Gridwalls - the game

What's more fun than learning stuff by implementing a game?

Current plan:

• Make a cool game
• Use microservices, docker and other stuff

Lots of commands goes here. TODO.

kubeadm init --apiserver-cert-extra-sans=<server public ip> --pod-network-cidr=10.244.0.0/16

From: https://github.com/kubernetes/helm/blob/master/docs/rbac.md

kubectl create namespace gw
kubectl create serviceaccount tiller --namespace gw
kubectl create clusterrolebinding tiller --clusterrole=cluster-admin --serviceaccount=gw:tiller
helm init --service-account tiller --tiller-namespace=gw
echo "export TILLER_NAMESPACE=gw" >> ~/.bashrc

helm install stable/docker-registry
POD_NAME=$(kubectl get pods --namespace gw -l "app=docker-registry,release=bold-cat" -o jsonpath="{.items[0].metadata.name}")
kubectl port-forward $POD_NAME 5000

Here we have 1 master and 1 slave.

On master host:

sudo echo " gluster2" >> /etc/hosts # replace with node ip

On slave host:

sudo echo " gluster1" >> /etc/hosts # replace with master ip

(Some steps from https://docs.gluster.org/en/latest/Install-Guide/Install/ TODO)

Source: https://serverfault.com/questions/204082/using-ntp-to-sync-a-group-of-linux-servers-to-a-common-time-source

Follow instructions here: http://gluster-documentations.readthedocs.io/en/master/Install-Guide/Quick_start/

Currently, that means doing (with a few deviations because no docs are pertfect):

apt update -y && apt upgrade -y

fdisk /dev/vdb # and create a single partition with n, e(xtended), the rest are deafult settings. Then w(rite and quit)
apt install -y xfsprogs # To install mkfs.xfs
mkfs.xfs -f -i size=512 /dev/vdb
mkdir -p /export/vdb && mount /dev/vdb /export/vdb && mkdir -p /export/vdb/brick
echo "/dev/vdb /export/vdb xfs defaults 0 0"  >> /etc/fstab

From http://gluster-documentations.readthedocs.io/en/master/Install-Guide/Install/:

apt install -y software-properties-common # To make add-apt-repository work
apt install -y python-software-properties
add-apt-repository ppa:gluster/glusterfs-4.0
apt-get update -y
apt install -y glusterfs-server

helm install stable/nginx-ingress --set rbac.create=true --set controller.hostNetwork=true --set controller.service.type=ClusterIP #--set controller.service.externalIPs={51.15.77.186} --set controller.service.nodePorts.http=80

Then make sure your ingress host is set correctly.

• (Opt) Watch pods: watch kubectl get po
• (Opt) Watch helm releases: watch helm list
• If using custom cloud (not updated, probably will fail)
  • export PRIVATE_DOCKER_REGISTRY=localhost:5000
  • make -C source/lib/helm_charts package host
  • Run helm repo
• If using gcloud, etc
  • ./run.sh