Skip to content

K3s

Jump to bottom
Mehdi Hadeli edited this page Apr 6, 2023 · 2 revisions

Articles

Videos

K3s Notes

  • A kubeconfig file will be written to /etc/rancher/k3s/k3s.yaml and the kubectl installed by K3s will automatically use it.
  • By default, kubectl looks for a file named config in the $HOME/.kube directory. You can specify other kubeconfig files by setting the KUBECONFIG environment variable or by setting the --kubeconfig flag.
  • installing traefik with k3s with k3s [ServiceLB (Klipper Load Balancer)](https://docs.k3s.io/networking#service-load-balancer) load-balancer or install traefik helm manually and install metallb for load-balancing
  • When we install traefik manullay with metallb as load-balancer for load-balancer, metallb reserved ips are not accessible outside of wsl because our wsl only accessible with its public ip wsl hostname -I, but by installing traefik and k3s load-balancer [ServiceLB (Klipper Load Balancer)](https://docs.k3s.io/networking#service-load-balancer) this problem will solve with assigning wsl hostname -I ip to load-balancer service type and traefik because this is ip of our wsl and it is accessible from the windows
  • traefik helm install as default in k3s but for activating dashboard we should define a custom HelmChartConfig for traefik helm, also we can install traefik ingress manually with helm by disable installing with installation scrip in yaml configuration with disable: "traefik". traefik ingress controller as default is behind of a external load-balancer service type, if we use could provider, cloud load-balancer service type else we should install metallb else this service type will be pending (we can use its created node-port because each load-balancer service type creates a node-port on each cluster internally), with both approach manually or installing with k3s will create a load-balancer service type so if we don't have a load-balancer like metallb traefik service status in EXTERNAL-IP section will be Pending
  • If the ServiceLB Pod runs on a node that has an external IP configured, the node's external IP is populated into the Service's status.loadBalancer.ingress address list. Otherwise, the node's internal IP is used. we can see detail of external and internal ip cluster node with kubectl get nodes -o wide and for k3s with one cluster external ip is empty.
NAME    STATUS   ROLES                       AGE    VERSION        INTERNAL-IP      EXTERNAL-IP  
mehdi   Ready    control-plane,etcd,master   2d5h   v1.25.6+k3s1   172.21.-.-   <none>
  • It is possible to expose multiple load-balancer Services type on the same node, as long as they use different ports.
  • If you try to create a LoadBalancer Service that listens on port 80, the ServiceLB will try to find a free host (cluster node internal ip) in the cluster for port 80. If no host with that port is available, the LB will remain Pending, so in this case we should use different port on the host for load-balancer service type.
  • we setup Cert-Manager as certificate manager
  • setup rancher as cluster management
  • According to the Kubernetes documentation, it is recommended to put resources related to the same microservice or application tier into the same file[1]. This helps in organizing and managing resources more efficiently. It also makes it easier to understand and maintain the configuration files.

Tools

Clone this wiki locally