[PYTHON] Operate Nutanix with REST API Part 2

Operate Nutanix with REST API Part 2

• When it comes to large-scale environments, there are at least requirements for automation and scripting, and so far we have used scripts such as perl and PowerShell, SNMP, etc., but I think that the demand for REST APIs has been increasing recently.
• Let's check the REST API from a state where you are not familiar with it.
• Last time, I looked at what you can do with the REST API from the Prism screen.
• This time, I will try to operate with Python.

table of contents

• It is summarized in the following flow
• Look at the REST API with Prism (previous)
• Try to operate REST API with Python (this time)

Script overview

Purpose

• Access a specific cluster and view the results of the specified REST API
• Since the cluster information is often changed, specify it as a configuration file in a separate file.

Program flow

• Read the configuration file
• conf.json
• rest-api_list.txt
• Issue REST API for Cluster
• Output only specific keys separated by commas
• Output until there are no more elements

Make little by little

Script for checking the main process

• Copy what you picked up

import requests
import urllib3
from urllib3.exceptions import InsecureRequestWarning
urllib3.disable_warnings(InsecureRequestWarning)
import json

IP = '10.42.8.37'
USER = 'admin'
PASSWORD = 'hogehoge'

# (1) Make Session
session = requests.Session()
session.auth = (USER, PASSWORD)
session.verify = False                              
session.headers.update({'Content-Type': 'application/json; charset=utf-8'})

# (2) Make URL
url = 'https://{}:9440/PrismGateway/services/rest/v1/cluster'.format(IP)

# (3) Send request and get Response
response = session.get(url)

# (4) Check response code
print('Response Code: {}'.format(response.status_code))
print('Response OK?: {}'.format(response.ok))

# (5) Check response body
print('Response Body:')
print(response.text)

Script description

import process. Set when doing something using http communication

• HTTP communication library ** Requests **.
• Urllib3, a collection of modules that handle URLs
• Ignore my certificate
• I haven't used it yet, but I will use it anyway.

import requests
import urllib3
from urllib3.exceptions import InsecureRequestWarning
urllib3.disable_warnings(InsecureRequestWarning)
import json

Creating and initializing session objects

• IP / user name / password will be finally entered by another file or user. Defined as a constant for the time being
• Create a session object from the requests module and add information (user ID password, etc.) to it.
• The URL will be described later as a variable or a separate file during processing.
• Not yet sent here

IP = '10.42.8.37'
USER = 'admin'
PASSWORD = 'hogehoge'

# (1) Make Session
session = requests.Session()
session.auth = (USER, PASSWORD)
session.verify = False                              
session.headers.update({'Content-Type': 'application/json; charset=utf-8'})

# (2) Make URL
url = 'https://{}:9440/PrismGateway/services/rest/v1/cluster'.format(IP)

Session execution

• Actually communicate with session.get (url).
• Get the response code from the server for debugging and conditional branching.
• print (response.text) Check the acquired result.

# (3) Send request and get Response
response = session.get(url)

# (4) Check response code
print('Response Code: {}'.format(response.status_code))
print('Response OK?: {}'.format(response.ok))

# (5) Check response body
print('Response Body:')
print(response.text)

The completed script

• get_vms.py
• It's a little different from the above script, but it's just the final output.
• Although it does not come out cleanly with standard output, it can be used as CSV.
• Some comments are left for future debugging

import requests
import urllib3
from urllib3.exceptions import InsecureRequestWarning
urllib3.disable_warnings(InsecureRequestWarning)
import json
import base64

#Read file
fin = open('cvm_info.json', 'r')

#Read server information
# IP = '10.42.8.37'
# USER = 'admin'
# PASSWORD = 'pass'

sv_info = json.load(fin)
IP = sv_info['cluster_address']
USER = sv_info['user_name']
PASSWORD = sv_info['cvm_password']
fin.close()

auth_str = USER + ':' + PASSWORD
auth_str = api_head = {
    'Authorization': 'Basic ' + base64.b64encode(auth_str.encode('utf-8')).decode(),
}

# (1) Make Session
session = requests.Session()
session.auth = (USER, PASSWORD)
session.verify = False                              
session.headers.update({'Content-Type': 'application/json; charset=utf-8'})

# (2) Make URL
url = 'https://{}:9440/PrismGateway/services/rest/v2.0/vms/'.format(IP)

# (3) Send request and get Response
response = session.get(url, headers=api_head)

# (4) Check response code
# print('Response Code: {}'.format(response.status_code))
# print('Response OK?: {}'.format(response.ok))

# (5) Check response body
# print('Response Body:')
# print(response.text)


# (6) text -> dict
dict_data = json.loads(response.text)

# (7)Output VM list CSV
i = 0
while i < dict_data['metadata']['total_entities']:
    vms= (
        dict_data['entities'][i]['name'], 
        dict_data['entities'][i]['description'], 
        dict_data['entities'][i]['num_cores_per_vcpu'],
        dict_data['entities'][i]['num_vcpus'],
        dict_data['entities'][i]['memory_mb'], 
        dict_data['entities'][i]['uuid']
        )
    print(*vms,sep=',')
    i += 1

Supplement

• Basic authentication had to be encoded in utf-8

auth_str = USER + ':' + PASSWORD
auth_str = api_head = {
    'Authorization': 'Basic ' + base64.b64encode(auth_str.encode('utf-8')).decode(),
}

• Since the original amount of information was large, only the required information VM Name, description, number of cpu, number of memory, and uuid are displayed.
• Since the number of VMs was obtained by total_entities, it is rotated based on that.
• I passed it to the tuple type (vms) once, but it seems that I can write it more beautifully.
• If the REST API specifications change, you will have to modify it directly, so it lacks maintainability ...

# (6) text -> dict
dict_data = json.loads(response.text)

# (7)Output VM list CSV
i = 0
while i < dict_data['metadata']['total_entities']:
    vms= (
        dict_data['entities'][i]['name'], 
        dict_data['entities'][i]['description'], 
        dict_data['entities'][i]['num_cores_per_vcpu'],
        dict_data['entities'][i]['num_vcpus'],
        dict_data['entities'][i]['memory_mb'], 
        dict_data['entities'][i]['uuid']
        )
    print(*vms,sep=',')
    i += 1

Execution result

• Basic authentication had to be encoded in utf-8
• Since the original amount of information was large, only the required information VM Name, description, number of cpu, number of memory, and uuid are displayed.
• It seems that you can use it to get it after construction work

PS C:\Users\oresama>  get_aos_rest.py
Move VM,RX-Autodeployed-VM,4,1,8192,ffac0fb7-6a0b-45b1-a3ea-7c8079af2364
RXAutomationPC,NutanixPrismCentral,1,4,16384,f6817073-c941-4504-8034-a1d776b03a4c
Windows 2012 VM,RX-Autodeployed-VM,4,1,8192,23647618-2d45-46ac-8909-ea1444be6d94
Windows 10 VM,RX-Autodeployed-VM,4,1,8192,59a768a1-556d-4c17-bd84-71875a666d99
AutoDC VM,RX-Autodeployed-VM,4,1,8192,68eef805-3671-4717-99d6-5a07723da365
CentOS VM,RX-Autodeployed-VM,4,1,8192,cd9d44b4-db7a-4582-91f0-44d5d27aff2b

• By the way, the original data

{"metadata":{"grand_total_entities":6,"total_entities":6,"count":6,"start_index":0,"end_index":6},"entities":[{"allow_live_migrate":true,"gpus_assigned":false,"boot":{"disk_address":{"device_bus":"scsi","device_index":0},"boot_device_type":"disk"},"description":"RX-Autodeployed-VM","ha_priority":0,"memory_mb":8192,"name":"Move VM","num_cores_per_vcpu":4,"num_vcpus":1,"power_state":"off","timezone":"UTC","uuid":"ffac0fb7-6a0b-45b1-a3ea-7c8079af2364","vm_features":{"AGENT_VM":false,"VGA_CONSOLE":true},"vm_logical_timestamp":1,"machine_type":"pc"},{"allow_live_migrate":true,"gpus_assigned":false,"description":"NutanixPrismCentral","ha_priority":0,"host_uuid":"fab5dfe4-f92d-4ce8-b296-4790dd19f4c8","memory_mb":16384,"name":"RXAutomationPC","num_cores_per_vcpu":1,"num_vcpus":4,"power_state":"on","timezone":"UTC","uuid":"f6817073-c941-4504-8034-a1d776b03a4c","vm_features":{"AGENT_VM":false,"VGA_CONSOLE":true},"vm_logical_timestamp":2,"machine_type":"pc"},{"allow_live_migrate":true,"gpus_assigned":false,"boot":{"disk_address":{"device_bus":"scsi","device_index":0},"boot_device_type":"disk"},"description":"RX-Autodeployed-VM","ha_priority":0,"memory_mb":8192,"name":"Windows 2012 VM","num_cores_per_vcpu":4,"num_vcpus":1,"power_state":"off","timezone":"UTC","uuid":"23647618-2d45-46ac-8909-ea1444be6d94","vm_features":{"AGENT_VM":false,"VGA_CONSOLE":true},"vm_logical_timestamp":1,"machine_type":"pc"},{"allow_live_migrate":true,"gpus_assigned":false,"boot":{"disk_address":{"device_bus":"scsi","device_index":0},"boot_device_type":"disk"},"description":"RX-Autodeployed-VM","ha_priority":0,"host_uuid":"fee3e23a-2c0f-4a51-963e-289601935fd3","memory_mb":8192,"name":"Windows 10 VM","num_cores_per_vcpu":4,"num_vcpus":1,"power_state":"on","timezone":"UTC","uuid":"59a768a1-556d-4c17-bd84-71875a666d99","vm_features":{"AGENT_VM":false,"VGA_CONSOLE":true},"vm_logical_timestamp":2,"machine_type":"pc"},{"allow_live_migrate":true,"gpus_assigned":false,"boot":{"disk_address":{"device_bus":"scsi","device_index":0},"boot_device_type":"disk"},"description":"RX-Autodeployed-VM","ha_priority":0,"host_uuid":"fee3e23a-2c0f-4a51-963e-289601935fd3","memory_mb":8192,"name":"AutoDC VM","num_cores_per_vcpu":4,"num_vcpus":1,"power_state":"on","timezone":"UTC","uuid":"68eef805-3671-4717-99d6-5a07723da365","vm_features":{"AGENT_VM":false,"VGA_CONSOLE":true},"vm_logical_timestamp":2,"machine_type":"pc"},{"allow_live_migrate":true,"gpus_assigned":false,"boot":{"disk_address":{"device_bus":"scsi","device_index":0},"boot_device_type":"disk"},"description":"RX-Autodeployed-VM","ha_priority":0,"memory_mb":8192,"name":"CentOS VM","num_cores_per_vcpu":4,"num_vcpus":1,"power_state":"off","timezone":"UTC","uuid":"cd9d44b4-db7a-4582-91f0-44d5d27aff2b","vm_features":{"AGENT_VM":false,"VGA_CONSOLE":true},"vm_logical_timestamp":1,"machine_type":"pc"}]}

Summary

• I was able to get Nutanix Cluster information using the REST API.
• This time I learned the following. This time it was a data reference, but it seems that you can take a snapshot in the same way.
• Execution of REST API (URL operation)
• Formatting JSON data
• Nutanix manages data based on the uuid of each VM, so I think that it can be reused because the VM name and uuid can be obtained by this method.