Self-healing

In the case of unexpected downtime, the remaining nodes can still serve traffic. If new data is added to the cluster while one of the nodes is down, this data needs to be synchronized once the node is brought back up.

GlusterFS will automatically self-heal these new files once they're accessed, triggering a self-heal operation between the nodes and copying the missing data. This is transparent to the users and clients.