OpenMOLE connect is a server meant to be deployed in a kubernetees cluster in order to provide a multi-user OpenMOLE service.
Here is the recepie to deploy a K3S cluster and deploy OpenMOLE connect on this cluster.
To complete the installation of the multi-user OpenMOLE server you'll need:
- a debian (or ubuntu) machine accesible via ssh and that is able to accept inbound connections on ports 80 and 443
- a root account on the machine
- a dns entry pointing to the IP of this machine
- optionnaly an account on an SMTP mail server accesible via start tls (more could be easily supported, please can ask for it) to send the notification to the users
- optionally a minio server to backup the user data stored in longhorn
The info needed are:
- MASTER_HOST: the dns name of the host mochine
- MASTER_USER: the name of the sudoer account on the machine
- EMAIL_SERVER: the dns name of the smtp server
- EMAIL_PORT: the port of the smtp start tls port (generally port 587)
- EMAIL_USER: the user to login to the email server
- EMAIL_PASSWORD: the password to login to the email server
- EMAIL_SENDER_ADDRESS: an email sender address
- MINIO_URL: the URL of the minio service
- MINIO_ACCESS_KEY: a minio API key
- MINIO_SECRET_KEY:: a minio API key secret
Install Helm on your machine
Instantiate a machine and make sure to open all the ports in your security group if you are on openstack.
Deploy the server node:
ssh ${MASTER_USER}@${MASTER_HOST}
sudo apt install open-iscsi # requiered for longhorn
curl -sfL https://get.k3s.io | INSTALL_K3S_VERSION="v1.30.6+k3s1" INSTALL_K3S_EXEC="server" sh -s -
Test that the server works. Copy the file /etc/rancher/k3s/k3s.yaml on your machine. In the file, replace the master IP address with MASTER_HOST address.
export KUBECONFIG=$PWD/k3s.yml
kubectl get node
helm repo add kubernetes-dashboard https://kubernetes.github.io/dashboard/
helm upgrade --install kubernetes-dashboard kubernetes-dashboard/kubernetes-dashboard --create-namespace --namespace kubernetes-dashboard
Create a file named token.yml on your local computer, and fill it with the following content:
apiVersion: v1
kind: ServiceAccount
metadata:
name: admin-user
namespace: kubernetes-dashboard
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: admin-user
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: cluster-admin
subjects:
- kind: ServiceAccount
name: admin-user
namespace: kubernetes-dashboard
Then you can create an token an a proxy for the dashboard:
kubectl create -f token.yml
kubectl -n kubernetes-dashboard create token admin-user
kubectl -n kubernetes-dashboard port-forward svc/kubernetes-dashboard-kong-proxy 8443:443
The dashboard is accesible from within your local browser at https://localhost:8443.
Longhorn is the storage system used by connect.
Install longhorn with the following command:
kubectl apply -f https://raw.githubusercontent.com/longhorn/longhorn/v1.7.1/deploy/longhorn.yaml
The installation documentation for longhorn can be found here.
Accesing the longhorn UI:
kubectl port-forward -n longhorn-system svc/longhorn-frontend 8080:80
helm repo add jetstack https://charts.jetstack.io --force-update
helm install cert-manager jetstack/cert-manager --namespace cert-manager --create-namespace --version v1.16.0 --set crds.enabled=true
Then apply the following yaml:
apiVersion: cert-manager.io/v1
kind: ClusterIssuer
metadata:
name: letsencrypt-prod
spec:
acme:
server: https://acme-v02.api.letsencrypt.org/directory
email: [email protected]
privateKeySecretRef:
name: letsencrypt-prod
solvers:
- http01:
ingress:
class: traefik
Generate a sercret and a salt at random and store back them up somewhere. You may want to use pwgen 20 to get two random strings.
apiVersion: v1
kind: Namespace
metadata:
name: openmole
labels:
name: openmole
---
apiVersion: v1
kind: ServiceAccount
metadata:
name: openmole-admin-sa
namespace: openmole
---
apiVersion: rbac.authorization.k8s.io/v1
kind: Role
metadata:
name: openmole-admin-role
namespace: openmole
rules:
- apiGroups:
- "*"
resources:
- "*" #["pods", "deployments" ]
verbs:
- "*" #["list"]
---
apiVersion: rbac.authorization.k8s.io/v1
kind: RoleBinding
metadata:
name: openmole-role-binding
namespace: openmole
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: Role
name: openmole-admin-role
subjects:
- kind: ServiceAccount
name: openmole-admin-sa
namespace: openmole
---
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
name: openmole-storage
namespace: openmole
provisioner: driver.longhorn.io
allowVolumeExpansion: true
parameters:
numberOfReplicas: "1"
dataLocality: "best-effort"
staleReplicaTimeout: "2880" # 48 hours in minutes
fromBackup: ""
---
apiVersion: v1
data:
config.yml: |
salt: "some-salt" # replace with the random salt
secret: "some-secret" # replace with the ranodm secret
kube:
storageClassName: openmole-storage
storageSize: 20480
openmole:
versionHistory: 5
minimumVersion: 17
smtp:
server: "smtp.mydomain.net" # replace with EMAIL_SERVER
port: 587 # replace with EMAIL_PORT
user: user # replace with EMAIL_USER
password: password # replace with EMAIL_PASSWORD
from: [email protected] # replace with EMAIL_SENDER_ADDRESS
shutdown:
days: 30
checkAt: 8
remind: [1, 3, 7]
kind: ConfigMap
metadata:
creationTimestamp: null
name: connect-config
namespace: openmole
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: connect-pv-claim
namespace: openmole
spec:
storageClassName: openmole-storage
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 20Gi
---
apiVersion: v1
kind: Pod
metadata:
name: openmole-connect
namespace: openmole
labels:
app.kubernetes.io/name: openmole-connect
spec:
containers:
- name: openmole-connect
image: openmole/openmole-connect:1.0-SNAPSHOT
imagePullPolicy: "Always"
args:
- --config-file
- "/etc/connect-config/config.yml"
ports:
- containerPort: 8080
volumeMounts:
- name: connect-config
mountPath: "/etc/connect-config"
readOnly: true
- mountPath: "/home/demiourgos728"
name: connect-pv-storage
securityContext:
fsGroup: 1001
volumes:
- name: connect-config
configMap:
name: connect-config
- name: connect-pv-storage
persistentVolumeClaim:
claimName: connect-pv-claim
serviceAccountName: openmole-admin-sa
---
apiVersion: v1
kind: Service
metadata:
name: openmole-connect-service
namespace: openmole
spec:
selector:
app.kubernetes.io/name: openmole-connect
ports:
- protocol: TCP
port: 8080
nodePort: 30080
type: NodePort
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: openmole-connect-ingress
namespace: openmole
annotations:
cert-manager.io/cluster-issuer: letsencrypt-prod
spec:
rules:
- host: host.mydomain.ext # replace with MASTER_HOST
http:
paths:
- backend:
service:
name: openmole-connect-service
port:
number: 8080
path: /
pathType: Prefix
tls:
- hosts:
- host.mydomain.ext # replace with MASTER_HOST
secretName: letsencrypt-prod
You can configure a minio server to be able to backup your service.
To the get the base64 info:
echo -n <URL> | base64
echo -n <Access Key> | base64
echo -n <Secret Key> | base64
--------------------------------------------------------
echo -n http://192.168.1.66:9000/ | base64 # replace with MINIO_URL
echo -n t6dstDsYQuf7pbSj | base64 # replacet with MINIO_ACCESS_KEY
echo -n Q9RPWKxZ4bUIqBSzuOY6TLkFkXcHszRU | base64 # replace with MINIO_SERCRET_KEY
Replace the base64 info in the followin yaml and apply it:
apiVersion: v1
kind: Secret
metadata:
name: longhorn-minio-credentials
namespace: longhorn-system
type: Opaque
data:
AWS_ENDPOINTS: aHR0cDovLzE5Mi4xNjguMS42Njo5MDAwLw==
AWS_ACCESS_KEY_ID: dDZkc3REc1lRdWY3cGJTag==
AWS_SECRET_ACCESS_KEY: UTlSUFdLeFo0YlVJcUJTenVPWTZUTGtGa1hjSHN6UlU=
Then you can use longhorn-minio-credentials as a credential in your backup field of longhorn UI
When your node is full you can add a node to the cluster.
Deploy an k3s node. Find the k3s token on the head node /var/lib/rancher/k3s/server/node-token. Store it in K3S_TOKEN environment variable on the node.
sudo apt install open-iscsi # requiered for longhorn
curl -sfL https://get.k3s.io | K3S_URL=https://${MASTER_HOST}:6443 K3S_TOKEN="${K3S_TOKEN}" sh -s -
Check that you have 2 nodes now.