Versions Compared

Old Version 1

changes.mady.by.user Abhijit Dasgupta

Saved on

compared with

New Version 2

changes.mady.by.user Abhijit Dasgupta

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

ELIOT Deployments in Virtual Environments Overview

This document provides a general approach to set up ELIOT ecosystem in virtual environments. The virtual environment taken are X86_64 , ARM64 and ARM32 servers. 

X86 Servers with CentOS - 7 and ARM - 64 with Ubuntu 17.10 versions are chosen for setup.  

ELIOT Deployment in X86_64/AMD64 Server with CentOS - 7.5 version.

Pre-Installation steps to be executed on ELIOT Manager and ELIOT Node

Disable SELinux:

Panel

      $ setenforce 0

      $ sed -i --follow-symlinks 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/sysconfig/selinux

Disable swap:

Panel

      $ swapoff -a

      Note : To make sure on reboot of server the swap isn't re-enabled, comment the swap line UUID in /etc/fstab file.

      $ vi /etc/fstab

       # /dev/mapper/centos-swap swap swap defaults 0 0

Enable br_netfilter:

Panel
               $ modprobe br_netfilter
               $ echo '1' > /proc/sys/net/bridge/bridge-nf-call-iptables



Install Docker:

Set up the repository
Install the required packages.

Panel
              $ sudo yum install -y yum-utils \
              device-mapper-persistent-data \
              lvm2     

  Set up stable repository

Panel
             $ sudo yum-config-manager \
             -add-repo \
              https://download.docker.com/linux/centos/docker-ce.repo     

        
  Install Docker CE
  
  To install latest execute step a or else if specific version has to be installed execute step b.

   a) Install the latest version of Docker CE and containerd

Panel
           $ sudo yum install docker-ce docker-ce-cli containerd.io

  
   b) List and sort specific available in your repo.

Panel
            $ yum install docker-ce --showduplicates | sort -r
            $ sudo yum install docker-ce-<VERSION_STRING> docker-ce-cli-<VERSION_STRING> containerd.io
       
            Example :
            $ yum list docker-ce --showduplicates | sort -r
            docker-ce.x86_64  3:18.09.1-3.el7                     docker-ce-stable
            docker-ce.x86_64  3:18.09.0-3.el7                     docker-ce-stable
            docker-ce.x86_64  18.06.1.ce-3.el7                    docker-ce-stable
           
           $ sudo yum install docker-ce-18.09.1 docker-ce-cli-18.09.1 containerd.io       

        
 Start Docker

Panel
          $ sudo systemctl start docker

Install Kubernetes 

Add kubernetes repository in Cent OS system 

Panel
        $ cat <<EOF > /etc/yum.repos.d/kubernetes.repo
        [kubernetes]
        name=Kubernetes
        baseurl=https://packages.cloud.google.com/yum/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

 Install kubernetes packages kubeadm, kubelet and kubectl.

Panel

   $ yum install -y kubelet kubeadm kubectl

 After installation is complete restart the ELIOT Manager and ELIOT Node

Panel
        $ sudo reboot  

Login to the servers and start the services, docker and kubelet

Panel
        $ systemctl start docker && systemctl enable docker
        $ systemctl start kubelet && systemctl enable kubelet

   
Change the cgroup-driver  (Need to make sure docker-ce and kubernetes are using same cgroup)   
  
Check docker cgroup

Panel
          $ docker info | grep -i cgroup

 It will display docker is using 'cgroupfs' as a cgroup-driver
 Run the below command to change the kubernetes cgroup-driver to 'cgroupfs' and Reload the systemd system and restart the kubelet service
   

Panel
    $ sed -i 's/cgroup-driver=systemd/cgroup-driver=cgroupfs/g' /etc/systemd/system/kubelet.service.d/10-kubeadm.conf
   
    $ systemctl daemon-reload
    $ systemctl restart kubelet   

Kubernetes Cluster Initialization

          Login to ELIOT Manager Server ; Initialize the Kubernetes Master
       

Panel
        $ kubeadm init --apiserver-advertise-address=<ELIOT Manager Server IP Address> --pod-network-cidr=10.244.0.0/16

      

Note

Note :

--apiserver-advertise-address = determines which IP Address Kubernetes should advertise its API server on.
--pod-network-cidr=10.244.0.0/16 - Pod network range.  Pod network must not overlap with any of the host networks as this can cause issues.
 If you find a collision between your network plugin?s preferred Pod network and some of your host networks,  you should think of a suitable CIDR replacement and use that during kubeadm init with --pod-network-cidr and as a  replacement in your network plugin?s YAML.

     
          To start using your cluster, you need to run (as a regular user)                 

Panel
             $ mkdir -p $HOME/.kube
             $ sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
             $ sudo chown $(id -u):$(id -g) $HOME/.kube/config         
                            or            
            (for root user)
             $ export KUBECONFIG=/etc/kubernetes/admin.conf

Adding ELIOT Node to the Cluster             

       Execute the command in the ELIOT Node 

Panel
        $ kubeadm join --token <token> <master-ip>:6443 --discovery-token-ca-cert-hash sha256:<hash>
       
       [The Kuberadm join command string is displayed after successfull installation of Kubernetes Master (ELIOT Manager)]

Deploy Flannel network in ELIOT Manager (Kubernetes Cluster)

Panel
        $ kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/a70459be0084506e4ec919aa1c114638878db11b/Documentation/kube-flannel.yml

        After join command is executed successfully , you could check the cluster node and pods

Panel
        $ kubectl get nodes
        $ kubectl get pods --all-namespaces


Note
        Note : You will get the 'k8s-master' node is running as a 'master' cluster with status 'ready', and you will get all pods that are needed for the cluster, including the 'kube-flannel-ds' for network pod configuration
        Make sure all kube-system pod status is running.