I have been trying to connect to the cluster using the connection string and specifying the cluster name. This should in theory connect me directly to the primary when the server specified in the connection string is one of the secondary nodes.
This failed to work and the message I received was that it could not connect to the primary node on 127.0.1.1. This is not the correct loopback address, so it failed. In the /etc/hosts file the name resolution for m103 was set to 127.0.1.1 instead of 127.0.0.1. Once I changed this, it connected correctly.
I assume this was set up as part of the vagrant configuration when the virtual machine was created, so can this be changed to help other users who have experienced the same problem? It causes some unnecessary confusion otherwise, despite the node configuration being correct.