Merge pull request #16841 from jiangyaoguo/keep-kubeproxy-hostname-consistent-with-kubelet

Auto commit by PR queue bot

cluster/aws/coreos/node.yaml

@@ -93,7 +93,8 @@ coreos:
         --master=https://${MASTER_IP} \
         --kubeconfig=/var/lib/kube-proxy/kubeconfig \
         --v=2 \
-        --logtostderr=true
+        --logtostderr=true \
+        --hostname-override=${HOSTNAME_OVERRIDE}
         Restart=always
         RestartSec=10
 

cluster/centos/node/scripts/kubelet.sh

@@ -48,9 +48,9 @@ EOF
 
 KUBE_PROXY_OPTS="   \${KUBE_LOGTOSTDERR}     \\
                     \${KUBE_LOG_LEVEL}       \\
-                    \${NODE_ADDRESS}       \\
+                    \${NODE_ADDRESS}         \\
-                    \${NODE_PORT}          \\
+                    \${NODE_PORT}            \\
-                    \${NODE_HOSTNAME}      \\
+                    \${NODE_HOSTNAME}        \\
                     \${KUBELET_API_SERVER}   \\
                     \${KUBE_ALLOW_PRIV}      \\
                     \${KUBELET_ARGS}"

cluster/centos/node/scripts/proxy.sh

@@ -16,6 +16,7 @@
 
 
 MASTER_ADDRESS=${1:-"8.8.8.18"}
+NODE_ADDRESS=${2:-"8.8.8.20"}
 
 cat <<EOF >/opt/kubernetes/cfg/kube-proxy
 # --logtostderr=true: log to standard error instead of files
@@ -24,12 +25,16 @@ KUBE_LOGTOSTDERR="--logtostderr=true"
 #  --v=0: log level for V logs
 KUBE_LOG_LEVEL="--v=4"
 
+# --hostname-override="": If non-empty, will use this string as identification instead of the actual hostname.
+NODE_HOSTNAME="--hostname-override=${NODE_ADDRESS}"
+
 # --master="": The address of the Kubernetes API server (overrides any value in kubeconfig)
 KUBE_MASTER="--master=http://${MASTER_ADDRESS}:8080"
 EOF
 
 KUBE_PROXY_OPTS="   \${KUBE_LOGTOSTDERR} \\
                     \${KUBE_LOG_LEVEL}   \\
+                    \${NODE_HOSTNAME}    \\
                     \${KUBE_MASTER}"
 
 cat <<EOF >/usr/lib/systemd/system/kube-proxy.service

cluster/juju/charms/trusty/kubernetes/files/proxy.upstart.tmpl

@@ -9,4 +9,5 @@ kill timeout 60 # wait 60s between SIGTERM and SIGKILL.
 
 exec /usr/local/bin/proxy \
      --master=%(kubeapi_server)s \
-     --logtostderr=true
+     --logtostderr=true \
+     --hostname-override=%(kubelet_bind_addr)s

cluster/libvirt-coreos/user_data_minion.yml

@@ -37,7 +37,8 @@ coreos:
 
         [Service]
         ExecStart=/opt/kubernetes/bin/kube-proxy \
-        --master=http://${MASTER_IP}:8080
+        --master=http://${MASTER_IP}:8080 \
+        --hostname-override=${NODE_IPS[$i]}
         Restart=always
         RestartSec=2
 

cluster/rackspace/cloud-config/node-cloud-config.yaml

@@ -179,6 +179,7 @@ coreos:
         --bind-address=$private_ipv4 \
         --kubeconfig=/var/lib/kube-proxy/kubeconfig \
         --logtostderr=true \
+        --hostname-override=$private_ipv4 \
         --master=${FIRST_APISERVER_URL}
         Restart=always
         RestartSec=5

cluster/ubuntu/util.sh

@@ -34,12 +34,12 @@ function test-build-release() {
 # From user input set the necessary k8s and etcd configuration information
 function setClusterInfo() {
   # Initialize NODE_IPS in setClusterInfo function
-  # NODE_IPS is defined as a global variable, and is concatenated with other nodeIP	
+  # NODE_IPS is defined as a global variable, and is concatenated with other nodeIP
   # When setClusterInfo is called for many times, this could cause potential problems
   # Such as, you will have NODE_IPS=192.168.0.2,192.168.0.3,192.168.0.2,192.168.0.3,
   # which is obviously wrong.
   NODE_IPS=""
-  
+
   local ii=0
   for i in $nodes; do
     nodeIP=${i#*@}
@@ -246,7 +246,8 @@ EOF
 function create-kube-proxy-opts() {
   cat <<EOF > ~/kube/default/kube-proxy
 KUBE_PROXY_OPTS="\
- --master=http://${1}:8080 \
+ --hostname-override=${1} \
+ --master=http://${2}:8080 \
  --logtostderr=true"
 EOF
 
@@ -267,7 +268,7 @@ EOF
 function detect-master() {
   source "${KUBE_ROOT}/cluster/ubuntu/${KUBE_CONFIG_FILE:-config-default.sh}"
   setClusterInfo
-  export KUBE_MASTER="${MASTER}" 
+  export KUBE_MASTER="${MASTER}"
   export KUBE_MASTER_IP="${MASTER_IP}"
   echo "Using master ${MASTER_IP}"
 }
@@ -348,7 +349,7 @@ function kube-up() {
 }
 
 function provision-master() {
-  
+
   echo -e "\nDeploying master on machine ${MASTER_IP}"
 
   ssh $SSH_OPTS "$MASTER" "mkdir -p ~/kube/default"
@@ -371,7 +372,7 @@ function provision-master() {
     DNS:kubernetes.default.svc
     DNS:kubernetes.default.svc.cluster.local
   )
-  
+
   EXTRA_SANS=$(echo "${EXTRA_SANS[@]}" | tr ' ' ,)
 
   # remote login to MASTER and configue k8s master
@@ -388,10 +389,10 @@ function provision-master() {
     create-kube-scheduler-opts
     create-flanneld-opts '127.0.0.1'
     sudo -E -p '[sudo] password to start master: ' -- /bin/bash -c '
-      cp ~/kube/default/* /etc/default/ 
+      cp ~/kube/default/* /etc/default/
-      cp ~/kube/init_conf/* /etc/init/ 
+      cp ~/kube/init_conf/* /etc/init/
       cp ~/kube/init_scripts/* /etc/init.d/
-      
+
       groupadd -f -r kube-cert
       ${PROXY_SETTING} ~/kube/make-ca-cert.sh \"${MASTER_IP}\" \"${EXTRA_SANS}\"
       mkdir -p /opt/bin/
@@ -402,10 +403,10 @@ function provision-master() {
       echo "Deploying master on machine ${MASTER_IP} failed"
       exit 1
     }
-} 
+}
 
 function provision-node() {
-  
+
   echo -e "\nDeploying node on machine ${1#*@}"
 
   ssh $SSH_OPTS $1 "mkdir -p ~/kube/default"
@@ -422,21 +423,23 @@ function provision-node() {
   # remote login to node and configue k8s node
   ssh $SSH_OPTS -t "$1" "
     source ~/kube/util.sh
-    
+
     setClusterInfo
     create-kubelet-opts \
       '${1#*@}' \
       '${MASTER_IP}' \
       '${DNS_SERVER_IP}' \
       '${DNS_DOMAIN}'
-    create-kube-proxy-opts '${MASTER_IP}'
+    create-kube-proxy-opts \
+	  '${1#*@}' \
+	  '${MASTER_IP}'
     create-flanneld-opts '${MASTER_IP}'
-                         
+
     sudo -E -p '[sudo] password to start node: ' -- /bin/bash -c '
       cp ~/kube/default/* /etc/default/
       cp ~/kube/init_conf/* /etc/init/
-      cp ~/kube/init_scripts/* /etc/init.d/ 
+      cp ~/kube/init_scripts/* /etc/init.d/
-      mkdir -p /opt/bin/ 
+      mkdir -p /opt/bin/
       cp ~/kube/minion/* /opt/bin
       service flanneld start
       ~/kube/reconfDocker.sh i
@@ -447,7 +450,7 @@ function provision-node() {
 }
 
 function provision-masterandnode() {
-  
+
   echo -e "\nDeploying master and node on machine ${MASTER_IP}"
 
   ssh $SSH_OPTS $MASTER "mkdir -p ~/kube/default"
@@ -464,7 +467,7 @@ function provision-masterandnode() {
     ubuntu/binaries/master/ \
     ubuntu/binaries/minion \
     "${MASTER}:~/kube"
-  
+
   EXTRA_SANS=(
     IP:${MASTER_IP}
     IP:${SERVICE_CLUSTER_IP_RANGE%.*}.1
@@ -473,13 +476,13 @@ function provision-masterandnode() {
     DNS:kubernetes.default.svc
     DNS:kubernetes.default.svc.cluster.local
   )
-  
+
   EXTRA_SANS=$(echo "${EXTRA_SANS[@]}" | tr ' ' ,)
 
   # remote login to the master/node and configue k8s
   ssh $SSH_OPTS -t "$MASTER" "
     source ~/kube/util.sh
-     
+
     setClusterInfo
     create-etcd-opts '${MASTER_IP}'
     create-kube-apiserver-opts \
@@ -493,17 +496,19 @@ function provision-masterandnode() {
       '${MASTER_IP}' \
       '${DNS_SERVER_IP}' \
       '${DNS_DOMAIN}'
-    create-kube-proxy-opts '${MASTER_IP}'
+    create-kube-proxy-opts \
+	  '${MASTER_IP}' \
+	  '${MASTER_IP}'
     create-flanneld-opts '127.0.0.1'
-    
+
-    sudo -E -p '[sudo] password to start master: ' -- /bin/bash -c ' 
+    sudo -E -p '[sudo] password to start master: ' -- /bin/bash -c '
-      cp ~/kube/default/* /etc/default/ 
+      cp ~/kube/default/* /etc/default/
-      cp ~/kube/init_conf/* /etc/init/ 
+      cp ~/kube/init_conf/* /etc/init/
       cp ~/kube/init_scripts/* /etc/init.d/
-      
+
       groupadd -f -r kube-cert
       ${PROXY_SETTING} ~/kube/make-ca-cert.sh \"${MASTER_IP}\" \"${EXTRA_SANS}\"
-      mkdir -p /opt/bin/ 
+      mkdir -p /opt/bin/
       cp ~/kube/master/* /opt/bin/
       cp ~/kube/minion/* /opt/bin/
 
@@ -512,7 +517,7 @@ function provision-masterandnode() {
       '" || {
       echo "Deploying master and node on machine ${MASTER_IP} failed"
       exit 1
-  }    
+  }
 }
 
 # check whether kubelet has torn down all of the pods
@@ -531,14 +536,14 @@ function check-pods-torn-down() {
 
 # Delete a kubernetes cluster
 function kube-down() {
-  
+
   export KUBECTL_PATH="${KUBE_ROOT}/cluster/ubuntu/binaries/kubectl"
   source "${KUBE_ROOT}/cluster/ubuntu/${KUBE_CONFIG_FILE:-config-default.sh}"
   source "${KUBE_ROOT}/cluster/common.sh"
 
   tear_down_alive_resources
-  check-pods-torn-down 
+  check-pods-torn-down
-  
+
   local ii=0
   for i in ${nodes}; do
       if [[ "${roles[${ii}]}" == "ai" || "${roles[${ii}]}" == "a" ]]; then
@@ -553,7 +558,7 @@ function kube-down() {
               /etc/init/etcd.conf \
               /etc/init.d/etcd \
               /etc/default/etcd
-      
+
             rm -rf /infra*
             rm -rf /srv/kubernetes
             '
@@ -562,20 +567,20 @@ function kube-down() {
         if [[ "${roles[${ii}]}" == "ai" ]]; then
           ssh $SSH_OPTS -t "$i" "sudo rm -rf /var/lib/kubelet"
         fi
-        
+
       elif [[ "${roles[${ii}]}" == "i" ]]; then
         echo "Cleaning on node ${i#*@}"
         ssh $SSH_OPTS -t "$i" "
           pgrep flanneld && \
           sudo -p '[sudo] password to stop node: ' -- /bin/bash -c '
             service flanneld stop
-            rm -rf /var/lib/kubelet            
+            rm -rf /var/lib/kubelet
             '
           " || echo "Cleaning on node ${i#*@} failed"
       else
         echo "unsupported role for ${i}"
       fi
-      
+
       ssh $SSH_OPTS -t "$i" "sudo -- /bin/bash -c '
         rm -f \
           /opt/bin/kube* \
@@ -586,7 +591,7 @@ function kube-down() {
           /etc/init.d/flanneld \
           /etc/default/kube* \
           /etc/default/flanneld
-        
+
         rm -rf ~/kube
         rm -f /run/flannel/subnet.env
       '" || echo "cleaning legacy files on ${i#*@} failed"
@@ -599,16 +604,16 @@ function kube-down() {
 function prepare-push() {
   # Use local binaries for kube-push
   if [[ -z "${KUBE_VERSION}" ]]; then
-    echo "Use local binaries for kube-push" 
+    echo "Use local binaries for kube-push"
     if [[ ! -d "${KUBE_ROOT}/cluster/ubuntu/binaries" ]]; then
       echo "No local binaries.Please check"
       exit 1
-    else 
+    else
       echo "Please make sure all the required local binaries are prepared ahead"
       sleep 3
     fi
   else
-    # Run download-release.sh to get the required release 
+    # Run download-release.sh to get the required release
     export KUBE_VERSION
     "${KUBE_ROOT}/cluster/ubuntu/download-release.sh"
   fi
@@ -617,13 +622,13 @@ function prepare-push() {
 # Update a kubernetes master with expected release
 function push-master() {
   source "${KUBE_ROOT}/cluster/ubuntu/${KUBE_CONFIG_FILE:-config-default.sh}"
-  
+
   if [[ ! -f "${KUBE_ROOT}/cluster/ubuntu/binaries/master/kube-apiserver" ]]; then
     echo "There is no required release of kubernetes, please check first"
     exit 1
   fi
   export KUBECTL_PATH="${KUBE_ROOT}/cluster/ubuntu/binaries/kubectl"
-  
+
   setClusterInfo
 
   local ii=0
@@ -651,8 +656,8 @@ function push-master() {
         rm -f /run/flannel/subnet.env
         rm -rf ~/kube
       '" || echo "Cleaning master ${i#*@} failed"
-    fi 
+    fi
-    
+
     if [[ "${roles[${ii}]}" == "a" ]]; then
       provision-master
     elif [[ "${roles[${ii}]}" == "ai" ]]; then
@@ -679,9 +684,9 @@ function push-node() {
   fi
 
   export KUBECTL_PATH="${KUBE_ROOT}/cluster/ubuntu/binaries/kubectl"
-  
+
   setClusterInfo
-  
+
   local node_ip=${1}
   local ii=0
   local existing=false
@@ -726,12 +731,12 @@ function push-node() {
     echo "node ${node_ip} does not exist"
   else
     verify-cluster
-  fi 
+  fi
-  
+
 }
 
 # Update a kubernetes cluster with expected source
-function kube-push() { 
+function kube-push() {
   prepare-push
   source "${KUBE_ROOT}/cluster/ubuntu/${KUBE_CONFIG_FILE:-config-default.sh}"
 
@@ -739,9 +744,9 @@ function kube-push() {
     echo "There is no required release of kubernetes, please check first"
     exit 1
   fi
-  
+
   export KUBECTL_PATH="${KUBE_ROOT}/cluster/ubuntu/binaries/kubectl"
-  #stop all the kube's process & etcd 
+  #stop all the kube's process & etcd
   local ii=0
   for i in ${nodes}; do
      if [[ "${roles[${ii}]}" == "ai" || "${roles[${ii}]}" == "a" ]]; then

hack/local-up-cluster.sh

@@ -309,6 +309,7 @@ function start_kubeproxy {
     PROXY_LOG=/tmp/kube-proxy.log
     sudo -E "${GO_OUT}/kube-proxy" \
       --v=${LOG_LEVEL} \
+      --hostname-override="127.0.0.1" \
       --master="http://${API_HOST}:${API_PORT}" >"${PROXY_LOG}" 2>&1 &
     PROXY_PID=$!