I just noticed this in my log files of an arbiter in a replica set with 2 other nodes.

I'm not exactly sure which action caused this, at the time of it occuring I was doing tests that involved stopping/starting the VM that run MongoDB. The stack trace looks similar but not identical to SERVER-28079.

The arbiter (and all other members) got restarted automatically, and seemingly the setup "recovered" without additional user interaction.

Set up:
This is a MongoDB 3.4.x unsharded replica set running inside kubernetes on Minikube (i.e. in a single VirtualBox VM). The MongoDB containers are managed using custom scripts.

I'm reporting this mainly to provide a data point, given that this is a test setup it is hard to reproduce the issue in any meaningful way.