Skip to content

Latest commit

 

History

History
209 lines (144 loc) · 5.12 KB

README.md

File metadata and controls

209 lines (144 loc) · 5.12 KB
Raw

KubernetesTesting

Just a personal testing repository

Install K8s

I used Docker Desktop

Install Nginx Controller

I used helm as I am lazy

helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx
helm repo update
helm install nginx-ingress ingress-nginx/ingress-nginx

Apply YAML files

kubectl apply -f .

How to make a pod not ready

I did this by setting the readiness probe to check for the host.html file. If the file vanishes the probe fails:

To debug this I was exec'ing onto the pod and running these commands:

# Make pod not ready
mv /usr/local/apache2/htdocs/host.html /usr/local/apache2/htdocs/host.html2
# Make pod ready
mv /usr/local/apache2/htdocs/host.html2 /usr/local/apache2/htdocs/host.html

Outcome Nginx does not work, as soon as the pod is not ready nginx removes it from the pool and does not route any traffic to that pod. Found out that nginx only uses the readiness probe and not the liveliness probe

Remove Nginx Ingress Controller

helm uninstall nginx-ingress

Install HAProxy (from HAProxyTeh)

helm repo add haproxytech https://haproxytech.github.io/helm-charts
helm repo update
helm install haproxy-ingress haproxytech/kubernetes-ingress \
  --set controller.kind=DaemonSet \
  --set controller.ingressClass=haproxy \
  --set controller.service.type=LoadBalancer

Same issue as Nginx = failed

Trying this HAProxy - https://github.com/jcmoraisjr/haproxy-ingress

First to install it:

kubectl create -f https://haproxy-ingress.github.io/resources/haproxy-ingress.yaml
kubectl label node docker-desktop role=ingress-controller

Check its running:

kubectl -n ingress-controller get daemonset
kubectl -n ingress-controller get pod

Now deploy the load balancer

kubectl apply -f jcmoraisjr-loadbalancer.yaml

Draining is still not working when tested. We need to enable drain-support flag

kubectl edit configmap haproxy-ingress -n ingress-controller

So the YAML looks like this:

apiVersion: v1
data:
  drain-support: "true"
kind: ConfigMap
metadata:
  creationTimestamp: "2020-09-12T13:57:35Z"
  name: haproxy-ingress
  namespace: ingress-controller
  resourceVersion: "192909"
  selfLink: /api/v1/namespaces/ingress-controller/configmaps/haproxy-ingress
  uid: beca0398-69d6-47eb-b9b7-bd7c318f49a4

Restart the ingress pod (not sure its required)

kubectl delete pod haproxy-ingress-wn9hd -n ingress-controller

It works

Istio

Download and install istioctl - https://istio.io/latest/docs/setup/getting-started/

curl -L https://istio.io/downloadIstio | sh -

Install istio and lavel namespace for auto proxy envoy

istioctl install --set profile=demo
kubectl label namespace default istio-injection=enabled

Deploy deployment and service:

kubectl apply -f deployment.yaml
kubectl apply -f service.yaml

Then apply istio config (virtaul service, gateway and destination rule)

kubectl apply -f istio-gateway.yaml

Test - session stickyness works - limits scaling (up or down) recreates all cookies and re-routes all traffic.

There is no way to drain sessions using this ingress.

Set up an AKS with AGIC

Followed this page: https://docs.microsoft.com/en-us/azure/application-gateway/tutorial-ingress-controller-add-on-new

Register for feature

az feature register --name AKS-IngressApplicationGatewayAddon --namespace Microsoft.ContainerService

Set up local CLI

Wait for feature to cahnge to Registered - about 15 minutes

az feature list -o table --query "[?contains(name, 'Microsoft.ContainerService/AKS-IngressApplicationGatewayAddon')].{Name:name,State:properties.state}"

Referesh registration

az provider register --namespace Microsoft.ContainerService

Add extenion to CLI

az extension add --name aks-preview
az extension list

Create Resource Group and Cluster

az group create --name myResourceGroup --location ukwest
az aks create -n myCluster -g myResourceGroup --network-plugin azure --enable-managed-identity -a ingress-appgw --appgw-name myApplicationGateway --appgw-subnet-prefix "10.2.0.0/16" --node-count 1 --kubernetes-version 1.18.8 --generate-ssh-keys

Deploy Apps

Deploy the Apps. Get Creds first:

az aks get-credentials -n myCluster -g myResourceGroup

Now deploy the components:

kubectl apply -f deployment.yaml
kubectl apply -f service.yaml
kubectl apply -f agic-ingress.yaml

Now to do persistance testing.

Helathness probe does NOT use the readiness probe - shame, it uses the liveness probe

Even if you set AGIC probe manually it still doesn't work. As soon as AGIC probe fails no more sessions can route to the pod (as its removed from the pool).

curl 'http://52.142.173.222/'  -H 'Cookie: ApplicationGatewayAffinity=7626ab90412fde0c1d28be613e212f55' ;
# Make pod not ready
mv /usr/local/apache2/htdocs/host.html /usr/local/apache2/htdocs/host.html2
# Make pod ready
mv /usr/local/apache2/htdocs/host.html2 /usr/local/apache2/htdocs/host.html