Openshift 4 vmware infra provisioner

This Role wants to provide the simplest way to install openshift 4 taking care of all the network configuration but DNS.

Features:

• You don't need DHCP
• check for your environment
• Validate your configuration
• Validate your DNS
• Configure an HaProxy (not deploy it, only configure ATM)
• Configure an Nginx web server (not deploy it, only configure ATM)
• Build the iPXE iso image
• Create the VM
• Boot the vm injecting network configuration
• Install the cluster

Additional documentation:

• Supported Methods
• tested Versions
• Known Problems
• Known Limitations

TODO:

• Check for yout environment
• Validate your configuration
• Validate your DNS
• Configure an HaProxy (not deploy it, only configure ATM)
• Configure an Nginx web server (not deploy it, only configure ATM)
• Build the iPXE iso image
• Create the VM
• Boot the vm injecting network configuration
• Install the cluster
• Calculate signature for fcos/rhcos
• Test all three methods with FCOS (OKD)
• Test all three methods with RHCOS (Openshift)
• Determine and download images automatically and build a fetch url
• Multiple interfaces transpile
• Proxy implementation
• Validate minimal requirements
• Deploy VM for HaProxy and NginX
• Build HA/Vrrp for HaProxy/NginX

Requirements

A Bastion node where execute thos tasks and where install a temporary webserver Rhel/Centos 8 Ansible 2.9

Role Variables

# Please lookup the variables in 
echo defaults/main.yml

Dependencies

This playbook is tested on RHEL8 and Centos8 with Ansible 2.9 All the dependencies are managed via dnf/yum.

Example Playbook

---
- hosts: all
  remote_user: root
  roles:
    - openshift4_infra_provisioner

Example Inventory

---
haproxy:
  hosts:
    bastion.ocp.okd01.lab.local
webserver:
  hosts:
    bastion.ocp.okd01.lab.local
all:
    vars:
      enable_reservation: False
      dns_fail_is_not_fatal: False
      ptr_fail_is_not_fatal: True
      checks_only: False
      #... look for the entire set of variables in defaults/main.yml ...

To run the installation please fill the variables

Check Installation status:

openshift-install \
  --dir={{playbook_dir}}/tmp/cluster_conf wait-for bootstrap-complete \
  --log-level=info

To configure the OC client export this variable: '

export KUBECONFIG={{playbook_dir}}/tmp/cluster_conf/auth/kubeconfig

oc login \
  -h api.{{ocp_installer.clustername}}{{ocp_installer.basedomain}} \
  -u kubeadmin \
  -p {{ lookup('file', playbook_dir + '/tmp/cluster_conf/auth/kubeadmin-password') }}

To approve the pending nodes you should run this (you may need to do this multiple times): "

oc get csr -o go-template='{{range .items}}{{if not .status}}{{.metadata.name}}{{"\n"}}{{end}}{{end}}' | xargs oc adm certificate approve

To configure the registry for non production cluster:'

oc patch configs.imageregistry.operator.openshift.io cluster --type merge --patch '{"spec":{"managementState":"Managed"}}'`
oc patch configs.imageregistry.operator.openshift.io cluster --type merge --patch '{"spec":{"storage":{"emptyDir":{}}}}'