I have an issue with Elastic allocation decider. Sometimes Elastic allocates new indices to nodes which are quite full for no reason. On the other nodes there is a plenty of space and it allocates indices on the full ones.
Would you please give me advice how to solve this unbalanced problem?
How many nodes do you have in the cluster? Do you have different zones configured? Are you using shard allocation awareness and/or filtering? Are all nodes using exactly the same version of Elasticsearch?
24 nodes
Im not sure what zones you are talking about.
Im using shard allocation awareness (you can see in my configuration file) - dc1 and dc2
All nodes have the same Elasticsearch version.
Can you please provide statistics, e.g. output from the cat nodes API, instead of images that are hard to read and interpret? Are all nodes exactly the same specification? Which Elasticsearch version are you using?
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.