Skip to content

Latest commit

 

History

History
216 lines (159 loc) · 6.49 KB

create-cluster.md

File metadata and controls

216 lines (159 loc) · 6.49 KB
Raw

Install Vagrant

ATTENTION!!! Tested commands on Ubuntu 20.04 in 2022.

wget -O- https://apt.releases.hashicorp.com/gpg | gpg --dearmor | sudo tee /usr/share/keyrings/hashicorp-archive-keyring.gpg
echo "deb [signed-by=/usr/share/keyrings/hashicorp-archive-keyring.gpg] https://apt.releases.hashicorp.com $(lsb_release -cs) main" | sudo tee /etc/apt/sources.list.d/hashicorp.list
sudo apt update && sudo apt install vagrant

Using Vagrant with VirtualBox

ATTENTION!!! The steps in this section should only be performed if you want to use vagrant with VirtualBox.

  • Install VirtualBox.

  • Access directory of Vagrantfile:

cd learning-cka/vagrant/virtualbox
  • Run the follow commands to install the plugins:
vagrant plugin install vagrant-vbguest
vagrant plugin install vagrant-disksize

Run steps of section GENERIC_STEPS - Create Cluster with Vagrant.

Using Vagrant with Libvirt/QEMU KVM

ATTENTION!!! The steps in this section should only be performed if you want to use vagrant with Libvirt/QEMU KVM. ATTENTION!!! Tested commands on Ubuntu 21.04 in 2022.

cd learning-cka/vagrant/libvirt
  • Run the follow commands to install the plugin:
vagrant plugin install vagrant-libvirt

Run steps of section GENERIC_STEPS - Create Cluster with Vagrant.

GENERIC_STEPS - Create Cluster with Vagrant

  • Start virtual machines (VMs) with commands:
vagrant up --provision
  • Access each VM with SSH
vagrant ssh master
vagrant ssh worker1
vagrant ssh worker2
  • Stop VMs:
vagrant halt
  • Destroy VMs:
vagrant destroy

  • In each VM install this tools.

  • Run the commands with type node:

#------- Specifics (master)
# Fix bug between containerd and kubeadm
# source: https://github.com/containerd/containerd/issues/4581

sudo rm /etc/containerd/config.toml
sudo systemctl restart containerd
sudo setfacl -m user:$USER:rw /var/run/containerd/containerd.sock

kubeadm config images pull
sudo kubeadm init --apiserver-advertise-address 192.168.56.10 --kubernetes-version 1.24.3
# Reset configuration
# sudo kubeadm reset
# sudo rm -rf /etc/cni/net.d
# sudo rm $HOME/.kube/config

# Config kubectl
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

# Enable modules of Kernel for pod network (weave-net)
sudo modprobe br_netfilter ip_vs ip_vs_rr ip_vs_wrr ip_vs_sh nf_conntrack_ipv4
# Troubleshooting
# https://github.com/kubernetes/kubeadm/issues/975#issuecomment-403081740

# Permanently enable modules of Kernel for pod network (weave-net)
cat << EOF | sudo tee -a /etc/modules-load.d/modules.conf
br_netfilter
ip_vs
ip_vs_rr
ip_vs_wrr
ip_vs_sh
nf_conntrack_ipv4 
EOF

cat /etc/modules-load.d/modules.conf

# Install weave-net
kubectl apply -f "https://cloud.weave.works/k8s/net?k8s-version=$(kubectl version | base64 | tr -d '\n')"
kubectl get pods -A
kubectl get nodes

#------- Specifics (worker1 and worker2)
# Allow all these ports: https://github.com/badtuxx/DescomplicandoKubernetes/blob/main/pt/day_one/descomplicando_kubernetes.md#portas-que-devemos-nos-preocupar
#
# In master node print a join command for add worker node in cluster
kubeadm token create --print-join-command
#
# Example of command to run in worker node:
# sudo kubeadm join 192.168.56.10:6443 --token 2o2t5a.t28j1hh4w21vti05 --discovery-token-ca-cert-hash sha256:cd84d6f4b8e975c7fcffa5bce7bdc2f19803647bc507bb0b06cc600d9fa72738

# Fix bug between containerd and kubeadm in worker1 and worker2
# source: https://github.com/containerd/containerd/issues/4581
sudo rm /etc/containerd/config.toml
sudo systemctl restart containerd
sudo setfacl -m user:$USER:rw /var/run/containerd/containerd.sock

# Fix problem in worker1 and worker2 with wave
# References:
# http://www.bennythejudge.com/kubernetes/cni/weaver/2020/05/02/weave-pod-crashes-on-worker-digitalocean.html
# https://stackoverflow.com/questions/54550285/readiness-probe-failed-error-in-weave-kubernetes
sudo modprobe br_netfilter ip_vs ip_vs_rr ip_vs_wrr ip_vs_sh nf_conntrack_ipv4
kubectl get svc kube-dns -n kube-system
# Change 10.96.0.10/32 to the network address in the output of the previous command
sudo iptables -t nat -I KUBE-SERVICES -d 10.96.0.1/32 -p tcp -m comment --comment "default/kubernetes:https cluster IP" -m tcp --dport 443 -j KUBE-MARK-MASQ


#------- Generic (master, worker1 and worker2)
# Save iptables rules
sudo apt install iptables-persistent
sudo iptables-save
# Reference: https://linuxconfig.org/how-to-make-iptables-rules-persistent-after-reboot-on-linux


# Permanently enable modules of Kernel for pod network (weave-net)
cat << EOF | sudo tee -a /etc/modules-load.d/modules.conf
br_netfilter
ip_vs
ip_vs_rr
ip_vs_wrr
ip_vs_sh
nf_conntrack_ipv4 
EOF

# Fix problem with nodes after reboot VMs
# Permanently configure kubelet in each VM
# Reference: https://stackoverflow.com/questions/51154911/kubectl-exec-results-in-error-unable-to-upgrade-connection-pod-does-not-exi

sudo iptables -t nat -I KUBE-SERVICES -d 10.96.0.1/32 -p tcp -m comment --comment "default/kubernetes:https cluster IP" -m tcp --dport 443 -j KUBE-MARK-MASQ

# Only master
cat << EOF | sudo tee /etc/default/kubelet
KUBELET_EXTRA_ARGS="--node-ip=192.168.56.10"
EOF
sudo systemctl daemon-reload
sudo systemctl restart kubelet

# Only worker1
cat << EOF | sudo tee /etc/default/kubelet
KUBELET_EXTRA_ARGS="--node-ip=192.168.56.11"
EOF
sudo systemctl daemon-reload
sudo systemctl restart kubelet

# Only worker2
cat << EOF | sudo tee /etc/default/kubelet
KUBELET_EXTRA_ARGS="--node-ip=192.168.56.12"
EOF
sudo systemctl daemon-reload
sudo systemctl restart kubelet

# List jobs running in background
#jobs -l

# Finish jobs running in background
#kill %1 %2 %3 %4 %5