To evict disks for a node,

1. Make sure the disk is disabled for scheduling and set to Disable.
2. Set Eviction Requested to true and save.

To evict a node,

1. Make sure the node is disabled for scheduling and set to Disable.
2. Set Eviction Requested to true, and save.

The number on the selected disks or nodes should be reduced to 0 once the eviction is a success.

If you click on the Replicas number, it will show the replica name on this disk. When you click on the replica name, the Longhorn UI will redirect the webpage to the corresponding volume page, and it will display the volume status. If there is any error, e.g. no space, or couldn’t find another schedulable disk (schedule failure), the error will be shown. All of the errors will be logged in the Event log.