/OKD_cluster

Deploys OKD cluster using vagrant

Primary LanguageShell

OKD_cluster

Deploys OKD cluster using vagrant and virtualbox.
Installation steps are based on this Spirited Engineering post

VMs deployment

Linux:

Validated with the following software versions:

  • Debian 10.7
  • Linux 4.19.0-13-amd64
  • Virtual Box 6.1.16-140961~Debian~buster
  • Vagrant 1:2.2.10

Execute from virtualbox host server:

host:repo_path$ vagrant up

Windows:

repo_path> vagrant.exe up

Validated with the following software versions:

NOTE: Some rsync problems have been detected with different Vagrant version than 2.2.5.

It takes long time to set up the cluster (around 20 minutes). After this, it's possible to check the openshift cluster nodes with:

Webconsole access:

Add to the hosts file the following entry (Update it with the corresponding one if it has been changed in config file file for master node)

192.168.50.11 master.okd.local

to access to the web console using the following URL:

https://master.okd.local:8443/

  • (NOTE I) A self-signed certificate warning may appears in the browser and it has to be accepted.
  • (NOTE II) Default credentials provisioned in these deployment scripts are admin / okdadmin123)

Troubleshooting:

  • VMs access:

host$ vagrant ssh master
host$ vagrant ssh infra
host$ vagrant ssh compute

  • Ansible connectivity tests (from master):

master$ sudo su -
master# ansible all -m ping

  • Manual ansible openshift deployment execution (from master):

master# cd /usr/share/ansible/openshift-ansible/playbooks/
master# ansible-playbook prerequisites.yml
master# ansible-playbook deploy_cluster.yml

  • Check openshift nodes (from master)

master# oc get nodes

NAME      STATUS    ROLES     AGE       VERSION  
compute   Ready     compute   23m       v1.9.1+a0ce1bc657  
infra     Ready     <none>    23m       v1.9.1+a0ce1bc657  
master    Ready     master    23m       v1.9.1+a0ce1bc657