There’s Flocker (https://clusterhq.com/flocker/introduction/) which is a volume plugin for Docker that persists/replicates data across all of the nodes in the cluster, so it doesn’t matter in which node your process is allocated, it will have the data available.
You might want to check that out.