Symptoms
Ensure virtual objects in the vSAN cluster are all healthy before proceeding disk replacement. If some virtual objects are not in a healthy state, there is a risk of data lost.
Cause
VxRail Manager first tries to remove the disk and ensure data accessibility. Typically, only partial data evacuation is required. It saves time compared with a full data migration. There is a timeout of the task. It fails if unable to complete within 30 minutes. Then, the VxRail manager removes the disk with no data migration action. If some objects are not in a healthy state, not fully in compliance with storage policy, the cluster might experience data loss.
Resolution
Before replacing the disk, check vSAN health status first. Verify that there is no vSAN object health alarm. And confirm that virtual objects are all in healthy state. Cluster > Monitor > vSAN > Virtual Objects. To understand vSAN object health state and corrective action, see VMware Knowledge Base article: https://knowledge.broadcom.com/external/article?legacyId=2108319
