Disk failure prediction in your Ceph cluster using SMART.
Keeping track of your disks in a Ceph cluster is important so you could replace them in time before data loss. In this video we look into how to use SMART data to run a machine learning prediction in order predict when a disk fails. SmartMonTools is used locally to fetch the data. The data is saved in the cluster and then the prediction will show you the estimate time until failure.