How will the ElasticSearch cluster behave if one of the nodes runs out of disk space?

Question

linux

There is an ElasticSearch cluster of 5 machines. Disk space (SSD disk) starts to be downloaded on two machines. How will the cluster behave if the free space runs out on these two nodes?
Next to the SSD is another HDD.
Is it possible to transfer part of the indexes to the HDD, and leave only fresh ones on the SSD? In which direction to dig to understand the issue?

Answer 1

Some information on setting up storage directories in ES: www.elastic.co/guide/en/elasticsearch/reference/1....
Might be interesting. Did a little experience. On one of the nodes filled the disk space. The cluster stopped creating new index shards on this node and the node remained as read/delete only.