Skip to content

Latest commit

 

History

History
655 lines (568 loc) · 14.8 KB

kubernetes.md

File metadata and controls

655 lines (568 loc) · 14.8 KB
Raw

useful links

Architecture

architecture nodes with software kubernetes

microk8s

installation

sudo snap install microk8s --classic
sudo snap install microk8s --classic --edge

enable addons

microk8s.start
microk8s.enable dns dashboard

check installation

microk8s.inspect

check journals for services

journalctl -u snap.microk8s.daemon-docker
  • snap.microk8s.daemon-apiserver
  • snap.microk8s.daemon-controller-manager
  • snap.microk8s.daemon-scheduler
  • snap.microk8s.daemon-kubelet
  • snap.microk8s.daemon-proxy
  • snap.microk8s.daemon-docker
  • snap.microk8s.daemon-etcd

minikube

installation

sudo snap install minikube

installation

curl -Lo minikube https://storage.googleapis.com/minikube/releases/latest/minikube-linux-amd64 && chmod +x minikube
curl -Lo kubectl https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/linux/amd64/kubectl && chmod +x kubectl

export MINIKUBE_WANTUPDATENOTIFICATION=false
export MINIKUBE_WANTREPORTERRORPROMPT=false
export MINIKUBE_HOME=$HOME
export CHANGE_MINIKUBE_NONE_USER=true
mkdir $HOME/.kube || true
touch $HOME/.kube/config

export KUBECONFIG=$HOME/.kube/config
sudo -E ./minikube start --vm-driver=none

# wait that Minikube has created
for i in {1..150}; do # timeout for 5 minutes
   ./kubectl get po &> /dev/null
   if [ $? -ne 1 ]; then
      break
  fi
  sleep 2
done

set up env

minikube completion bash

start

minikube start

uninstall kubectl, uninstall minikube

kubectl delete node --all
kubectl delete pods --all
kubectl stop
kubectl delete

launchctl stop '*kubelet*.mount'
launchctl stop localkube.service
launchctl disable localkube.service

sudo kubeadm reset
rm -rf ~/.kube ~/.minikube
sudo rm -rf /usr/local/bin/localkube /usr/local/bin/minikube
sudo rm -rf /etc/kubernetes/

# sudo apt-get purge kubeadm kubectl kubelet kubernetes-cni kube*
sudo apt-get purge kube*
sudo apt-get autoremove

docker system prune -af --volumes

start without VirtualBox/KVM

export MINIKUBE_WANTUPDATENOTIFICATION=false
export MINIKUBE_WANTREPORTERRORPROMPT=false
export MINIKUBE_HOME=$HOME
export CHANGE_MINIKUBE_NONE_USER=true

export KUBECONFIG=$HOME/.kube/config
sudo -E minikube start --vm-driver=none

kubectl using minikube context

permanently

kubectl config use-context minikube

temporary

kubectl get pods --context=minikube

check namespaces

kubectl get namespaces

at least three namespaces will be provided

default       Active    15m
kube-public   Active    15m
kube-system   Active    15m

create namespace

kubectl create namespace my-own-namespace

or via yaml file

kubectl apply -f {filename}

kind: Namespace
apiVersion: v1
metadata:
  name: test

delete namespace

kubectl delete namespace {name of namespace}

get configurations, read configuration in specific format

kubectl get configmap 
kubectl get configmap --namespace kube-system 
kubectl get configmap --namespace kube-system kube-proxy --output json

start readiness, check cluster

kubectl cluster-info dump
kubectl get node
minikube dashboard

addons

minikube addons list
minikube addons enable ingress

deployment

main schema

start dummy container

kubectl run hello-minikube --image=k8s.gcr.io/echoserver:1.4 --port=8080

create deployment ( with replica set )

kubectl run http --image=katacoda/docker-http-server:latest --replicas=1

scale deployment

kubectl scale --replicas=3 deployment {name of the deployment}

create from yaml file, update yaml file

kubectl apply -f /path/to/controller.yml
kubectl create -f /path/to/controller.yml

create service fastly

kubectl expose deployment helloworld-deployment --type=NodePort --name=helloworld-service
kubectl expose deployment helloworld-deployment --external-ip="172.17.0.13" --port=8000 --target-port=80

reach out service

minikube service helloworld-service
minikube service helloworld-service --url

service port range

kube-apiserver --service-node-port-range=30000-40000

describe resources

kubectl describe deployment {name of deployment}
kubectl describe service {name of service}

describe users, user token

kubectl --namespace kube-system describe secret admin-user

get resources

kubectl get all --all-namespaces
kubectl get pods
kubectl get pods --namespace kube-system
kubectl get pods --show-labels
kubectl get pods --output=wide --selector="run=load-balancer-example" 
kubectl get service --output=wide
kubectl get service --output=wide --selector="app=helloworld"
kubectl get deployments
kubectl get replicasets
kubectl get nodes
kubectl get cronjobs
kubectl get daemonsets
kubectl get pods,deployments,services,rs,cm,pv,pvc -n demo

determinate cluster 'hostIP' to reach out application(s)

minikube ip

open 'kube-dns-....'/hostIP open 'kube-proxy-....'/hostIP

edit configuration of controller

kubectl edit pod hello-minikube-{some random hash}
kubectl edit deploy hello-minikube
kubectl edit ReplicationControllers helloworld-controller
kubectl set image deployment/helloworld-deployment {name of image}

rollout status

kubectl rollout status  deployment/helloworld-deployment

rollout history

kubectl rollout history  deployment/helloworld-deployment
kubectl rollout undo deployment/helloworld-deployment
kubectl rollout undo deployment/helloworld-deployment --to-revision={number of revision from 'history'}

delete running container

kubectl delete pod hello-minikube-6c47c66d8-td9p2

delete deployment

kubectl delete deploy hello-minikube

delete ReplicationController

kubectl delete rc helloworld-controller

port forwarding from local to pod/deployment/service

next receipts allow to redirect 127.0.0.1:8080 to pod:6379

kubectl port-forward redis-master-765d459796-258hz      8080:6379 
kubectl port-forward pods/redis-master-765d459796-258hz 8080:6379
kubectl port-forward deployment/redis-master            8080:6379 
kubectl port-forward rs/redis-master                    8080:6379 
kubectl port-forward svc/redis-master                   8080:6379

NodeSelector for certain host

spec:
  nodeSelector: 
    kubernetes.io/hostname: gtxmachine1-ev

persistent volume

kind: PersistentVolume
apiVersion: v1
metadata:
  name: pv-volume3
  labels:
    type: local
spec:
  capacity:
    storage: 10Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/mnt/data3"

to access created volume

ls /mnt/data3

list of existing volumes

kubectl get pv 
kubectl get pvc

Serverless

  • OpenFaas
  • Kubeless
  • Fission
  • OpenWhisk

labels

show labels for each node

kubectl get nodes --show-labels

add label to Node

kubectl label nodes {node name} my_label=my_value

remove label from Node

kubectl label nodes {node name} my_label-

deploy Pod on Node with label

apiVersion: v1
kind: Pod
metadata:
...
spec:
...
  nodeSelector:
    my_label=my_value

create Deployment for specific node

apiVersion: some-version
kind: Deployment
metadata:
...
spec:
...
  nodeSelector:
    my_label=my_value

resolving destination node

when label was not found

  • nodeAffinity
    • preferred - deploy in any case, with preferrence my_label=my_value
spec:
  affinity:
    nodeAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 1
        preference:
          matchExpressions:
          - key: my_label
            operator: In
            values:
            - my_value
    • required - deploy only when label matched my_label=my_value
spec:
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: my_label
            operator: In
            values:
            - my_value
  • nodeAntiAffinity
spec:
  affinity:
    nodeAntiAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
  • podAffinity
    • preferred spec.affinity.podAffinity.preferredDuringSchedulingIgnoredDuringExecution
    • required spec.affinity.podAffinity.requiredDuringSchedulingIgnoredDuringExecution
  • podAntiAffinity
    • preferred spec.affinity.podAntiAffinity.preferredDuringSchedulingIgnoredDuringExecution
    • required spec.affinity.podAntiAffinity.requiredDuringSchedulingIgnoredDuringExecution

delete node from cluster

kubectl get nodes
kubectl delete {node name}

add node to cluster

ssh {master node}
kubeadm token create --print-join-command  --ttl 0

expected result from previous command

kubeadm join 10.14.26.210:6443 --token 7h0dmx.2v5oe1jwed --discovery-token-ca-cert-hash sha256:1d28ebf950316b8f3fdf680af5619ea2682707f2e966fc0

go to node, clean up and apply token

ssh {node address}
rm -rf /etc/kubernetes
# apply token from previous step with additional flag: --ignore-preflight-errors=all
kubeadm join 10.14.26.210:6443 --token 7h0dmx.2v5oe1jwed --discovery-token-ca-cert-hash sha256:1d28ebf950316b8f3fdf680af5619ea2682707f2e966fc0 --ignore-preflight-errors=all

expected result from previous command

...
This node has joined the cluster:
* Certificate signing request was sent to master and a response
  was received.
* The Kubelet was informed of the new secure connection details.
 
Run 'kubectl get nodes' on the master to see this node join the cluster.

next block is not mandatory in most cases

systemctl restart kubelet

logs

kubectl logs <name of pod>

create dashboard

kubectl create -f https://raw.githubusercontent.com/kubernetes/dashboard/master/aio/deploy/recommended/kubernetes-dashboard.yaml

access dashboard

kubectl -n kube-system describe secret admin-user
http://127.0.0.1:8001/api/v1/namespaces/kube-system/services/https:kubernetes-dashboard:/proxy/#!/overview?namespace=default
kubectl proxy

common

execute command on specific pod

kubectl exec -it {name of a pod}  -- bash -c "echo hi > /path/to/output/test.txt"

Flannel

restart nodes

# remove died pods
kubectl delete pods kube-flannel-ds-amd64-zsfz  --grace-period=0 --force
kubectl delete -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml
kubectl create  -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml

read logs

kubectl logs --namespace kube-system kube-flannel-ds-amd64-j4frw -c kube-flannel

read logs from all pods

for each_node in $(kubectl get pods --namespace kube-system | grep flannel | awk '{print $1}');do echo $each_node;kubectl logs --namespace kube-system $each_node -c kube-flannel;done

read settings

kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp ls /etc/kube-flannel/
kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp cat /etc/kube-flannel/cni-conf.json
kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp cat /etc/kube-flannel/net-conf.json

kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp ls /run/flannel/
kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp cat /run/flannel/subnet.env

kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp ls /etc/cni/net.d
kubectl --namespace kube-system exec kube-flannel-ds-amd64-wc4zp cat /etc/cni/net.d/10-flannel.conflist

read DNS logs

kubectl get svc --namespace=kube-system | grep kube-dns
kubectl logs --namespace=kube-system coredns-78fcd94-7tlpw | tail

Helm

documentation

Architecture

main components

installation

sudo snap install helm --classic
curl https://raw.githubusercontent.com/kubernetes/helm/master/scripts/get | bash

de-installation

helm reset

initialization

helm init
# sync latest available packages
helm repo update

useful variables

* $HELM_HOME: the location of Helm's configuration
* $TILLER_HOST: the host and port that Tiller is listening on
* $HELM_BIN: the path to the helm command on your system
* $HELM_PLUGIN_DIR: the full path to this plugin (not shown above, but we'll see it in a moment).

analyze local package

helm inspect { folder }
helm lint { folder }

search remote package

helm search 
helm describe {full name of the package}

information about remote package

helm info {name of resource}
helm status {name of resource}

create package locally

helm create

folder structure

create package with local templates

ls -la ~/.helm/starters/

install package

helm install { full name of the package }
helm install --name {my name for new package} { full name of the package }
helm install --name {my name for new package} --namespace {namespace} -f values.yml --debug --dry-run { full name of the package }

install aws plugin

helm plugin install https://github.com/hypnoglow/helm-s3.git

list of installed packages

helm list
helm list --all
helm ls

package upgrade

local package

helm upgrade  {deployment/svc/rs/rc name} . --set replicas=2,maria.db.password="new password"

package by name

helm upgrade {name of package} {folder with helm scripts} --set replicas=2

check upgrade

helm history
helm rollback {name of package} {revision of history}

remove packageHelm

helm delete --purge {name of package}

trouble shooting

issue with 'helm list'

E1209 22:25:57.285192    5149 portforward.go:331] an error occurred forwarding 40679 -> 44134: error forwarding port 44134 to pod de4963c7380948763c96bdda35e44ad8299477b41b5c4958f0902eb821565b19, uid : unable to do port forwarding: socat not found.
Error: transport is closing

solution

sudo apt install socat

incompatible version of client and server

Error: incompatible versions client[v2.12.3] server[v2.11.0]

solution

helm init --upgrade
kubectl get pods --namespace kube-system # waiting for start Tiller
helm version

template frameworks