test_check_pods_status_after_node_failure fails on 6 nodes VSphere IPI cluster due to random node selection

[am-agrawa]

When cluster has 6 or more nodes, random node selection leads to shutting down a node which doesn't have rook-ceph-xxxx pods hence pod status remains the same and test fails. We need to ensure that we select the node where rook-ceph-xxxx pods are running.