test: remove Clear Linux networking workaround

We had to ensure that the controller didn't run on the same node as
the node drivers. We no longer test on Clear Linux, so this workaround
is no longer necessary.

test/start-kubernetes.sh

@@ -424,12 +424,6 @@ function init_kubernetes_cluster() (
         sed -e 's/.*: v\([0-9]*\)\.\([0-9]*\)\..*/\1.\2/' \
         >"${CLUSTER_DIRECTORY}/kubernetes.version"
 
-    # Allow all pods to run also on the master node *and* prevent the PMEM-CSI controller
-    # from running on the normal nodes. Workaround for networking issues on
-    # Clear Linux (https://github.com/intel/pmem-csi/issues/555).
-    ssh $SSH_ARGS ${CLOUD_USER}@${master_ip} kubectl taint nodes pmem-csi-${CLUSTER}-master node-role.kubernetes.io/master:NoSchedule-
-    ssh $SSH_ARGS ${CLOUD_USER}@${master_ip} kubectl label node -l storage=pmem pmem-csi.intel.com/controller=no
-
     kubernetes_usage
 )