I have an index of around 4.1TB~ total size in a cluster of 12 nodes (9 DATA & INGEST, 3 DATA & MASTER) and all of them are set to "hot" category - we don't currently have any "warm" node in this cluster.
The problem is I have 10 replica shards in state UNASSIGNED - I added in the first comment of this post the output from Explain API command regarding this (topic doesn't allow me to enter too many chars):
Every shard is around 300GB~ (above the recommended threshold of 50GB) and initially there were disk space issues in many nodes but then I performed clean up of garbage files there. After this, the explain output stopped complaining about disk watermark hit but keeps complaining about the box_type setting.
Appreciate any help on this.
Regards,
Vinicio