Skip to content

Latest commit

 

History

History
257 lines (192 loc) · 7.73 KB

pouch_with_kubernetes_deploying.md

File metadata and controls

257 lines (192 loc) · 7.73 KB
Raw

Deploy Kubernetes With Pouch

Updated: 2018.3.30

Overview

This document shows how to easily install a Kubernetes cluster with Pouch as the container runtime.For Chinese customer, it's might be faster and easier to download Kubernetes components and various kinds of container images from Aliyun Kubernetes + Pouch + Aliyun.

pouch_with_kubernetes

Restriction

Kubernetes: Version 1.5+ is recommanded.

Pouch: Version 0.4.0 is recommended.

Install and Configure

An all-in-one kubernetes cluster with pouch runtime could be deployed by running:

hack/kubernetes/allinone.sh

Please refer to allinone .

Install Pouch

You can easily setup a basic Pouch environment, see INSTALLATION.md.

Configure Pouch

On Ubuntu 16.04+:

sed -i 's/ExecStart=\/usr\/bin\/pouchd/ExecStart=\/usr\/bin\/pouchd --enable-cri=true/g' /usr/lib/systemd/system/pouch.service
systemctl daemon-reload
systemctl restart pouch

On CentOS 7:

sed -i 's/ExecStart=\/usr\/local\/bin\/pouchd/ExecStart=\/usr\/local\/bin\/pouchd --enable-cri=true/g' /lib/systemd/system/pouch.service
systemctl daemon-reload
systemctl restart pouch

Install CNI

On Ubuntu 16.04+:

apt-get update && apt-get install -y apt-transport-https
curl -s https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key add -
cat <<EOF > /etc/apt/sources.list.d/kubernetes.list
deb http://apt.kubernetes.io/ kubernetes-xenial main
EOF
apt-get update
apt-get install -y kubernetes-cni

On CentOS 7:

cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=http://yum.kubernetes.io/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg
       https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
EOF
setenforce 0
yum install -y kubernetes-cni

Configure CNI networks:

  • If you want to use CNI plugins like Flannel, Weave, Calico etc, please skip this section.
  • Otherwise, you can use bridge network plugin, it's the simplest way.
    • Subnets should be different on different nodes. e.g. 10.244.1.0/24 for the master node and 10.244.2.0/24 for the first minion node.
mkdir -p /etc/cni/net.d
cat >/etc/cni/net.d/10-mynet.conf <<-EOF
{
    "cniVersion": "0.3.0",
    "name": "mynet",
    "type": "bridge",
    "bridge": "cni0",
    "isGateway": true,
    "ipMasq": true,
    "ipam": {
        "type": "host-local",
        "subnet": "10.244.1.0/24",
        "routes": [
            { "dst": "0.0.0.0/0"  }
        ]
    }
}
EOF
cat >/etc/cni/net.d/99-loopback.conf <<-EOF
{
    "cniVersion": "0.3.0",
    "type": "loopback"
}
EOF

Install Kubernetes Components

On Ubuntu 16.04+:

apt-get install -y kubelet kubeadm kubectl

On CentOS 7:

yum install -y kubelet kubeadm kubectl

Configure kubelet with Pouch as its runtime:

sed -i '2 i\Environment="KUBELET_EXTRA_ARGS=--container-runtime=remote --container-runtime-endpoint=unix:///var/run/pouchcri.sock --image-service-endpoint=unix:///var/run/pouchcri.sock"' /etc/systemd/system/kubelet.service.d/10-kubeadm.conf
systemctl daemon-reload

For more details, please check install kubelet.

Setting up the master node

For more detailed Kubernetes cluster installation, please check Using kubeadm to Create a Cluster

kubeadm init

NOTE: If you want to use CNI plugin other than bridge, please check Installing a pod network.

Optional: enable schedule pods on the master node

export KUBECONFIG=/etc/kubernetes/admin.conf
kubectl taint nodes --all node-role.kubernetes.io/master:NoSchedule-

Setting up the minion nodes

After initializing the master node, you may get the following prompt:

You can now join any number of machines by running the following on each node
as root:

  kubeadm join --token $token ${master_ip:port} --discovery-token-ca-cert-hash $ca-cert

Copy & Run it in all your minion nodes.

Setting up CNI network routes

If your CNI plugin is bridge, you could use direct routes to connect the containers across multi-node.Suppose you have one master node and one minion node:

NODE   IP_ADDRESS   CONTAINER_CIDR
master 10.148.0.1  10.244.1.0/24
minion 10.148.0.2  10.244.2.0/24

Setting up routes:

# master node
ip route add 10.244.2.0/24 via 10.148.0.2

# minion node
ip route add 10.244.1.0/24 via 10.148.0.1

Run and Verify

Create a deployment named Pouch:

# cat pouch.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: pouch
  labels:
    pouch: pouch
spec:
  selector:
    matchLabels:
      pouch: pouch
  template:
    metadata:
      labels:
        pouch: pouch
    spec:
      containers:
      - name: pouch
        image: docker.io/library/busybox:latest

# kubectl create -f pouch.yaml
deployment "pouch" created

Confirm the pod of deployment is really running:

# kubectl get pods -o wide
NAME                     READY     STATUS    RESTARTS   AGE       IP           NODE
pouch-7dcd875d69-gq5r9   1/1       Running   0          44m       10.244.1.4   master
# ping 10.244.1.4
PING 10.244.1.4 (10.244.1.4) 56(84) bytes of data.
64 bytes from 10.244.1.4: icmp_seq=1 ttl=64 time=0.065 ms
64 bytes from 10.244.1.4: icmp_seq=2 ttl=64 time=0.068 ms
64 bytes from 10.244.1.4: icmp_seq=3 ttl=64 time=0.041 ms
64 bytes from 10.244.1.4: icmp_seq=4 ttl=64 time=0.047 ms
^C
--- 10.244.1.4 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3048ms
rtt min/avg/max/mdev = 0.041/0.055/0.068/0.012 ms

Troubleshooting

  • Because kubeadm still assumes docker as the only container runtime which can be used with kubernetes. When you use kubeadm to initialize the master node or join the minion node to the cluster, you may encounter the following error message:[ERROR SystemVerification]: failed to get docker info: Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?. Use the flag --skip-preflight-checks to skip the check, like kubeadm init --skip-preflight-checks.

  • By default Pouch will support CRI v1alpha2,which means that using a version of Kubernetes prior to 1.10 will not work. As the NOTE mentioned above, we could start pouchd with the configuration like pouchd --cri-version v1alpha1 to specify the version of CRI to support the version of Kubernetes below 1.10.

  • By default Pouch will not enable the CRI. If you'd like to deploy Kubernetes with Pouch, you should start pouchd with the configuration like pouchd --enable-cri.

  • By default Pouch will use registry.cn-hangzhou.aliyuncs.com/google-containers/pause-amd64:3.0 as the image of infra container. If you'd like use image other than that, you could start pouchd with the configuration like pouchd --enable-cri --sandbox-image XXX.

  • Any other troubles? Make an issue to connect with us!