I just setup a new HDP Cluster out of the box. And there are two disks on seperate datanodes have a failure according to the dfshealth.
I checked them on Linux level and they are fine. Is there a way to bring them back?
May be this article can help. It has details on how can a disk be replaced in datanode.