8:02 p.m.
I am trying to eat own dog food. Infrastructure as code...
So.. I have taken three physical servers. Setup all the base packages, set
NICs, NTP, DNS, collected disk UUIDs etc..
The servers are prepped.
And one server has cockpit and ovirt-engine installed... and service
started.
#############################
---
# This task takes nodes prepaired and deploys cockpit, ovirt
deployment and scripts cluster initilization. ONLY RUN THIS ON FIRST
NODE OF CLUSTER
# v0.001 20210222
# file: /roles/task_hci_setup.yml
# Documentation
# Overview:
https://www.ovirt.org/documentation/gluster-hyperconverged/chap-Deploying...
- name: Check if rpm ovirt-release44 is already installed
yum:
list: ovirt-release44
register: yum_list
- name: Conditionally do next thing
debug:
msg: "ovirt-release44 Not installed"
when:
yum_list.results | selectattr("yumstate", "match",
"installed") | list
| length == 0
# Add oVirt Repo manual.
- name: Download oVirt Repo RPM
get_url:
url: https://resources.ovirt.org/pub/yum-repo/ovirt-release44.rpm
dest: ~/ovirt-release44.rpm
mode: '0440'
when:
yum_list.results | selectattr("yumstate", "match",
"installed") | list
| length == 0
- name: Install Python3 as this is needed for oVirt.
yum:
name:
- python3
state: latest
become: true
- name:
For some reason install of oVirt Repo fails as package so have to run as shell
command: "rpm -ivh ~/ovirt-release44.rpm"
when:
yum_list.results | selectattr("yumstate", "match",
"installed") | list
| length == 0
- name: With new oVirt repo.. install rest of HCI dependancies
yum:
name:
- cockpit-ovirt-dashboard
- vdsm-gluster
- ovirt-host
- ovirt-ansible-roles
- ovirt-ansible-infra
state: latest
become: true
# Set variable for first node in cluster list to set as ignition master
- name: Only on first node thor deploy ovirt-engine-appliance
yum:
name:
- ovirt-engine-appliance
state: latest
become: true
when: "'{{ hci_ignition_node }}' in inventory_hostname"
# on local ansible host install galaxy modules
# Broken: I need a means to check if installed. If so .. don't install again.
# - name: Install Galaxy Modules on local ansible host
# local_action:
# module: shell
# _raw_params: ansible-galaxy install oVirt.ovirt-ansible-roles
# # when: <check for something to show galaxy installed module
- name: Start cockpit service
service:
name=cockpit
state=started
enabled=yes
when: "'{{ hci_ignition_node }}' in inventory_hostname"
################################
####### HCI Deploy #############
################################
# https://www.openvirtualization.pro/rhv-ovirt-ansible-getting-started/
#############################
What I am missing is the logic and though to get the last section of
playbook.. Which is ... what I would do in the wizard... ... but "step away
from the mouse".
# Ansible section: from bastion node (laptop windows 10 WSL with ansible
and galaxy installed
# 1) build playbook from my site variables... based on some ovirt master
sharing template of what that UI builds
# 2) push the playbook that installs ansible on the first node. Install
ansible as a first node... Run ansible playbook from customized template
file
# 3) Create vlans.
# 4) Setup new VM gluster shared volume from three other disk
# 5) Push a template VM into cluster, attached to new VLAN and attached to
new volume
#6) Deploy VM from this template
#7) Drink a cold beer
###############################
Question:
1) Can you do ansible deployment from "bastion host" or does the ansible
have to run on the node that is installed with the "ovirt-engine". If so..
how do you collect feedback etc..
2) I assume there is a template. Maybe it is published but I could not
find it. That I replace (or match) variables of my site that fill in that
file.
3) I installed galaxy oVirt.ovirt-ansible-roles but.. how do I know it
is deployed on my bastion host? I poked around in the ansible folder.. not
seeing anything. And I assume this only is of help POST the above seven
steps being done. But maybe this has deployment components and I am
missing documents.
Thanks,
--
p <jeremey.wise(a)gmail.com>enguinpages