kubernetes-lab-setup

Kubernetes setup

Before you begin

1. Consider creating terraform.tfvars files in every Terraform directory not to pass the variables each time apply is run
2. If you see that an instance has more than one private IP it may indicate that this error has occured. If that is the case run terraform destroy and attemp terraform apply again
3. If an instance is stuck with BUILD or ERROR status do the same
4. If an instance fails to complete its user_data script do the same
5. Usage of one SSH key default for all machines is assumed (besides your local machine)
6. If you get a warning WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! when connecting via SSH do as the suggestion says: ssh-keygen -f "/home/ubuntu/.ssh/known_hosts" -R "<ip>"
7. Ignore sudo: unable to resolve host <machine-hostname>: Name or service not known errors

From your local machine

1. Provide required keys in configs/keys
2. Run generate_pub_keys.sh in that directory
3. Provide required VM images in vm_images
4. Run generate_scripts.sh in scripts/gen
5. Run terraform init && terraform apply in terraform/00_from_local
6. Wait until the access machine is up and connect to it via SSH: ssh -i <path_to_default.pem> -o "StrictHostKeyChecking=no" ubuntu@<access_fip>

From the access machine

1. Make sure that in scripts/gen/constants.sh there is BRANCH="origin/master" (unless you're testing changes from a different branch)
2. Run generate_scripts.sh in that directory
3. Run terraform init && terraform apply in terraform/01_from_access

From the kube-master machine

1. Copy ~/.kube/config to your local machine, put in the same path. It will allow remote connection to the Kubernetes cluster

Back to the local machine

1. Change server's value in the kube config. Currently there's master's private IP but since we want to connect from outside of that network we need to provide its public IP
2. Confirm that you're able to use kubectl from your local machine: kubectl get nodes
3. Confirm that the command returned a list containing kube-master and some kube-worker-{id}, their state should be Ready; wait some time if it's not

Setup within Kubernetes cluster

It can be performed from your local machine or from kube-master

microservices-demo

git clone git@github.com:k8loud/microservices-demo.git
cd microservices-demo

kubectl apply -f deploy/kubernetes/manifests
kubectl apply -f deploy/kubernetes/manifests-monitoring

If there are no SockShop dashboards available in Grafana

kubectl delete job/grafana-import-dashboards
kubectl apply -f deploy/kubernetes/manifests-monitoring

Metrics-Deployment

git clone git@github.com:k8loud/Metrics-Deployment.git
cd Metrics-Deployment

kubectl apply -f volume-creation/
kubectl apply -f manifests/