diff --git a/hack/e2e-suite/guestbook.sh b/hack/e2e-suite/guestbook.sh
index bc2956bb763..8b348b36fe2 100755
--- a/hack/e2e-suite/guestbook.sh
+++ b/hack/e2e-suite/guestbook.sh
@@ -29,23 +29,21 @@ source "${KUBE_ROOT}/cluster/$KUBERNETES_PROVIDER/util.sh"
 GUESTBOOK="${KUBE_ROOT}/examples/guestbook"
 
 # Launch the guestbook example
-$KUBECFG -c "${GUESTBOOK}/redis-master.json" create /pods
-$KUBECFG -c "${GUESTBOOK}/redis-master-service.json" create /services
-$KUBECFG -c "${GUESTBOOK}/redis-slave-controller.json" create /replicationControllers
+${KUBECTL} create -f  "${GUESTBOOK}"
 
-sleep 5
+sleep 15
 
-POD_LIST_1=$($KUBECFG '-template={{range.items}}{{.id}} {{end}}' list pods)
+POD_LIST_1=$(${KUBECTL} get pods -o template '--template={{range.items}}{{.id}} {{end}}')
 echo "Pods running: ${POD_LIST_1}"
 
-$KUBECFG stop redis-slave-controller
-# Needed until issue #103 gets fixed
-sleep 25
-$KUBECFG rm redis-slave-controller
-$KUBECFG delete services/redis-master
-$KUBECFG delete pods/redis-master
+# TODO make this an actual test. Open up a firewall and use curl to post and
+# read a message via the frontend
 
-POD_LIST_2=$($KUBECFG '-template={{range.items}}{{.id}} {{end}}' list pods)
+${KUBECTL} stop rc redis-slave-controller
+${KUBECTL} delete services redis-master
+${KUBECTL} delete pods redis-master
+
+POD_LIST_2=$(${KUBECTL} get pods -o template '--template={{range.items}}{{.id}} {{end}}')
 echo "Pods running after shutdown: ${POD_LIST_2}"
 
 exit 0
diff --git a/hack/e2e-suite/liveness.sh b/hack/e2e-suite/liveness.sh
index 32f0913b871..7178430289f 100755
--- a/hack/e2e-suite/liveness.sh
+++ b/hack/e2e-suite/liveness.sh
@@ -37,12 +37,12 @@ fi
 function teardown() {
   echo "Cleaning up test artifacts"
   for test in ${liveness_tests}; do
-    ${KUBECFG} delete pods/liveness-${test}
+    ${KUBECTL} delete pods liveness-${test}
   done
 }
 
 function waitForNotPending() {
-  pod_id_list=$($KUBECFG '-template={{range.items}}{{.id}} {{end}}' -l test=liveness list pods)
+  pod_id_list=$(${KUBECTL} get pods -o template '--template={{range.items}}{{.id}} {{end}}' -l test=liveness)
   # Pod turn up on a clean cluster can take a while for the docker image pull.
   all_running=0
   for i in $(seq 1 24); do
@@ -50,7 +50,7 @@ function waitForNotPending() {
     sleep 5
     all_running=1
     for id in $pod_id_list; do
-      current_status=$($KUBECFG -template '{{.currentState.status}}' get pods/$id) || true
+      current_status=$(${KUBECTL} get pods $id -o template '--template={{.currentState.status}}') || true
       if [[ "$current_status" == "Pending" ]]; then
         all_running=0
         break
@@ -70,18 +70,18 @@ trap "teardown" EXIT
 
 for test in ${liveness_tests}; do
   echo "Liveness test: ${test}"
-  ${KUBECFG} -c ${KUBE_ROOT}/examples/liveness/${test}-liveness.yaml create pods
+  ${KUBECTL} create -f ${KUBE_ROOT}/examples/liveness/${test}-liveness.yaml
   waitForNotPending
 
-  before=$(${KUBECFG} '-template={{.currentState.info.liveness.restartCount}}' get pods/liveness-${test})
+  before=$(${KUBECTL} get pods "liveness-${test}" -o template '--template={{.currentState.info.liveness.restartCount}}')
   while [[ "${before}" == "<no value>" ]]; do
-    before=$(${KUBECFG} '-template={{.currentState.info.liveness.restartCount}}' get pods/liveness-${test})
+    before=$(${KUBECTL} get pods "liveness-${test}" -o template '--template={{.currentState.info.liveness.restartCount}}')
   done
-  
+
   echo "Waiting for restarts."
   for i in $(seq 1 24); do
-    sleep 10 
-    after=$(${KUBECFG} '-template={{.currentState.info.liveness.restartCount}}' get pods/liveness-${test})
+    sleep 10
+    after=$(${KUBECTL} get pods "liveness-${test}" -o template '--template={{.currentState.info.liveness.restartCount}}')
     echo "Restarts: ${after} > ${before}"
     if [[ "${after}" == "<no value>" ]]; then
       continue
diff --git a/hack/e2e-suite/monitoring.sh b/hack/e2e-suite/monitoring.sh
index 73206842ee7..78df2d8a67d 100755
--- a/hack/e2e-suite/monitoring.sh
+++ b/hack/e2e-suite/monitoring.sh
@@ -28,7 +28,6 @@ source "${KUBE_ROOT}/cluster/$KUBERNETES_PROVIDER/util.sh"
 
 MONITORING="${KUBE_ROOT}/cluster/addons/cluster-monitoring"
 KUBECTL="${KUBE_ROOT}/cluster/kubectl.sh"
-KUBECFG="${KUBE_ROOT}/cluster/kubecfg.sh"
 BIGRAND=$(printf "%x\n" $(( $RANDOM << 16 | $RANDOM ))) # random 2^32 in hex
 MONITORING_FIREWALL_RULE="monitoring-test-${BIGRAND}"
 
@@ -49,13 +48,9 @@ function setup {
 }
 
 function cleanup {
-  "${KUBECFG}" resize monitoring-influx-grafana-controller 0 &> /dev/null || true
-  "${KUBECFG}" resize monitoring-heapster-controller 0 &> /dev/null || true
-  while "${KUBECTL}" get pods -l "name=influxGrafana" -o template -t {{range.items}}{{.id}}:{{end}} | grep -c . &> /dev/null \
-    || "${KUBECTL}" get pods -l "name=heapster" -o template -t {{range.items}}{{.id}}:{{end}} | grep -c . &> /dev/null; do
-    sleep 2
-  done
-  "${KUBECTL}" delete -f "${MONITORING}/" &> /dev/null || true
+  "${KUBECTL}" stop rc monitoring-influx-grafana-controller &> /dev/null || true
+  "${KUBECTL}" stop rc monitoring-heapster-controller &> /dev/null || true
+r "${KUBECTL}" delete -f "${MONITORING}/" &> /dev/null || true
 
   # This only has work to do on gce and gke
   if [[ "${KUBERNETES_PROVIDER}" == "gce" ]] || [[ "${KUBERNETES_PROVIDER}" == "gke" ]]; then
diff --git a/hack/e2e-suite/pd.sh b/hack/e2e-suite/pd.sh
index d6e34f8b141..673e69d2fe4 100755
--- a/hack/e2e-suite/pd.sh
+++ b/hack/e2e-suite/pd.sh
@@ -35,13 +35,13 @@ config="/tmp/${disk_name}.yaml"
 
 function delete_pd_pod() {
   # Delete the pod this should unmount the PD
-  ${KUBECFG} delete pods/testpd
+  ${KUBECTL} delete pods testpd
   for i in $(seq 1 30); do
     echo "Waiting for pod to be deleted."
     sleep 5
     all_running=0
     for id in $pod_id_list; do
-      current_status=$($KUBECFG -template '{{.currentState.status}}' get pods/$id) || true
+      current_status=$(${KUBECTL} get pods $id -o template '--template={{.currentState.status}}') || true
       if [[ "$current_status" == "Running" ]]; then
         all_running=1
         break
@@ -99,17 +99,17 @@ perl -p -e "s/%.*%/${disk_name}/g" ${KUBE_ROOT}/examples/gce-pd/testpd.yaml > ${
 # Create and format the disk.
 "${GCLOUD}" compute disks create --zone="${ZONE}" --size=10GB "${disk_name}"
 "${GCLOUD}" compute instances attach-disk --zone="${ZONE}" --disk="${disk_name}" \
-  --device-name temp-data "${MASTER_NAME}"
+  --device-name tempdata "${MASTER_NAME}"
 "${GCLOUD}" compute ssh --zone="${ZONE}" "${MASTER_NAME}" --command "sudo rm -rf /mnt/tmp"
 "${GCLOUD}" compute ssh --zone="${ZONE}" "${MASTER_NAME}" --command "sudo mkdir -p /mnt/tmp"
-"${GCLOUD}" compute ssh --zone="${ZONE}" "${MASTER_NAME}" --command "sudo /usr/share/google/safe_format_and_mount /dev/disk/by-id/google-temp-data /mnt/tmp"
+"${GCLOUD}" compute ssh --zone="${ZONE}" "${MASTER_NAME}" --command "sudo /usr/share/google/safe_format_and_mount /dev/disk/by-id/google-tempdata /mnt/tmp"
 "${GCLOUD}" compute ssh --zone="${ZONE}" "${MASTER_NAME}" --command "sudo umount /mnt/tmp"
 "${GCLOUD}" compute instances detach-disk --zone="${ZONE}" --disk "${disk_name}" "${MASTER_NAME}"
 
 # Create a pod that uses the PD
-${KUBECFG} -c ${config} create pods
+${KUBECTL} create -f ${config}
 
-pod_id_list=$($KUBECFG '-template={{range.items}}{{.id}} {{end}}' -l test=testpd list pods)
+pod_id_list=$(${KUBECTL} get pods -o template '--template={{range.items}}{{.id}} {{end}}' -l test=testpd)
 # Pod turn up on a clean cluster can take a while for the docker image
 # pull, and even longer if the PD mount takes a bit.
 all_running=0
@@ -118,7 +118,7 @@ for i in $(seq 1 30); do
   sleep 5
   all_running=1
   for id in $pod_id_list; do
-    current_status=$($KUBECFG -template '{{.currentState.status}}' get pods/$id) || true
+    current_status=$(${KUBECTL} get pods $id -o template '--template={{.currentState.status}}') || true
     if [[ "$current_status" != "Running" ]]; then
       all_running=0
       break
@@ -141,9 +141,9 @@ sleep 20
 
 
 # Recreate the pod, this should re-mount the PD
-${KUBECFG} -c ${config} create pods
+${KUBECTL} create -f ${config}
 
-pod_id_list=$($KUBECFG '-template={{range.items}}{{.id}} {{end}}' -l test=testpd list pods)
+pod_id_list=$(${KUBECTL} get pods -o template '--template={{range.items}}{{.id}} {{end}}' -l test=testpd)
 # Pod turn up on a clean cluster can take a while for the docker image pull.
 all_running=0
 for i in $(seq 1 30); do
@@ -151,7 +151,7 @@ for i in $(seq 1 30); do
   sleep 5
   all_running=1
   for id in $pod_id_list; do
-    current_status=$($KUBECFG -template '{{.currentState.status}}' get pods/$id) || true
+    current_status=$(${KUBECTL} get pods $id -o template '--template={{.currentState.status}}') || true
     if [[ "$current_status" != "Running" ]]; then
       all_running=0
       break
diff --git a/hack/e2e-suite/services.sh b/hack/e2e-suite/services.sh
index ba39a37daed..ed8926e51d3 100755
--- a/hack/e2e-suite/services.sh
+++ b/hack/e2e-suite/services.sh
@@ -135,9 +135,8 @@ __EOF__
 #   $1: service name
 function stop_service() {
   echo "Stopping service '$1'"
-  ${KUBECFG} stop "$1" || true
-  ${KUBECFG} delete "/replicationControllers/$1" || true
-  ${KUBECFG} delete "/services/$1" || true
+  ${KUBECTL} stop rc "$1" || true
+  ${KUBECTL} delete services "$1" || true
 }
 
 # Args:
@@ -309,9 +308,9 @@ svc1_pods=$(query_pods "${svc1_name}" "${svc1_count}")
 svc2_pods=$(query_pods "${svc2_name}" "${svc2_count}")
 
 # Get the portal IPs.
-svc1_ip=$(${KUBECFG} -template '{{.portalIP}}' get "services/${svc1_name}")
+svc1_ip=$(${KUBECTL} get services -o template '--template={{.portalIP}}' "${svc1_name}")
 test -n "${svc1_ip}" || error "Service1 IP is blank"
-svc2_ip=$(${KUBECFG} -template '{{.portalIP}}' get "services/${svc2_name}")
+svc2_ip=$(${KUBECTL} get services -o template '--template={{.portalIP}}' "${svc2_name}")
 test -n "${svc2_ip}" || error "Service2 IP is blank"
 if [[ "${svc1_ip}" == "${svc2_ip}" ]]; then
   error "Portal IPs conflict: ${svc1_ip}"
@@ -381,7 +380,7 @@ wait_for_pods "${svc3_name}" "${svc3_count}"
 svc3_pods=$(query_pods "${svc3_name}" "${svc3_count}")
 
 # Get the portal IP.
-svc3_ip=$(${KUBECFG} -template '{{.portalIP}}' get "services/${svc3_name}")
+svc3_ip=$(${KUBECTL} get services -o template '--template={{.portalIP}}' "${svc3_name}")
 test -n "${svc3_ip}" || error "Service3 IP is blank"
 
 echo "Verifying the portals from the host"
@@ -437,7 +436,7 @@ wait_for_pods "${svc4_name}" "${svc4_count}"
 svc4_pods=$(query_pods "${svc4_name}" "${svc4_count}")
 
 # Get the portal IP.
-svc4_ip=$(${KUBECFG} -template '{{.portalIP}}' get "services/${svc4_name}")
+svc4_ip=$(${KUBECTL} get services -o template '--template={{.portalIP}}' "${svc4_name}")
 test -n "${svc4_ip}" || error "Service4 IP is blank"
 if [[ "${svc4_ip}" == "${svc2_ip}" || "${svc4_ip}" == "${svc3_ip}" ]]; then
   error "Portal IPs conflict: ${svc4_ip}"
diff --git a/hack/e2e-suite/update.sh b/hack/e2e-suite/update.sh
index 39eb69e3774..6ddc1ab32e8 100755
--- a/hack/e2e-suite/update.sh
+++ b/hack/e2e-suite/update.sh
@@ -53,7 +53,7 @@ function validate() {
     for id in "${pod_id_list[@]+${pod_id_list[@]}}"; do
       local template_string current_status current_image host_ip
 
-      # NB: kubectl & kubecfg add the "exists" function to the standard template functions.
+      # NB: kubectl adds the "exists" function to the standard template functions.
       # This lets us check to see if the "running" entry exists for each of the containers
       # we care about. Exists will never return an error and it's safe to check a chain of
       # things, any one of which may not exist. In the below template, all of info,
diff --git a/hack/e2e.go b/hack/e2e.go
index 6f2b6eced69..b5d8da8a8d3 100644
--- a/hack/e2e.go
+++ b/hack/e2e.go
@@ -57,7 +57,6 @@ var (
 		"You can explicitly set to false if you're, e.g., testing client changes "+
 		"for which the server version doesn't make a difference.")
 
-	cfgCmd = flag.String("cfg", "", "If nonempty, pass this as an argument, and call kubecfg. Implies -v.")
 	ctlCmd = flag.String("ctl", "", "If nonempty, pass this as an argument, and call kubectl. Implies -v. (-test, -cfg, -ctl are mutually exclusive)")
 )
 
@@ -164,8 +163,6 @@ func main() {
 
 	failure := false
 	switch {
-	case *cfgCmd != "":
-		failure = !runBash("'kubecfg "+*cfgCmd+"'", "$KUBECFG "+*cfgCmd)
 	case *ctlCmd != "":
 		failure = !runBash("'kubectl "+*ctlCmd+"'", "$KUBECTL "+*ctlCmd)
 	case *tests != "":
@@ -536,15 +533,6 @@ func printPrefixedLines(prefix, s string) {
 	}
 }
 
-// returns either "", or a list of args intended for appending with the
-// kubecfg or kubectl commands (begining with a space).
-func kubecfgArgs() string {
-	if *checkVersionSkew {
-		return " -expect_version_match"
-	}
-	return ""
-}
-
 // returns either "", or a list of args intended for appending with the
 // kubectl command (begining with a space).
 func kubectlArgs() string {
@@ -564,7 +552,6 @@ export KUBE_CONFIG_FILE="config-test.sh"
 
 # TODO(jbeda): This will break on usage if there is a space in
 # ${KUBE_ROOT}.  Convert to an array?  Or an exported function?
-export KUBECFG="` + versionRoot + `/cluster/kubecfg.sh` + kubecfgArgs() + `"
 export KUBECTL="` + versionRoot + `/cluster/kubectl.sh` + kubectlArgs() + `"
 
 source "` + *root + `/cluster/kube-env.sh"