Cause: A node's attribute manager writes all its transient node attributes from memory to the CIB after winning the election for attribute writer, even if its node has requested shutdown. Consequence: If a node is DC, requests shutdown, and wins the attribute writer election after its controller has left the cluster but before its attribute manager has left, it can write out its shutdown attribute to the CIB. The next time it rejoins the cluster, it will be immediately shut down. Fix: A node's attribute manager should not write out its attributes after winning an election if shutdown has been requested for its node. Result: A leaving DC node does not have an unexpected shutdown the next time it rejoins.