I have 10 data nodes of which 3 have additional ML and transform role to it. I could observe that almost all the primary shards are assigned to nodes with ml role assigned to it. And replicas are distributed. This is causing disk full issues and ES is not able to rebalance. Does node roles affect shard allocation.
PS: all nodes have "data" role assigned to it
Do all data nodes in the cluster have the same amount of disk space?
Are all nodes running exactly the same version of Elasticsearch?
yes all nodes have same disk capacity and all are running in same version(8.15.1).