mirror of https://github.com/k3s-io/k3s
693 lines
28 KiB
Bash
693 lines
28 KiB
Bash
#!/bin/bash
|
|
|
|
# Copyright 2015 The Kubernetes Authors All rights reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
# This script contains functions for configuring instances to run kubernetes
|
|
# master and nodes. It is uploaded as GCE instance metadata. The upstart jobs
|
|
# in cluster/gce/trusty/<node.yaml, master.yaml> download it and make use
|
|
# of needed functions. The script itself is not supposed to be executed in
|
|
# other manners.
|
|
|
|
config_hostname() {
|
|
# Set the hostname to the short version.
|
|
short_hostname=$(hostname -s)
|
|
hostname $short_hostname
|
|
}
|
|
|
|
config_ip_firewall() {
|
|
# We have seen that GCE image may have strict host firewall rules which drop
|
|
# most inbound/forwarded packets. In such a case, add rules to accept all
|
|
# TCP/UDP packets.
|
|
if iptables -L INPUT | grep "Chain INPUT (policy DROP)" > /dev/null; then
|
|
echo "Add rules to accpet all inbound TCP/UDP packets"
|
|
iptables -A INPUT -w -p TCP -j ACCEPT
|
|
iptables -A INPUT -w -p UDP -j ACCEPT
|
|
fi
|
|
if iptables -L FORWARD | grep "Chain FORWARD (policy DROP)" > /dev/null; then
|
|
echo "Add rules to accpet all forwarded TCP/UDP packets"
|
|
iptables -A FORWARD -w -p TCP -j ACCEPT
|
|
iptables -A FORWARD -w -p UDP -j ACCEPT
|
|
fi
|
|
}
|
|
|
|
create_dirs() {
|
|
# Create required directories.
|
|
mkdir -p /var/lib/kubelet
|
|
mkdir -p /etc/kubernetes/manifests
|
|
if [ "${KUBERNETES_MASTER:-}" = "false" ]; then
|
|
mkdir -p /var/lib/kube-proxy
|
|
fi
|
|
}
|
|
|
|
download_kube_env() {
|
|
# Fetch kube-env from GCE metadata server.
|
|
readonly tmp_install_dir="/var/cache/kubernetes-install"
|
|
mkdir -p ${tmp_install_dir}
|
|
curl --fail --silent --show-error \
|
|
-H "X-Google-Metadata-Request: True" \
|
|
-o "${tmp_install_dir}/kube_env.yaml" \
|
|
http://metadata.google.internal/computeMetadata/v1/instance/attributes/kube-env
|
|
# Convert the yaml format file into a shell-style file.
|
|
eval $(python -c '''
|
|
import pipes,sys,yaml
|
|
for k,v in yaml.load(sys.stdin).iteritems():
|
|
print("readonly {var}={value}".format(var = k, value = pipes.quote(str(v))))
|
|
''' < "${tmp_install_dir}/kube_env.yaml" > /etc/kube-env)
|
|
}
|
|
|
|
create_kubelet_kubeconfig() {
|
|
# Create the kubelet kubeconfig file.
|
|
if [ -z "${KUBELET_CA_CERT:-}" ]; then
|
|
KUBELET_CA_CERT="${CA_CERT}"
|
|
fi
|
|
cat > /var/lib/kubelet/kubeconfig << EOF
|
|
apiVersion: v1
|
|
kind: Config
|
|
users:
|
|
- name: kubelet
|
|
user:
|
|
client-certificate-data: ${KUBELET_CERT}
|
|
client-key-data: ${KUBELET_KEY}
|
|
clusters:
|
|
- name: local
|
|
cluster:
|
|
certificate-authority-data: ${KUBELET_CA_CERT}
|
|
contexts:
|
|
- context:
|
|
cluster: local
|
|
user: kubelet
|
|
name: service-account-context
|
|
current-context: service-account-context
|
|
EOF
|
|
}
|
|
|
|
create_kubeproxy_kubeconfig() {
|
|
# Create the kube-proxy config file.
|
|
cat > /var/lib/kube-proxy/kubeconfig << EOF
|
|
apiVersion: v1
|
|
kind: Config
|
|
users:
|
|
- name: kube-proxy
|
|
user:
|
|
token: ${KUBE_PROXY_TOKEN}
|
|
clusters:
|
|
- name: local
|
|
cluster:
|
|
certificate-authority-data: ${CA_CERT}
|
|
contexts:
|
|
- context:
|
|
cluster: local
|
|
user: kube-proxy
|
|
name: service-account-context
|
|
current-context: service-account-context
|
|
EOF
|
|
}
|
|
|
|
# Installs the critical packages that are required by spinning up a cluster.
|
|
install_critical_packages() {
|
|
apt-get update
|
|
# Install docker and brctl if they are not in the image.
|
|
if ! which docker > /dev/null; then
|
|
echo "Do not find docker. Install it."
|
|
# We should install the latest qualified docker, which is version 1.8.3 at present.
|
|
curl -sSL https://get.docker.com/ | DOCKER_VERSION=1.8.3 sh
|
|
fi
|
|
if ! which brctl > /dev/null; then
|
|
echo "Do not find brctl. Install it."
|
|
apt-get install --yes bridge-utils
|
|
fi
|
|
}
|
|
|
|
# Install the packages that are useful but not required by spinning up a cluster.
|
|
install_additional_packages() {
|
|
# Socat and nsenter are not required for spinning up a cluster. We move the
|
|
# installation here to be in parallel with the cluster creation.
|
|
if ! which socat > /dev/null; then
|
|
echo "Do not find socat. Install it."
|
|
apt-get install --yes socat
|
|
fi
|
|
if ! which nsenter > /dev/null; then
|
|
echo "Do not find nsenter. Install it."
|
|
# Note: this is an easy way to install nsenter, but may not be the fastest
|
|
# way. In addition, this may not be a trusted source. So, replace it if
|
|
# we have a better solution.
|
|
docker run --rm -v /usr/local/bin:/target jpetazzo/nsenter
|
|
fi
|
|
}
|
|
|
|
# Retry a download until we get it.
|
|
#
|
|
# $1 is the file to create
|
|
# $2 is the URL to download
|
|
download_or_bust() {
|
|
rm -f $1 > /dev/null
|
|
until curl --ipv4 -Lo "$1" --connect-timeout 20 --retry 6 --retry-delay 10 "$2"; do
|
|
echo "Failed to download file ($2). Retrying."
|
|
done
|
|
}
|
|
|
|
# Downloads kubernetes binaries and kube-system manifest tarball, unpacks them,
|
|
# and places them into suitable directories.
|
|
install_kube_binary_config() {
|
|
# In anyway we have to download the release tarball as docker_tag files and
|
|
# kube-proxy image file are there.
|
|
cd /tmp
|
|
k8s_sha1="${SERVER_BINARY_TAR_URL##*/}.sha1"
|
|
echo "Downloading k8s tar sha1 file ${k8s_sha1}"
|
|
download_or_bust "${k8s_sha1}" "${SERVER_BINARY_TAR_URL}.sha1"
|
|
k8s_tar="${SERVER_BINARY_TAR_URL##*/}"
|
|
echo "Downloading k8s tar file ${k8s_tar}"
|
|
download_or_bust "${k8s_tar}" "${SERVER_BINARY_TAR_URL}"
|
|
# Validate hash.
|
|
actual=$(sha1sum ${k8s_tar} | awk '{ print $1 }') || true
|
|
if [ "${actual}" != "${SERVER_BINARY_TAR_HASH}" ]; then
|
|
echo "== ${k8s_tar} corrupted, sha1 ${actual} doesn't match expected ${SERVER_BINARY_TAR_HASH} =="
|
|
else
|
|
echo "Validated ${SERVER_BINARY_TAR_URL} SHA1 = ${SERVER_BINARY_TAR_HASH}"
|
|
fi
|
|
tar xzf "/tmp/${k8s_tar}" -C /tmp/ --overwrite
|
|
# Copy docker_tag and image files to /run/kube-docker-files.
|
|
mkdir -p /run/kube-docker-files
|
|
cp /tmp/kubernetes/server/bin/*.docker_tag /run/kube-docker-files/
|
|
if [ "${KUBERNETES_MASTER:-}" = "false" ]; then
|
|
cp /tmp/kubernetes/server/bin/kube-proxy.tar /run/kube-docker-files/
|
|
else
|
|
cp /tmp/kubernetes/server/bin/kube-apiserver.tar /run/kube-docker-files/
|
|
cp /tmp/kubernetes/server/bin/kube-controller-manager.tar /run/kube-docker-files/
|
|
cp /tmp/kubernetes/server/bin/kube-scheduler.tar /run/kube-docker-files/
|
|
cp -r /tmp/kubernetes/addons /run/kube-docker-files/
|
|
fi
|
|
# For a testing cluster, we use kubelet, kube-proxy, and kubectl binaries
|
|
# from the release tarball and place them in /usr/local/bin. For a non-test
|
|
# cluster, we use the binaries pre-installed in the image, or pull and place
|
|
# them in /usr/bin if they are not pre-installed.
|
|
BINARY_PATH="/usr/bin/"
|
|
if [ "${TEST_CLUSTER:-}" = "true" ]; then
|
|
BINARY_PATH="/usr/local/bin/"
|
|
fi
|
|
if ! which kubelet > /dev/null || ! which kube-proxy > /dev/null || [ "${TEST_CLUSTER:-}" = "true" ]; then
|
|
cp /tmp/kubernetes/server/bin/kubelet ${BINARY_PATH}
|
|
cp /tmp/kubernetes/server/bin/kubectl ${BINARY_PATH}
|
|
fi
|
|
# Clean up.
|
|
rm -rf "/tmp/kubernetes"
|
|
rm "/tmp/${k8s_tar}"
|
|
rm "/tmp/${k8s_sha1}"
|
|
|
|
# Put kube-system pods manifests in /etc/kube-manifests/.
|
|
mkdir -p /run/kube-manifests
|
|
cd /run/kube-manifests
|
|
manifests_sha1="${KUBE_MANIFESTS_TAR_URL##*/}.sha1"
|
|
echo "Downloading kube-system manifests tar sha1 file ${manifests_sha1}"
|
|
download_or_bust "${manifests_sha1}" "${KUBE_MANIFESTS_TAR_URL}.sha1"
|
|
manifests_tar="${KUBE_MANIFESTS_TAR_URL##*/}"
|
|
echo "Downloading kube-manifest tar file ${manifests_tar}"
|
|
download_or_bust "${manifests_tar}" "${KUBE_MANIFESTS_TAR_URL}"
|
|
# Validate hash.
|
|
actual=$(sha1sum ${manifests_tar} | awk '{ print $1 }') || true
|
|
if [ "${actual}" != "${KUBE_MANIFESTS_TAR_HASH}" ]; then
|
|
echo "== ${manifests_tar} corrupted, sha1 ${actual} doesn't match expected ${KUBE_MANIFESTS_TAR_HASH} =="
|
|
else
|
|
echo "Validated ${KUBE_MANIFESTS_TAR_URL} SHA1 = ${KUBE_MANIFESTS_TAR_HASH}"
|
|
fi
|
|
tar xzf "/run/kube-manifests/${manifests_tar}" -C /run/kube-manifests/ --overwrite
|
|
rm "/run/kube-manifests/${manifests_sha1}"
|
|
rm "/run/kube-manifests/${manifests_tar}"
|
|
}
|
|
|
|
restart_docker_daemon() {
|
|
# Assemble docker deamon options
|
|
DOCKER_OPTS="-p /var/run/docker.pid --bridge=cbr0 --iptables=false --ip-masq=false"
|
|
if [ "${TEST_CLUSTER:-}" = "true" ]; then
|
|
DOCKER_OPTS="${DOCKER_OPTS} --log-level=debug"
|
|
fi
|
|
echo "DOCKER_OPTS=\"${DOCKER_OPTS} ${EXTRA_DOCKER_OPTS:-}\"" > /etc/default/docker
|
|
# Make sure the network interface cbr0 is created before restarting docker daemon
|
|
while ! [ -L /sys/class/net/cbr0 ]; do
|
|
echo "Sleep 1 second to wait for cbr0"
|
|
sleep 1
|
|
done
|
|
initctl restart docker
|
|
# Remove docker0
|
|
ifconfig docker0 down
|
|
brctl delbr docker0
|
|
}
|
|
|
|
# Create the log file and set its properties.
|
|
#
|
|
# $1 is the file to create
|
|
prepare_log_file() {
|
|
touch $1
|
|
chmod 644 $1
|
|
chown root:root $1
|
|
}
|
|
|
|
# It monitors the health of several master and node components.
|
|
health_monitoring() {
|
|
sleep_seconds=10
|
|
max_seconds=10
|
|
# We simply kill the process when there is a failure. Another upstart job will automatically
|
|
# restart the process.
|
|
while [ 1 ]; do
|
|
if ! timeout 10 docker version > /dev/null; then
|
|
echo "Docker daemon failed!"
|
|
pkill docker
|
|
fi
|
|
if ! curl --insecure -m ${max_seconds} -f -s https://127.0.0.1:${KUBELET_PORT:-10250}/healthz > /dev/null; then
|
|
echo "Kubelet is unhealthy!"
|
|
pkill kubelet
|
|
fi
|
|
# TODO(andyzheng0831): Add master side health monitoring.
|
|
sleep ${sleep_seconds}
|
|
done
|
|
}
|
|
|
|
|
|
########## The functions below are for master only ##########
|
|
|
|
# Mounts a persistent disk (formatting if needed) to store the persistent data
|
|
# on the master -- etcd's data, a few settings, and security certs/keys/tokens.
|
|
# safe_format_and_mount only formats an unformatted disk, and mkdir -p will
|
|
# leave a directory be if it already exists.
|
|
mount_master_pd() {
|
|
readonly pd_path="/dev/disk/by-id/google-master-pd"
|
|
readonly mount_point="/mnt/disks/master-pd"
|
|
|
|
# TODO(zmerlynn): GKE is still lagging in master-pd creation
|
|
if [ ! -e ${pd_path} ]; then
|
|
return
|
|
fi
|
|
# Format and mount the disk, create directories on it for all of the master's
|
|
# persistent data, and link them to where they're used.
|
|
mkdir -p ${mount_point}
|
|
/usr/share/google/safe_format_and_mount -m "mkfs.ext4 -F" ${pd_path} ${mount_point} >/var/log/master-pd-mount.log || \
|
|
{ echo "!!! master-pd mount failed, review /var/log/master-pd-mount.log !!!"; return 1; }
|
|
# Contains all the data stored in etcd
|
|
mkdir -m 700 -p "${mount_point}/var/etcd"
|
|
# Contains the dynamically generated apiserver auth certs and keys
|
|
mkdir -p "${mount_point}/etc/srv/kubernetes"
|
|
# Directory for kube-apiserver to store SSH key (if necessary)
|
|
mkdir -p /"${mount_point}/etc/srv/sshproxy"
|
|
ln -s -f "${mount_point}/var/etcd" /var/etcd
|
|
mkdir -p /etc/srv
|
|
ln -s -f /"${mount_point}/etc/srv/kubernetes" /etc/srv/kubernetes
|
|
ln -s -f /"${mount_point}/etc/srv/sshproxy" /etc/srv/sshproxy
|
|
|
|
if ! id etcd &>/dev/null; then
|
|
useradd -s /sbin/nologin -d /var/etcd etcd
|
|
fi
|
|
chown -R etcd /"${mount_point}/var/etcd"
|
|
chgrp -R etcd "${mount_point}/var/etcd"
|
|
}
|
|
|
|
# A helper function that adds an entry to a token file.
|
|
# $1: account information
|
|
# $2: token file
|
|
add_token_entry() {
|
|
current_token=$(dd if=/dev/urandom bs=128 count=1 2>/dev/null | base64 | tr -d "=+/" | dd bs=32 count=1 2>/dev/null)
|
|
echo "${tcurrent_token},$1,$1" >> $2
|
|
}
|
|
|
|
# After the first boot and on upgrade, these files exists on the master-pd
|
|
# and should never be touched again (except perhaps an additional service
|
|
# account, see NB below.)
|
|
create_master_auth() {
|
|
readonly auth_dir="/etc/srv/kubernetes"
|
|
if [ ! -e "${auth_dir}/ca.crt" ]; then
|
|
if [ ! -z "${CA_CERT:-}" ] && [ ! -z "${MASTER_CERT:-}" ] && [ ! -z "${MASTER_KEY:-}" ]; then
|
|
echo "${CA_CERT}" | base64 -d > "${auth_dir}/ca.crt"
|
|
echo "${MASTER_CERT}" | base64 -d > "${auth_dir}/server.cert"
|
|
echo "${MASTER_KEY}" | base64 -d > "${auth_dir}/server.key"
|
|
# Kubecfg cert/key are optional and included for backwards compatibility.
|
|
# TODO(roberthbailey): Remove these two lines once GKE no longer requires
|
|
# fetching clients certs from the master VM.
|
|
echo "${KUBECFG_CERT:-}" | base64 -d > "${auth_dir}/kubecfg.crt"
|
|
echo "${KUBECFG_KEY:-}" | base64 -d > "${auth_dir}/kubecfg.key"
|
|
fi
|
|
fi
|
|
readonly basic_auth_csv="${auth_dir}/basic_auth.csv"
|
|
if [ ! -e "${basic_auth_csv}" ]; then
|
|
echo "${KUBE_PASSWORD},${KUBE_USER},admin" > "${basic_auth_csv}"
|
|
fi
|
|
readonly known_tokens_csv="${auth_dir}/known_tokens.csv"
|
|
if [ ! -e "${known_tokens_csv}" ]; then
|
|
echo "${KUBE_BEARER_TOKEN},admin,admin" > "${known_tokens_csv}"
|
|
echo "${KUBELET_TOKEN},kubelet,kubelet" >> "${known_tokens_csv}"
|
|
echo "${KUBE_PROXY_TOKEN},kube_proxy,kube_proxy" >> "${known_tokens_csv}"
|
|
|
|
# Generate tokens for other "service accounts". Append to known_tokens.
|
|
#
|
|
# NB: If this list ever changes, this script actually has to
|
|
# change to detect the existence of this file, kill any deleted
|
|
# old tokens and add any new tokens (to handle the upgrade case).
|
|
add_token_entry "system:scheduler" ${known_tokens_csv}
|
|
add_token_entry "system:controller_manager" ${known_tokens_csv}
|
|
add_token_entry "system:logging" ${known_tokens_csv}
|
|
add_token_entry "system:monitoring" ${known_tokens_csv}
|
|
add_token_entry "system:dns" ${known_tokens_csv}
|
|
fi
|
|
|
|
if [ -n "${PROJECT_ID:-}" ] && [ -n "${TOKEN_URL:-}" ] && [ -n "${TOKEN_BODY:-}" ] && [ -n "${NODE_NETWORK:-}" ]; then
|
|
cat <<EOF >/etc/gce.conf
|
|
[global]
|
|
token-url = ${TOKEN_URL}
|
|
token-body = ${TOKEN_BODY}
|
|
project-id = ${PROJECT_ID}
|
|
network-name = ${NODE_NETWORK}
|
|
EOF
|
|
fi
|
|
}
|
|
|
|
# Uses KUBELET_CA_CERT (falling back to CA_CERT), KUBELET_CERT, and KUBELET_KEY
|
|
# to generate a kubeconfig file for the kubelet to securely connect to the apiserver.
|
|
create_master_kubelet_auth() {
|
|
# Only configure the kubelet on the master if the required variables are
|
|
# set in the environment.
|
|
if [ -n "${KUBELET_APISERVER:-}" ] && [ -n "${KUBELET_CERT:-}" ] && [ -n "${KUBELET_KEY:-}" ]; then
|
|
create_kubelet_kubeconfig
|
|
fi
|
|
}
|
|
|
|
# Replaces the variables in the etcd manifest file with the real values, and then
|
|
# copy the file to the manifest dir
|
|
# $1: value for variable 'suffix'
|
|
# $2: value for variable 'port'
|
|
# $3: value for variable 'server_port'
|
|
# $4: value for variable 'cpulimit'
|
|
# $5: pod name, which should be either etcd or etcd-events
|
|
prepare_etcd_manifest() {
|
|
readonly etcd_temp_file="/tmp/$5"
|
|
cp /run/kube-manifests/kubernetes/trusty/etcd.manifest "${etcd_temp_file}"
|
|
sed -i -e "s@{{ *suffix *}}@$1@g" "${etcd_temp_file}"
|
|
sed -i -e "s@{{ *port *}}@$2@g" "${etcd_temp_file}"
|
|
sed -i -e "s@{{ *server_port *}}@$3@g" "${etcd_temp_file}"
|
|
sed -i -e "s@{{ *cpulimit *}}@\"$4\"@g" "${etcd_temp_file}"
|
|
# Replace the volume host path
|
|
sed -i -e "s@/mnt/master-pd/var/etcd@/mnt/disks/master-pd/var/etcd@g" "${etcd_temp_file}"
|
|
mv "${etcd_temp_file}" /etc/kubernetes/manifests
|
|
}
|
|
|
|
# Starts etcd server pod (and etcd-events pod if needed).
|
|
# More specifically, it prepares dirs and files, sets the variable value
|
|
# in the manifests, and copies them to /etc/kubernetes/manifests.
|
|
start_etcd_servers() {
|
|
if [ -d /etc/etcd ]; then
|
|
rm -rf /etc/etcd
|
|
fi
|
|
if [ -e /etc/default/etcd ]; then
|
|
rm -f /etc/default/etcd
|
|
fi
|
|
if [ -e /etc/systemd/system/etcd.service ]; then
|
|
rm -f /etc/systemd/system/etcd.service
|
|
fi
|
|
if [ -e /etc/init.d/etcd ]; then
|
|
rm -f /etc/init.d/etcd
|
|
fi
|
|
prepare_log_file /var/log/etcd.log
|
|
prepare_etcd_manifest "" "4001" "2380" "200m" "etcd.manifest"
|
|
|
|
# Switch on the second etcd instance if there are more than 50 nodes.
|
|
if [ -n "${NUM_NODES:-}" ] && [ "${NUM_NODES}" -gt 50 ]; then
|
|
prepare_log_file /var/log/etcd-events.log
|
|
prepare_etcd_manifest "-events" "4002" "2381" "100m" "etcd-events.manifest"
|
|
fi
|
|
}
|
|
|
|
# Calculates the following variables based on env variables, which will be used
|
|
# by the manifests of several kube-master components.
|
|
# CLOUD_CONFIG_VOLUME
|
|
# CLOUD_CONFIG_MOUNT
|
|
# DOCKER_REGISTRY
|
|
compute_master_manifest_variables() {
|
|
CLOUD_CONFIG_VOLUME=""
|
|
CLOUD_CONFIG_MOUNT=""
|
|
if [ -n "${PROJECT_ID:-}" ] && [ -n "${TOKEN_URL:-}" ] && [ -n "${TOKEN_BODY:-}" ] && [ -n "${NODE_NETWORK:-}" ]; then
|
|
CLOUD_CONFIG_VOLUME="{\"name\": \"cloudconfigmount\",\"hostPath\": {\"path\": \"/etc/gce.conf\"}},"
|
|
CLOUD_CONFIG_MOUNT="{\"name\": \"cloudconfigmount\",\"mountPath\": \"/etc/gce.conf\", \"readOnly\": true},"
|
|
fi
|
|
DOCKER_REGISTRY="gcr.io/google_containers"
|
|
if [ -n "${KUBE_DOCKER_REGISTRY:-}" ]; then
|
|
DOCKER_REGISTRY=${KUBE_DOCKER_REGISTRY}
|
|
fi
|
|
}
|
|
|
|
# Starts k8s apiserver.
|
|
# It prepares the log file, loads the docker image, calculates variables, sets them
|
|
# in the manifest file, and then copies the manifest file to /etc/kubernetes/manifests.
|
|
#
|
|
# Assumed vars (which are calculated in function compute_master_manifest_variables)
|
|
# CLOUD_CONFIG_VOLUME
|
|
# CLOUD_CONFIG_MOUNT
|
|
# DOCKER_REGISTRY
|
|
start_kube_apiserver() {
|
|
prepare_log_file /var/log/kube-apiserver.log
|
|
# Load the docker image from file.
|
|
echo "Try to load docker image file kube-apiserver.tar"
|
|
timeout 30 docker load -i /run/kube-docker-files/kube-apiserver.tar
|
|
|
|
# Calculate variables and assemble the command line.
|
|
params="--cloud-provider=gce --address=127.0.0.1 --etcd-servers=http://127.0.0.1:4001 --tls-cert-file=/etc/srv/kubernetes/server.cert --tls-private-key-file=/etc/srv/kubernetes/server.key --secure-port=443 --client-ca-file=/etc/srv/kubernetes/ca.crt --token-auth-file=/etc/srv/kubernetes/known_tokens.csv --basic-auth-file=/etc/srv/kubernetes/basic_auth.csv --allow-privileged=true"
|
|
if [ -n "${NUM_NODES:-}" ] && [ "${NUM_NODES}" -gt 50 ]; then
|
|
params="${params} --etcd-servers-overrides=/events#http://127.0.0.1:4002"
|
|
fi
|
|
if [ -n "${SERVICE_CLUSTER_IP_RANGE:-}" ]; then
|
|
params="${params} --service-cluster-ip-range=${SERVICE_CLUSTER_IP_RANGE}"
|
|
fi
|
|
if [ -n "${ADMISSION_CONTROL:-}" ]; then
|
|
params="${params} --admission-control=${ADMISSION_CONTROL}"
|
|
fi
|
|
if [ -n "${KUBE_APISERVER_REQUEST_TIMEOUT:-}" ]; then
|
|
params="${params} --min-request-timeout=${KUBE_APISERVER_REQUEST_TIMEOUT}"
|
|
fi
|
|
if [ -n "${RUNTIME_CONFIG:-}" ]; then
|
|
params="${params} --runtime-config=${RUNTIME_CONFIG}"
|
|
fi
|
|
if [ -n "${APISERVER_TEST_ARGS:-}" ]; then
|
|
params="${params} ${APISERVER_TEST_ARGS}"
|
|
fi
|
|
log_level="--v=2"
|
|
if [ -n "${API_SERVER_TEST_LOG_LEVEL:-}" ]; then
|
|
log_level="${API_SERVER_TEST_LOG_LEVEL}"
|
|
fi
|
|
params="${params} ${log_level}"
|
|
|
|
if [ -n "${PROJECT_ID:-}" ] && [ -n "${TOKEN_URL:-}" ] && [ -n "${TOKEN_BODY:-}" ] && [ -n "${NODE_NETWORK:-}" ]; then
|
|
readonly vm_external_ip=$(curl --fail --silent -H 'Metadata-Flavor: Google' "http://metadata/computeMetadata/v1/instance/network-interfaces/0/access-configs/0/external-ip")
|
|
params="${params} --cloud-config=/etc/gce.conf --advertise-address=${vm_external_ip} --ssh-user=${PROXY_SSH_USER} --ssh-keyfile=/etc/srv/sshproxy/.sshkeyfile"
|
|
fi
|
|
readonly kube_apiserver_docker_tag=$(cat /run/kube-docker-files/kube-apiserver.docker_tag)
|
|
|
|
src_file="/run/kube-manifests/kubernetes/trusty/kube-apiserver.manifest"
|
|
sed -i -e "s@{{params}}@${params}@g" ${src_file}
|
|
sed -i -e "s@{{cloud_config_mount}}@${CLOUD_CONFIG_MOUNT}@g" ${src_file}
|
|
sed -i -e "s@{{cloud_config_volume}}@${CLOUD_CONFIG_VOLUME}@g" ${src_file}
|
|
sed -i -e "s@{{kube_docker_registry}}@${DOCKER_REGISTRY}@g" ${src_file}
|
|
sed -i -e "s@{{kube-apiserver_docker_tag}}@${kube_apiserver_docker_tag}@g" ${src_file}
|
|
cp ${src_file} /etc/kubernetes/manifests
|
|
}
|
|
|
|
# Starts k8s controller manager.
|
|
# It prepares the log file, loads the docker image, calculates variables, sets them
|
|
# in the manifest file, and then copies the manifest file to /etc/kubernetes/manifests.
|
|
#
|
|
# Assumed vars (which are calculated in function compute_master_manifest_variables)
|
|
# CLOUD_CONFIG_VOLUME
|
|
# CLOUD_CONFIG_MOUNT
|
|
# DOCKER_REGISTRY
|
|
start_kube_controller_manager() {
|
|
prepare_log_file /var/log/kube-controller-manager.log
|
|
# Load the docker image from file.
|
|
echo "Try to load docker image file kube-controller-manager.tar"
|
|
timeout 30 docker load -i /run/kube-docker-files/kube-controller-manager.tar
|
|
|
|
# Calculate variables and assemble the command line.
|
|
params="--master=127.0.0.1:8080 --cloud-provider=gce --root-ca-file=/etc/srv/kubernetes/ca.crt --service-account-private-key-file=/etc/srv/kubernetes/server.key"
|
|
if [ -n "${INSTANCE_PREFIX:-}" ]; then
|
|
params="${params} --cluster-name=${INSTANCE_PREFIX}"
|
|
fi
|
|
if [ -n "${CLUSTER_IP_RANGE:-}" ]; then
|
|
params="${params} --cluster-cidr=${CLUSTER_IP_RANGE}"
|
|
fi
|
|
if [ "${ALLOCATE_NODE_CIDRS:-}" = "true" ]; then
|
|
params="${params} --allocate-node-cidrs=${ALLOCATE_NODE_CIDRS}"
|
|
fi
|
|
if [ -n "${TERMINATED_POD_GC_THRESHOLD:-}" ]; then
|
|
params="${params} --terminated-pod-gc-threshold=${TERMINATED_POD_GC_THRESHOLD}"
|
|
fi
|
|
log_level="--v=2"
|
|
if [ -n "${CONTROLLER_MANAGER_TEST_LOG_LEVEL:-}" ]; then
|
|
log_level="${CONTROLLER_MANAGER_TEST_LOG_LEVEL}"
|
|
fi
|
|
params="${params} ${log_level}"
|
|
if [ -n "${CONTROLLER_MANAGER_TEST_ARGS:-}" ]; then
|
|
params="${params} ${CONTROLLER_MANAGER_TEST_ARGS}"
|
|
fi
|
|
readonly kube_rc_docker_tag=$(cat /run/kube-docker-files/kube-controller-manager.docker_tag)
|
|
|
|
src_file="/run/kube-manifests/kubernetes/trusty/kube-controller-manager.manifest"
|
|
sed -i -e "s@{{params}}@${params}@g" ${src_file}
|
|
sed -i -e "s@{{cloud_config_mount}}@${CLOUD_CONFIG_MOUNT}@g" ${src_file}
|
|
sed -i -e "s@{{cloud_config_volume}}@${CLOUD_CONFIG_VOLUME}@g" ${src_file}
|
|
sed -i -e "s@{{kube_docker_registry}}@${DOCKER_REGISTRY}@g" ${src_file}
|
|
sed -i -e "s@{{kube-controller-manager_docker_tag}}@${kube_rc_docker_tag}@g" ${src_file}
|
|
cp ${src_file} /etc/kubernetes/manifests
|
|
}
|
|
|
|
# Start k8s scheduler.
|
|
# It prepares the log file, loads the docker image, calculates variables, sets them
|
|
# in the manifest file, and then copies the manifest file to /etc/kubernetes/manifests.
|
|
#
|
|
# Assumed vars (which are calculated in compute_master_manifest_variables())
|
|
# DOCKER_REGISTRY
|
|
start_kube_scheduler() {
|
|
prepare_log_file /var/log/kube-scheduler.log
|
|
# Load the docker image from file.
|
|
echo "Try to load docker image file kube-scheduler.tar"
|
|
timeout 30 docker load -i /run/kube-docker-files/kube-scheduler.tar
|
|
|
|
# Calculate variables and set them in the manifest.
|
|
params=""
|
|
log_level="--v=2"
|
|
if [ -n "${SCHEDULER_TEST_LOG_LEVEL:-}" ]; then
|
|
log_level="${SCHEDULER_TEST_LOG_LEVEL}"
|
|
fi
|
|
params="${params} ${log_level}"
|
|
if [ -n "${SCHEDULER_TEST_ARGS:-}" ]; then
|
|
params="${params} ${SCHEDULER_TEST_ARGS}"
|
|
fi
|
|
readonly kube_scheduler_docker_tag=$(cat /run/kube-docker-files/kube-scheduler.docker_tag)
|
|
|
|
# Remove salt comments and replace variables with values
|
|
src_file="/run/kube-manifests/kubernetes/trusty/kube-scheduler.manifest"
|
|
sed -i "/^ *{%/d" ${src_file}
|
|
sed -i -e "s@{{params}}@${params}@g" ${src_file}
|
|
sed -i -e "s@{{pillar\['kube_docker_registry'\]}}@${DOCKER_REGISTRY}@g" ${src_file}
|
|
sed -i -e "s@{{pillar\['kube-scheduler_docker_tag'\]}}@${kube_scheduler_docker_tag}@g" ${src_file}
|
|
cp ${src_file} /etc/kubernetes/manifests
|
|
}
|
|
|
|
# Start a fluentd static pod for logging.
|
|
start_fluentd() {
|
|
if [ "${ENABLE_NODE_LOGGING:-}" = "true" ]; then
|
|
if [ "${LOGGING_DESTINATION:-}" = "gcp" ]; then
|
|
cp /run/kube-manifests/kubernetes/fluentd-gcp.yaml /etc/kubernetes/manifests/
|
|
elif [ "${LOGGING_DESTINATION:-}" = "elasticsearch" ]; then
|
|
cp /run/kube-manifests/kubernetes/fluentd-es.yaml /etc/kubernetes/manifests/
|
|
fi
|
|
fi
|
|
}
|
|
|
|
# A helper function for copying addon manifests and set dir/files
|
|
# permissions.
|
|
# $1: addon category under /etc/kubernetes
|
|
# $2: manifest source dir
|
|
setup_addon_manifests() {
|
|
src_dir="/run/kube-manifests/kubernetes/trusty/$2"
|
|
dst_dir="/etc/kubernetes/$1/$2"
|
|
if [ ! -d "${dst_dir}" ]; then
|
|
mkdir -p "${dst_dir}"
|
|
fi
|
|
files=$(find "${src_dir}" -name "*.yaml")
|
|
if [ -n "${files}" ]; then
|
|
cp "${src_dir}/"*.yaml "${dst_dir}"
|
|
fi
|
|
files=$(find "${src_dir}" -name "*.json")
|
|
if [ -n "${files}" ]; then
|
|
cp "${src_dir}/"*.json "${dst_dir}"
|
|
fi
|
|
files=$(find "${src_dir}" -name "*.yaml.in")
|
|
if [ -n "${files}" ]; then
|
|
cp "${src_dir}/"*.yaml.in "${dst_dir}"
|
|
fi
|
|
chown -R root:root "${dst_dir}"
|
|
chmod 755 "${dst_dir}"
|
|
chmod 644 "${dst_dir}"/*
|
|
}
|
|
|
|
# Start k8s addons static pods.
|
|
prepare_kube_addons() {
|
|
addon_src_dir="/run/kube-manifests/kubernetes/trusty"
|
|
addon_dst_dir="/etc/kubernetes/addons"
|
|
# Set up manifests of other addons.
|
|
if [ "${ENABLE_CLUSTER_MONITORING:-}" = "influxdb" ] || \
|
|
[ "${ENABLE_CLUSTER_MONITORING:-}" = "google" ] || \
|
|
[ "${ENABLE_CLUSTER_MONITORING:-}" = "standalone" ] || \
|
|
[ "${ENABLE_CLUSTER_MONITORING:-}" = "googleinfluxdb" ]; then
|
|
file_dir="cluster-monitoring/${ENABLE_CLUSTER_MONITORING}"
|
|
setup_addon_manifests "addons" "${file_dir}"
|
|
# Replace the salt configurations with variable values.
|
|
heapster_memory="200Mi"
|
|
if [ -n "${NUM_NODES:-}" ] && [ "${NUM_NODES}" -gt 1 ]; then
|
|
heapster_memory="$((${NUM_NODES} * 3 + 200))Mi"
|
|
fi
|
|
controller_yaml="${addon_dst_dir}/${file_dir}"
|
|
if [ "${ENABLE_CLUSTER_MONITORING:-}" = "googleinfluxdb" ]; then
|
|
controller_yaml="${controller_yaml}/heapster-controller-combined.yaml"
|
|
else
|
|
controller_yaml="${controller_yaml}/heapster-controller.yaml"
|
|
fi
|
|
sed -i "/^ *{%/d" "${controller_yaml}"
|
|
sed -i -e "s@{{ *heapster_memory *}}@${heapster_memory}@g" "${controller_yaml}"
|
|
fi
|
|
cp "${addon_src_dir}/namespace.yaml" "${addon_dst_dir}"
|
|
if [ "${ENABLE_L7_LOADBALANCING:-}" = "glbc" ]; then
|
|
setup_addon_manifests "addons" "cluster-loadbalancing/glbc"
|
|
fi
|
|
if [ "${ENABLE_CLUSTER_DNS:-}" = "true" ]; then
|
|
setup_addon_manifests "addons" "dns"
|
|
dns_rc_file="${addon_dst_dir}/dns/skydns-rc.yaml"
|
|
dns_svc_file="${addon_dst_dir}/dns/skydns-svc.yaml"
|
|
mv "${addon_dst_dir}/dns/skydns-rc.yaml.in" "${dns_rc_file}"
|
|
mv "${addon_dst_dir}/dns/skydns-svc.yaml.in" "${dns_svc_file}"
|
|
# Replace the salt configurations with variable values.
|
|
sed -i -e "s@{{ *pillar\['dns_replicas'\] *}}@${DNS_REPLICAS}@g" "${dns_rc_file}"
|
|
sed -i -e "s@{{ *pillar\['dns_domain'\] *}}@${DNS_DOMAIN}@g" "${dns_rc_file}"
|
|
sed -i -e "s@{{ *pillar\['dns_server'\] *}}@${DNS_SERVER_IP}@g" "${dns_svc_file}"
|
|
fi
|
|
if [ "${ENABLE_CLUSTER_REGISTRY:-}" = "true" ]; then
|
|
setup_addon_manifests "addons" "registry"
|
|
registry_pv_file="${addon_dst_dir}/registry/registry-pv.yaml"
|
|
registry_pvc_file="${addon_dst_dir}/registry/registry-pvc.yaml"
|
|
mv "${addon_dst_dir}/registry/registry-pv.yaml.in" "${registry_pv_file}"
|
|
mv "${addon_dst_dir}/registry/registry-pvc.yaml.in" "${registry_pvc_file}"
|
|
# Replace the salt configurations with variable values.
|
|
sed -i "/^ *{%/d" "${registry_pv_file}"
|
|
sed -i -e "s@{{ *pillar\['cluster_registry_disk_size'\] *}}@${CLUSTER_REGISTRY_DISK_SIZE}@g" "${registry_pv_file}"
|
|
sed -i -e "s@{{ *pillar\['cluster_registry_disk_size'\] *}}@${CLUSTER_REGISTRY_DISK_SIZE}@g" "${registry_pvc_file}"
|
|
sed -i -e "s@{{ *pillar\['cluster_registry_disk_name'\] *}}@${CLUSTER_REGISTRY_DISK}@g" "${registry_pvc_file}"
|
|
fi
|
|
if [ "${ENABLE_NODE_LOGGING:-}" = "true" ] && \
|
|
[ "${LOGGING_DESTINATION:-}" = "elasticsearch" ] && \
|
|
[ "${ENABLE_CLUSTER_LOGGING:-}" = "true" ]; then
|
|
setup_addon_manifests "addons" "fluentd-elasticsearch"
|
|
fi
|
|
if [ "${ENABLE_CLUSTER_UI:-}" = "true" ]; then
|
|
setup_addon_manifests "addons" "dashboard"
|
|
fi
|
|
if echo "${ADMISSION_CONTROL:-}" | grep -q "LimitRanger"; then
|
|
setup_addon_manifests "admission-controls" "limit-range"
|
|
fi
|
|
|
|
# Prepare the scripts for running addons.
|
|
addon_script_dir="/var/lib/cloud/scripts/kubernetes"
|
|
mkdir -p "${addon_script_dir}"
|
|
cp "${addon_src_dir}/kube-addons.sh" "${addon_script_dir}"
|
|
cp "${addon_src_dir}/kube-addon-update.sh" "${addon_script_dir}"
|
|
chmod 544 "${addon_script_dir}/"*.sh
|
|
# In case that some GCE customized trusty may have a read-only /root.
|
|
mount -t tmpfs tmpfs /root
|
|
mount --bind -o remount,rw,noexec /root
|
|
}
|