diff --git a/build/kube-dns/Changelog b/build/kube-dns/Changelog
index 623db8e00d..6d7afd7462 100644
--- a/build/kube-dns/Changelog
+++ b/build/kube-dns/Changelog
@@ -15,3 +15,6 @@
 
  ## Version 1.7 (Wed August 24 2016 Zihong Zheng <zihongz@google.com>)
  - Add support for ExternalName services (pr #31159)
+
+ ## Version 1.8 (Thu September 29 2016 Zihong Zheng <zihongz@google.com>)
+ - Add support for graceful termination (issue #31807)
diff --git a/build/kube-dns/Makefile b/build/kube-dns/Makefile
index c71c09c8c9..0628ecd30f 100644
--- a/build/kube-dns/Makefile
+++ b/build/kube-dns/Makefile
@@ -22,7 +22,7 @@
 # Default registry, arch and tag. This can be overwritten by arguments to make
 PLATFORM?=linux
 ARCH?=amd64
-TAG?=1.7
+TAG?=1.8
 REGISTRY?=gcr.io/google_containers
 
 GOLANG_VERSION=1.6
diff --git a/cluster/addons/dns/skydns-rc.yaml.base b/cluster/addons/dns/skydns-rc.yaml.base
index 4b5e39c24b..22cdda791c 100644
--- a/cluster/addons/dns/skydns-rc.yaml.base
+++ b/cluster/addons/dns/skydns-rc.yaml.base
@@ -19,29 +19,29 @@
 apiVersion: v1
 kind: ReplicationController
 metadata:
-  name: kube-dns-v19
+  name: kube-dns-v20
   namespace: kube-system
   labels:
     k8s-app: kube-dns
-    version: v19
+    version: v20
     kubernetes.io/cluster-service: "true"
 spec:
   replicas: __PILLAR__DNS__REPLICAS__
   selector:
     k8s-app: kube-dns
-    version: v19
+    version: v20
   template:
     metadata:
       labels:
         k8s-app: kube-dns
-        version: v19
+        version: v20
       annotations:
         scheduler.alpha.kubernetes.io/critical-pod: ''
         scheduler.alpha.kubernetes.io/tolerations: '[{"key":"CriticalAddonsOnly", "operator":"Exists"}]'
     spec:
       containers:
       - name: kubedns
-        image: gcr.io/google_containers/kubedns-amd64:1.7
+        image: gcr.io/google_containers/kubedns-amd64:1.8
         resources:
           # TODO: Set memory limits when we've profiled the container for large
           # clusters, then set request = limit to keep this container in
@@ -83,7 +83,7 @@ spec:
           name: dns-tcp-local
           protocol: TCP
       - name: dnsmasq
-        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.3
+        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.4
         livenessProbe:
           httpGet:
             path: /healthz-dnsmasq
diff --git a/cluster/addons/dns/skydns-rc.yaml.in b/cluster/addons/dns/skydns-rc.yaml.in
index 507b47d886..ee098c4c2e 100644
--- a/cluster/addons/dns/skydns-rc.yaml.in
+++ b/cluster/addons/dns/skydns-rc.yaml.in
@@ -19,30 +19,29 @@
 apiVersion: v1
 kind: ReplicationController
 metadata:
-  name: kube-dns-v19
+  name: kube-dns-v20
   namespace: kube-system
   labels:
     k8s-app: kube-dns
-    version: v19
+    version: v20
     kubernetes.io/cluster-service: "true"
 spec:
   replicas: {{ pillar['dns_replicas'] }}
   selector:
     k8s-app: kube-dns
-    version: v19
+    version: v20
   template:
     metadata:
       labels:
         k8s-app: kube-dns
-        version: v19
-        kubernetes.io/cluster-service: "true"
+        version: v20
       annotations:
         scheduler.alpha.kubernetes.io/critical-pod: ''
         scheduler.alpha.kubernetes.io/tolerations: '[{"key":"CriticalAddonsOnly", "operator":"Exists"}]'
     spec:
       containers:
       - name: kubedns
-        image: gcr.io/google_containers/kubedns-amd64:1.7
+        image: gcr.io/google_containers/kubedns-amd64:1.8
         resources:
           # TODO: Set memory limits when we've profiled the container for large
           # clusters, then set request = limit to keep this container in
@@ -84,7 +83,16 @@ spec:
           name: dns-tcp-local
           protocol: TCP
       - name: dnsmasq
-        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.3
+        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.4
+        livenessProbe:
+          httpGet:
+            path: /healthz-dnsmasq
+            port: 8080
+            scheme: HTTP
+          initialDelaySeconds: 60
+          timeoutSeconds: 5
+          successThreshold: 1
+          failureThreshold: 5
         args:
         - --cache-size=1000
         - --no-resolv
@@ -97,15 +105,6 @@ spec:
         - containerPort: 53
           name: dns-tcp
           protocol: TCP
-        livenessProbe:
-          httpGet:
-            path: /healthz-dnsmasq
-            port: 8080
-            scheme: HTTP
-          initialDelaySeconds: 60
-          timeoutSeconds: 5
-          successThreshold: 1
-          failureThreshold: 5
       - name: healthz
         image: gcr.io/google_containers/exechealthz-amd64:1.2
         resources:
diff --git a/cluster/addons/dns/skydns-rc.yaml.sed b/cluster/addons/dns/skydns-rc.yaml.sed
index cb60972aa4..831282cab3 100644
--- a/cluster/addons/dns/skydns-rc.yaml.sed
+++ b/cluster/addons/dns/skydns-rc.yaml.sed
@@ -19,30 +19,29 @@
 apiVersion: v1
 kind: ReplicationController
 metadata:
-  name: kube-dns-v19
+  name: kube-dns-v20
   namespace: kube-system
   labels:
     k8s-app: kube-dns
-    version: v19
+    version: v20
     kubernetes.io/cluster-service: "true"
 spec:
   replicas: $DNS_REPLICAS
   selector:
     k8s-app: kube-dns
-    version: v19
+    version: v20
   template:
     metadata:
       labels:
         k8s-app: kube-dns
-        version: v19
-        kubernetes.io/cluster-service: "true"
+        version: v20
       annotations:
         scheduler.alpha.kubernetes.io/critical-pod: ''
         scheduler.alpha.kubernetes.io/tolerations: '[{"key":"CriticalAddonsOnly", "operator":"Exists"}]'
     spec:
       containers:
       - name: kubedns
-        image: gcr.io/google_containers/kubedns-amd64:1.7
+        image: gcr.io/google_containers/kubedns-amd64:1.8
         resources:
           # TODO: Set memory limits when we've profiled the container for large
           # clusters, then set request = limit to keep this container in
@@ -83,7 +82,16 @@ spec:
           name: dns-tcp-local
           protocol: TCP
       - name: dnsmasq
-        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.3
+        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.4
+        livenessProbe:
+          httpGet:
+            path: /healthz-dnsmasq
+            port: 8080
+            scheme: HTTP
+          initialDelaySeconds: 60
+          timeoutSeconds: 5
+          successThreshold: 1
+          failureThreshold: 5
         args:
         - --cache-size=1000
         - --no-resolv
@@ -96,15 +104,6 @@ spec:
         - containerPort: 53
           name: dns-tcp
           protocol: TCP
-        livenessProbe:
-          httpGet:
-            path: /healthz-dnsmasq
-            port: 8080
-            scheme: HTTP
-          initialDelaySeconds: 60
-          timeoutSeconds: 5
-          successThreshold: 1
-          failureThreshold: 5
       - name: healthz
         image: gcr.io/google_containers/exechealthz-amd64:1.2
         resources:
diff --git a/cluster/gce/coreos/kube-manifests/addons/dns/skydns-rc.yaml b/cluster/gce/coreos/kube-manifests/addons/dns/skydns-rc.yaml
index 86d338511a..1fcdbedb2c 100644
--- a/cluster/gce/coreos/kube-manifests/addons/dns/skydns-rc.yaml
+++ b/cluster/gce/coreos/kube-manifests/addons/dns/skydns-rc.yaml
@@ -1,29 +1,29 @@
 apiVersion: v1
 kind: ReplicationController
 metadata:
-  name: kube-dns-v19
+  name: kube-dns-v20
   namespace: kube-system
   labels:
     k8s-app: kube-dns
-    version: v19
+    version: v20
     kubernetes.io/cluster-service: "true"
 spec:
   replicas: ${DNS_REPLICAS}
   selector:
     k8s-app: kube-dns
-    version: v19
+    version: v20
   template:
     metadata:
       labels:
         k8s-app: kube-dns
-        version: v19
+        version: v20
       annotations:
         scheduler.alpha.kubernetes.io/critical-pod: ''
         scheduler.alpha.kubernetes.io/tolerations: '[{"key":"CriticalAddonsOnly", "operator":"Exists"}]'
     spec:
       containers:
       - name: kubedns
-        image: gcr.io/google_containers/kubedns-amd64:1.7
+        image: gcr.io/google_containers/kubedns-amd64:1.8
         resources:
           # TODO: Set memory limits when we've profiled the container for large
           # clusters, then set request = limit to keep this container in
@@ -64,7 +64,7 @@ spec:
           name: dns-tcp-local
           protocol: TCP
       - name: dnsmasq
-        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.3
+        image: gcr.io/google_containers/kube-dnsmasq-amd64:1.4
         livenessProbe:
           httpGet:
             path: /healthz-dnsmasq
diff --git a/cluster/juju/layers/kubernetes/templates/kubedns-rc.yaml b/cluster/juju/layers/kubernetes/templates/kubedns-rc.yaml
index 51e743bd2c..ff430f8c27 100644
--- a/cluster/juju/layers/kubernetes/templates/kubedns-rc.yaml
+++ b/cluster/juju/layers/kubernetes/templates/kubedns-rc.yaml
@@ -17,29 +17,29 @@
 apiVersion: v1
 kind: ReplicationController
 metadata:
-  name: kube-dns-v19
+  name: kube-dns-v20
   namespace: kube-system
   labels:
     k8s-app: kube-dns
-    version: v19
+    version: v20
     kubernetes.io/cluster-service: "true"
 spec:
   replicas: {{ pillar['dns_replicas'] }}
   selector:
     k8s-app: kube-dns
-    version: v19
+    version: v20
   template:
     metadata:
       labels:
         k8s-app: kube-dns
-        version: v19
+        version: v20
       annotations:
         scheduler.alpha.kubernetes.io/critical-pod: ''
         scheduler.alpha.kubernetes.io/tolerations: '[{"key":"CriticalAddonsOnly", "operator":"Exists"}]'
     spec:
       containers:
       - name: kubedns
-        image: gcr.io/google_containers/kubedns-{{ arch }}:1.7
+        image: gcr.io/google_containers/kubedns-{{ arch }}:1.8
         resources:
           # TODO: Set memory limits when we've profiled the container for large
           # clusters, then set request = limit to keep this container in
@@ -82,7 +82,7 @@ spec:
           name: dns-tcp-local
           protocol: TCP
       - name: dnsmasq
-        image: gcr.io/google_containers/kube-dnsmasq-{{ arch }}:1.3
+        image: gcr.io/google_containers/kube-dnsmasq-{{ arch }}:1.4
         livenessProbe:
           httpGet:
             path: /healthz-dnsmasq