexes on servers with the HOT role ?
Moreover, on all other servers with the roles of WARM and COULD, the distribution of the index record is even.
Please don't post pictures of text, logs or code. They are difficult to read, impossible to search and replicate (if it's code), and some people may not be even able to see them
At first, on the first and then on the second HOT servers, I had more than 85% full.
New shards could not be recognized (unsigned).
found ~ 1000 unsigned shards.
After I corrected the ILM policy by expanding storage towards WARM servers, all shards recovered, in less than a week the space on XOT servers stabilized.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.