How to deploy a new Virtual Machine with PyVSS

Owned by JM Lopez
Last updated: May 01, 2020
2 min read

Deploying a virtual machine interacting directly with thee ITS Private Cloud API could somehow be overwhelming due to its multiple resources to check such as networks, folders, disks, etc. In order to simplify the interaction with the API we created a Python Client named pyvss with many self-descriptive methods  and available in the Python Package Index. 

This tutorial uses pyvss to create a virtual machine with no operating system and which will be installed manually by mounting an ISO image from the VSS Catalog. The full specification is listed below:

  • name: VMfromPyVss
  • built process: os install
  • client department: VSS
  • description: virtual machine created from vss
  • usage: QA
  • cpu and memory: 2 
  • disks: 1GB, 2GB and 3GB
  • networks: any available network in your account
  • folders: any folder available in your account
  • iso: ubuntu 12.04 i386
  • os: ubuntuGuest

Pre-requirements

  • Python 2.7.10+
  • Ptpython
  • Access to the ITS Private Cloud API

  • Python client for the ITS Private Cloud API: pyvss

    pip install pyvss

  • ptpython (optional) - usual python console can be used

    pip install ptpython

  • httpie (optional) - curl or wget can be used as well

    pip install httpie

Step-by-step guide

  1. Generate an API access token and save it to tk.json file:

    http POST https://vss-api.eis.utoronto.ca/auth/request-token -a jm > tk.json
http: password for jm@vss-api.eis.utoronto.ca:

  2. Start ptpython or python console:

    ptpython
  3. Import VssManager and json libraries and load tk.json into a variable


    from pyvss.manager import VssManager
import json 
 
token = json.load(open('tk.json'))
  4. Create a VssManager instance with the previously loaded token:


    vss = VssManager(tk=token.get('token'))
  5.  Refer to the official docs or use the help built in function to display what parameters are required by the create_vm function:


    help(vss.create_vm)

  6. Define the following variables

    name = 'VMfromPyVss'
built = 'os_install' # it could be either clone or image as well
client = 'VSS'
description = 'Virtual machine created from pyVss'
usage = 'QA' # could be either Prod, Test, QA or Dev
cpu = 2
memory = 2 # this value is in GB
disks = [1, 2, 3]  # three disks of 1GB, 2GB and 3GB

  7. Get target folder, networks, operating system and iso image to mount for the installation

    networks = vss.get_networks(name='1102') # filtering by name
networks = [{'network': net.get('moref')} for net in networks]
 
folders = vss.get_folders(name='APIDemo') # filtering by name
folder = folders[0].get('moref')

isos = vss.get_isos(name='ubuntu') # filtering by name
iso = isos[0].get('path')

os_list = vss.get_os(name='ubuntu') # filtering by name
os = os_list[0].get('guestFullName')
 

  8. Create new vm request 

    r = vss.create_vm(name=name, os=os, built=built, client=client, description=description, 
				  folder=folder, networks=networks, disks=disks, usage=usage, 
				  cpu=cpu, memoryGB=memory, iso=iso, power_on=True)

  9. We will use wait_for_request function to get new virtual machine Uuid when provisioned:

    vm_id = vss.wait_for_request(r.get('_links').get('request'), 'vm_moref', 'Processed')

  10. Once you get the id, get a summary of the VM:

    vm = vss.get_vm(vm_id)
vm.get('state').get('power_state')
u'poweredOn'

  11. Generate a one-time VM vSphere client link to start the OS installation

    link = vss.get_vm_vsphere_link(vm_id)