Skip to content

Latest commit

 

History

History

Lab3.InstanceForKubernetes

README.md

Kubernetes Install With Kubespray

0. 변수 셋팅 파일 생성 후 실행

  • 아래와 같은 내용을 ~/.bashrc에 추가하고 실행해 줍니다.
su - ubuntu

echo '
export TF_VAR_AWS_ACCESS_KEY="xxxxxxx"
export TF_VAR_AWS_SECRET_KEY="xxxxxxxxxxxxxxx"
export TF_VAR_AWS_REGION="ap-northeast-2"
'>> ~/.bashrc
. ~/.bashrc

1. OS key 생성 [있으면 생략]

ssh-keygen -f ~/.ssh/id_rsa -N ''
  • cf) 설치 대상 host에 Public-key 배포 ssh-copy-id root@10.0.2.10

2. Terrform 으로 host 셋팅

cd
git clone https://github.com/Finfra/sreMsa
cd ~/sreMsa/Lab1.Kubespray/Lab3.InstanceForKubernetes
terraform init
terraform apply --auto-approve

3. Hosts파일 셋팅

aws configure
  # security setting
    AWS Access Key ID [None]: xxxxxxxxxx
    AWS Secret Access Key [None]: xxxxxxxxxxxxxxxxxxx
    Default region name [None]: ap-northeast-2
    Default output format [None]: text
cd ~/sreMsa/Lab1.Kubespray/Lab3.InstanceForKubernetes
# rm -rf ~/.ssh/known_hosts
bash doSetHosts.sh

3.5. 모든 호스트 repo update

  • 필수 아님. repo상태 안좋을때만,
hosts=("vm01" "vm02" "vm03")
for host in "${hosts[@]}"; do
  echo "Connecting to $host ..."
  ssh "$host" << EOF
    # apt update 및 repository 추가
    sudo apt update
    REPO_LINE="deb http://mirror.kakao.com/ubuntu/ noble main universe"
    if ! grep -Fxq "\$REPO_LINE" /etc/apt/sources.list; then
      echo "\$REPO_LINE" | sudo tee -a /etc/apt/sources.list
      echo "Repository line added successfully."
    else
      echo "Repository line already exists."
    fi
    sudo apt-get update
EOF
  if [ $? -eq 0 ]; then
    echo "Script executed successfully on $host."
  else
    echo "Failed to execute script on $host."
  fi
done
#cf : deb http://ftp.daum.net/ubuntu/ noble main universe
  • cf) 아래와 같이 /etc/hosts파일을 직접 셋팅 해도 됨
52.213.183.141 vm01
54.75.118.15   vm02
54.75.118.154  vm03

4. git Clone

cd
git clone -b release-2.25 https://github.com/kubernetes-sigs/kubespray

5. inventory파일 생성

cd kubespray
cat > inventory/inventory.ini <<EOF
[all]
vm01 etcd_member_name=etcd1
vm02 etcd_member_name=etcd2
vm03 etcd_member_name=etcd3

[kube_control_plane]
vm01
vm02

[etcd]
vm01
vm02
vm03

[kube-node]
vm01
vm02
vm03

[k8s-cluster:children]
kube_control_plane
kube-node
EOF

6. kubesparyInstall.sh 실행

  • pip error시 requirements.txt파일에서 에러가 발생하는 페키지의 "=="부터 줄의 끝까지 제거.
  • python3.12버전에서는 --break-system-packages 옵션 필요.
# ubuntu24.04에서만 : sudo apt remove -y python3-jsonschema
sudo python3 -m pip install -r requirements.txt

ansible-playbook --flush-cache -u ubuntu -b --become --become-user=root \
  -i inventory/inventory.ini \
  cluster.yml

Admin

Shutdown All Instance

for i in $(seq 3); do ssh ubuntu@vm0$i sudo sh -c 'shutdown -h now'; done

Startup all Instance

ids=$(aws ec2 describe-instances  --filters "Name=tag-value,Values=vm0*" --query "Reservations[].Instances[].InstanceId" --output text)
for i in $ids; do     aws ec2 start-instances --instance-ids $i ;done

vm destroy

terraform destroy --auto-approve