Docker Swarm not clearing entries from etcd on reboot, causing cluster to fail

I’m running Docker Swarm Standalone, and Docker isn’t clearing entries from etcd when it’s shutdown. This is causing horrible breakage in the cluster when a CoreOS node reboots.

This is the closest related issue I’ve been able to find with Docker so far:

At the moment, my thought is to disable automatic reboots on the CoreOS cluster until I can switch to a more reliable platform like Kubernetes.

Does anyone have ideas for a better fix?

Running CoreOS Stable 1409.5.0

Docker version 1.12.6, build a82d35e
Docker Swarm Standalone 1.2.6