During the test, we expect a node to switch its sync source to be the primary. The test times out waiting for this to happen, but the node is never able to sync from the expected node since it isn't eligible. It looks like node 1 denylists the primary due to network issues, but by the time it re-evaluates the primary as a sync source, it has already caught up to it.

The periodic noop writer should mean that the primary will keep advancing its optime, so it will be an eligible sync source after such network issues.