/*@jsxRuntime automatic @jsxImportSource react*/
const {Fragment: _Fragment, jsx: _jsx, jsxs: _jsxs} = arguments[0];
function MDXContent(props = {}) {
  const {wrapper: MDXLayout} = props.components || ({});
  return MDXLayout ? _jsx(MDXLayout, Object.assign({}, props, {
    children: _jsx(_createMdxContent, {})
  })) : _createMdxContent();
  function _createMdxContent() {
    const _components = Object.assign({
      p: "p",
      a: "a",
      h2: "h2",
      blockquote: "blockquote",
      h3: "h3",
      pre: "pre",
      code: "code",
      span: "span",
      img: "img"
    }, props.components);
    return _jsxs(_Fragment, {
      children: [_jsxs(_components.p, {
        children: ["We now have ", _jsx(_components.a, {
          href: "/post/2021/5/30/infrastructure-nvidia-nvidia-jetson-cluster",
          children: "a beautiful cluster assembled and running"
        }), ", but no software running on it yet. So let's get started and provision Kubernetes on top of it so we can manage our nodes easier!"]
      }), "\n", _jsx(_components.h2, {
        children: "Introducing K3s"
      }), "\n", _jsxs(_components.p, {
        children: ["For our Kubernetes setup we want a lightweight Kubernetes installation seeing the required Edge computation. For this, we can utilize ", _jsx(_components.a, {
          href: "https://k3s.io/",
          children: "K3S"
        }), " which is a distribution built specifically to run on IoT & Edge devices."]
      }), "\n", _jsxs(_components.blockquote, {
        children: ["\n", _jsx(_components.p, {
          children: "Personally, the installation feels similar to how we installed and configured Docker Swarm before. Install it everywhere and join nodes to the master - quick and easy!"
        }), "\n"]
      }), "\n", _jsx(_components.h2, {
        children: "Installing K3s"
      }), "\n", _jsx(_components.p, {
        children: "Installing K3S is super trivial! There is however an easier way if you want to automate all of it, by utilizing the CLI and adding just 1 parameter."
      }), "\n", _jsx(_components.h3, {
        children: "Master Node Setup"
      }), "\n", _jsx(_components.p, {
        children: "To configure our Master Node, we execute the following command that will pass the secret we define:"
      }), "\n", _jsx(_components.pre, {
        className: "language-bash",
        children: _jsx(_components.code, {
          className: "language-bash code-highlight",
          children: _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "curl"
            }), " -sfL https://get.k3s.io ", _jsx(_components.span, {
              className: "token operator",
              children: "|"
            }), " ", _jsx(_components.span, {
              className: "token function",
              children: "sh"
            }), " -s server --cluster-init --token ", _jsx(_components.span, {
              className: "token string",
              children: "\"XAVIER_HIS_SECRET_TOKEN\""
            }), " --write-kubeconfig-mode ", _jsx(_components.span, {
              className: "token number",
              children: "644"
            }), " --docker\n"]
          })
        })
      }), "\n", _jsxs(_components.blockquote, {
        children: ["\n", _jsxs(_components.p, {
          children: ["💡 As you may have noticed, we are utilizing ", _jsx(_components.code, {
            children: "--write-kubeconfig-mode 644"
          }), " to allow our cluster to be imported in rancher later."]
        }), "\n"]
      }), "\n", _jsxs(_components.blockquote, {
        children: ["\n", _jsxs(_components.p, {
          children: ["💡 ", _jsx(_components.code, {
            children: "--docker"
          }), " will enable us to use our host's Docker registry and use local images"]
        }), "\n"]
      }), "\n", _jsx(_components.h3, {
        children: "Worker Node Setup"
      }), "\n", _jsx(_components.p, {
        children: "Finally, we can add our Worker (or Agent) nodes by running the following line on each of the nodes:"
      }), "\n", _jsx(_components.pre, {
        className: "language-bash",
        children: _jsx(_components.code, {
          className: "language-bash code-highlight",
          children: _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "curl"
            }), " -sfL https://get.k3s.io ", _jsx(_components.span, {
              className: "token operator",
              children: "|"
            }), " ", _jsx(_components.span, {
              className: "token function",
              children: "sh"
            }), " -s agent --server https://MASTER_NODE_IP_ADDRESS:6443 --token ", _jsx(_components.span, {
              className: "token string",
              children: "\"XAVIER_HIS_SECRET_TOKEN\""
            }), "\n"]
          })
        })
      }), "\n", _jsx(_components.h3, {
        children: "Verification"
      }), "\n", _jsxs(_components.p, {
        children: ["When we now execute ", _jsx(_components.code, {
          children: "sudo k3s kubectl get nodes -w"
        }), " on our master node, we will see our nodes appearing."]
      }), "\n", _jsx(_components.p, {
        children: _jsx(_components.img, {
          src: "./k3s-node-setup.png",
          alt: ""
        })
      }), "\n", _jsx(_components.h2, {
        children: "Installing and Configuring Kubectl"
      }), "\n", _jsxs(_components.blockquote, {
        children: ["\n", _jsx(_components.p, {
          children: "⚠ Execute this on your local machine! (WSL in this case)"
        }), "\n"]
      }), "\n", _jsx(_components.h3, {
        children: "Installing Kubectl"
      }), "\n", _jsxs(_components.p, {
        children: ["Since we now have a Kubernetes cluster running, it is time to configure it to run on our local system. For that I am utilizing the ", _jsx(_components.code, {
          children: "WSL"
        }), " or ", _jsx(_components.code, {
          children: "Windows Subsystem for Linux"
        }), " system as explained in a ", _jsx(_components.a, {
          href: "/post/2021/1/4/coding-python-windows-development-environment-setup",
          children: "previous article"
        }), "."]
      }), "\n", _jsx(_components.p, {
        children: "To install WSL we execute the following:"
      }), "\n", _jsx(_components.pre, {
        className: "language-bash",
        children: _jsxs(_components.code, {
          className: "language-bash code-highlight",
          children: [_jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Download Kubectl"
            }), "\n"]
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "curl"
            }), " -LO ", _jsxs(_components.span, {
              className: "token string",
              children: ["\"https://dl.k8s.io/release/", _jsxs(_components.span, {
                className: "token variable",
                children: [_jsx(_components.span, {
                  className: "token variable",
                  children: "$("
                }), _jsx(_components.span, {
                  className: "token function",
                  children: "curl"
                }), " -L -s https://dl.k8s.io/release/stable.txt", _jsx(_components.span, {
                  className: "token variable",
                  children: ")"
                })]
              }), "/bin/linux/amd64/kubectl\""]
            }), "\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "\n"
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Install it"
            }), "\n"]
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "sudo"
            }), " ", _jsx(_components.span, {
              className: "token function",
              children: "install"
            }), " -o root -g root -m 0755 kubectl /usr/local/bin/kubectl\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "\n"
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Verify"
            }), "\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "kubectl version --client\n"
          })]
        })
      }), "\n", _jsx(_components.h3, {
        children: "Configuring Kubectl"
      }), "\n", _jsx(_components.p, {
        children: "Now this is installed, we can configure it towards our remote cluster:"
      }), "\n", _jsx(_components.pre, {
        className: "language-bash",
        children: _jsxs(_components.code, {
          className: "language-bash code-highlight",
          children: [_jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Copy over the Kubeconfig file to our local machine"
            }), "\n"]
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "scp"
            }), " xavier@xavier-nvidia-node-1.local:/etc/rancher/k3s/k3s.yaml nvidia-node-1-config.yaml\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "\n"
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Replace localhost with the location of the master node"
            }), "\n"]
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "sed"
            }), " -ie s/127.0.0.1/xavier-nvidia-node-1.local/g nvidia-node-1-config.yaml\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "\n"
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Verify that it works"
            }), "\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "kubectl --kubeconfig nvidia-node-1-config.yaml get pods -A\n"
          }), _jsx(_components.span, {
            className: "code-line",
            children: "\n"
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Then copy it over to kubectl directory and activate it"
            }), "\n"]
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# we can now run kubectl get pods -A"
            }), "\n"]
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token function",
              children: "cp"
            }), " nvidia-node-1-config.yaml ~/.kube/config\n"]
          })]
        })
      }), "\n", _jsx(_components.h2, {
        children: "Troubleshooting"
      }), "\n", _jsx(_components.h3, {
        children: "Agent does not want to join"
      }), "\n", _jsx(_components.p, {
        children: "Sometimes an agent does not want to join, in that case try utilizing the IP address directly."
      }), "\n", _jsx(_components.h3, {
        children: "Removing K3S"
      }), "\n", _jsx(_components.p, {
        children: "⚠ In case something happened, you can uninstall K3S with the following commands:"
      }), "\n", _jsx(_components.pre, {
        className: "language-bash",
        children: _jsxs(_components.code, {
          className: "language-bash code-highlight",
          children: [_jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Uninstall Server Node"
            }), "\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "/usr/local/bin/k3s-uninstall.sh\n"
          }), _jsx(_components.span, {
            className: "code-line",
            children: "\n"
          }), _jsxs(_components.span, {
            className: "code-line",
            children: [_jsx(_components.span, {
              className: "token comment",
              children: "# Uninstall Agent Node"
            }), "\n"]
          }), _jsx(_components.span, {
            className: "code-line",
            children: "/usr/local/bin/k3s-agent-uninstall.sh\n"
          })]
        })
      }), "\n", _jsx(_components.h2, {
        children: "Conclusion"
      }), "\n", _jsxs(_components.p, {
        children: ["We have now set-up our Nvidia Jetson nodes and are running a Kubernetes cluster on it! To keep this article light, I have decided to move the demonstration to a new article where you will learn to ", _jsx(_components.a, {
          href: "/post/2021/5/30/infrastructure-nvidia-nvidia-jetson-k3s-dapr-and-grafana",
          children: "run Dapr as your distributed application runtime as well as Grafana to monitor your nodes"
        })]
      })]
    });
  }
}
return {
  default: MDXContent
};


Learn how you can install Kubernetes on a small GPU Cluster for edge computation purposes

Xavier Geerinck

Installing Kubernetes on an Nvidia Jetson Cluster

Are you struggling with testing and integrating AI into your product?

Are you struggling with testing and integrating AI into your product?