Install Gluu CN with Rancher or RKE

#Overview

This is a tutorial to walk you through installing Gluu cloud native edition on RKE on a ubuntu 20.04.

#Instructions

1. Create instances to cover at minimum 8 cpu and 16 GB ram.

2. Follow instructions here to install RKE. Please do not run ./rke up until next step is executed.

3. You should have generated a cluster.yml following instructions above. Please open that file and add the following under the services --> kubelet --> extra_binds, as we use OpenEBS for volumes on RKE.

   services:
     kubelet:
       extra_binds:
        - /var/openebs/local:/var/openebs/local
        - /etc/iscsi:/etc/iscsi
        - /bin/iscsiadm:/bin/iscsiadm
        - /var/lib/iscsi:/var/lib/iscsi
        - /lib/modules
        - /var/openebs/sparse:/var/openebs/sparse
        - /var/openebs/local:/var/openebs/local
        - /mnt/openebs/local:/mnt/openebs/local
        - /opt/openebs/local:/opt/openebs/local

   Copy

4. Now you can bring your cluster up ./rke up.

5. You should have a running Kubernetes cluster now. You may add or remove nodes as necessary.

6. Follow instructions here to install openEBS.

7. Make sure that the instances has permissions as required by RKE to fully communicate with the intended cloud.

8. Download pygluu-kubernetes.pyz. This package can be built manually.

9. Optional: If using couchbase as the persistence backend. Download the couchbase kubernetes operator package for linux and place it in the same directory as pygluu-kubernetes.pyz

10. Run :

    ./pygluu-kubernetes.pyz install

    Copy

#Exposing the UI:

Exposing the UI is not necessary but you may do so by following these steps:

1. Create a single point of entry ( Load balancer) for your cluster. Please note that if you have given the instances permissions to act on the cloud the load balancer should have already been created and hence the following steps are not needed, otherwise a NodePort would have been created and you would need to follow with the next steps..

2. Get the port number nginx is using for forwarding 4.4 connections, here that would be 31822

   kubectl get svc -n ingress-nginx
    NAME                                          TYPE           CLUSTER-IP     EXTERNAL-IP   PORT(S)                      AGE
    default-http-backend                          ClusterIP      10.43.60.162   <none>        80/TCP                       131m
    ningress-ingress-nginx-controller             LoadBalancer   10.43.55.137   <pending>     80:30925/TCP,4.4:31822/TCP   18m
    ningress-ingress-nginx-controller-admission   ClusterIP      10.43.8.231    <none>        4.4/TCP                      18m

   Copy

   Forexample, Using AWS UI create an LB with TCP port 4.4 pointing to your nodeport i.e TCP 31822. If this is a FQDN create an https listener rule instead and attach your arn certificate. If you used TCP, add to your local computer ip of your loadbalancer and map it to your FQDN, then head to your browser at https://<gluu-FQDN> i.e https://demoexample.gluu.org

3. You must add a record to point your FQDN to the loadbalancer / ip used as a the single point of entry.

#Non-FQDN setup:

oxTrust will likely fail as it calls to check the ssl certificate. You will have to make sure to add to oxtrust statefulset a hostAlias entry to point your FQDN to the ip of the loadbalancer.

#Using Rancher

1. Follow instructions here to install Rancher in the environment of your choice

2. In the Rancher UI ,create your cluster by going to Global --> Add Cluster and choosing from any of the providers, AKS, EKS, GKE, EC2, Azure, Linode, and vSphere

3. After the cluster has been created, copy kubeconfig to a vm or laptop with kubectl , helm and python3 installed.

4. Download pygluu-kubernetes.pyz. This package can be built manually.

5. Optional: If using couchbase as the persistence backend. Download the couchbase kubernetes operator package for linux and place it in the same directory as pygluu-kubernetes.pyz

6. Run :

   ./pygluu-kubernetes.pyz install

   Copy