Something to enjoy with Prim Pro (X-Play) and Python

Something to enjoy with Prim Pro (X-Play) and Python

• This article is the 17th day article of Nutanix Advent Calendar 2020.
• Keywords: Nutanix, python, X-Play, automation

Introduction

• Recently, I used to script it because I often typed the same command as getting Nutanix health after work.
• The idea is to combine that script with Prism Pro's X-Play for even more fun.
• It would be great if you could see this content and get an image of how Prism Pro and scripts work together.

Thing you want to do

• When an alert occurs in Nutanix, various information is automatically fetched from CVM.
• In other words, kick the script from X-Play
• ↓ I want you to get such a result automatically. I'm told to get support anyway

===========< id
uid=1000(nutanix) gid=1000(nutanix) groups=1000(nutanix),160(snmp),600(java),990(docker)context=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023

===========< uname -a
Linux ntnx-20sm6k250202-c-cvm 3.10.0-1062.12.1.el7.nutanix.20200216.cvm.x86_64 #1 SMP Sun Feb 16 07:02:50 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

===========< hostssh date
============= 10.55.87.27 ============
Sat Dec 12 11:17:09 UTC 2020
============= 10.55.87.25 ============
Sat Dec 12 11:17:10 UTC 2020
============= 10.55.87.26 ============
Sat Dec 12 11:17:10 UTC 2020

===========< hostssh uptime
============= 10.55.87.27 ============
 11:17:12 up 1 day,  9:40,  0 users,  load average: 1.90, 2.36, 2.38
============= 10.55.87.25 ============
 11:17:12 up 1 day,  9:41,  0 users,  load average: 6.04, 6.07, 5.98
============= 10.55.87.26 ============
 11:17:12 up 1 day,  9:41,  0 users,  load average: 1.47, 1.69, 1.80

===========< ncli cluster info
    Cluster Id                : 0005b626-72dd-8051-3f5b-ac1f6b3d7d4b::4565431898213285195
    Cluster Uuid              : 0005b626-72dd-8051-3f5b-ac1f6b3d7d4b
    Cluster Name              : RTP-POC087
    Cluster Version           : 5.15
    Cluster Full Version      : el7.3-release-euphrates-5.15-stable-4fbdd4d9de331230bb468b3549f530e80ab53bb9

・ ・ ・ The following is omitted

What is X-Play?

• before that
• X-Play is one of the operation support functions that realizes easy automation, and is a function provided under the Prism Pro license.
• Prism Pro is packed with features that make it easier to operate.
• It solves the problems shown in this figure. Today, I would like to talk about the automation in red.

X-Play setting method image

• Regarding the setting image of X-Play, you may be able to imagine the video around here as a demo. It would be nice to be able to create an automation mechanism without coding.

• https://www.youtube.com/watch?v=UmAWG0UHpuU

• The nice thing about Nutanix is ​​that you can understand it even if you don't understand English.

• X-Play can automatically take various actions triggered by alerts and schedules on Nutanix. You can also combine it.

• Various things (typical). I can do more and more every day. When I saw it a while ago, it increased again.

• VM power ON / Off

• Take a snapshot of the VM

• Add or remove resources

• Send email / Notify Slack

• Report creation

• etc

• Furthermore, you can execute scripts as shown below, so you can do anything depending on the combination. So for Linux, command from ssh. In the case of Windows, it feels like you can run Powershell, so you can kick python in it.

• SSH to command execution

• Run Powershell

• Run REST API

• The figure below is a playbook that "takes a backup (snapshot)", "adds CPU", and "emails the administrator" when the "VM CPU usage alert" goes up.

• Just click from this list of actions

• You can do this kind of flow. Super easy.

Script linkage (Powershell)

List of commands you want to get

• Now, let's create a script that links the script.
• Here, I would like to fetch the commands that are often executed by AOS according to the trigger.
• Commands you want to take at once (again)

id
uname -a
hostssh date
hostssh uptime
ncli cluster info
ncli host ls
ncli cluster get-domain-fault-tolerance-status type=node
ncli alert history duration=14
ncli license get-license
ncli license get-allowances
cluster status
zeus_config_printer

• There was something else I wanted to take, but can't I get environment variables with Python's Paramiko? There were some commands that I couldn't get well. Like nodetool -h 0 ring.
• You have to write with the full path.
• Hmm, incomplete
• After all, write this list with the full path. Hmm. Well, let's make it because it's a big deal

id
uname -a
/usr/local/nutanix/cluster/bin/hostssh date
/usr/local/nutanix/cluster/bin/hostssh uptime
/home/nutanix/prism/cli/ncli cluster info
/home/nutanix/prism/cli/ncli host ls
/home/nutanix/prism/cli/ncli cluster get-domain-fault-tolerance-status type=node
/home/nutanix/prism/cli/ncli alert history duration=14
/home/nutanix/prism/cli/ncli license get-license
/home/nutanix/prism/cli/ncli license get-allowances
/usr/local/nutanix/cluster/bin/cluster status
/usr/local/nutanix/bin/zeus_config_printer

Script to execute

• Script overview

• Get the list of commands listed from the Nutanix cluster

• Divert the script created before https://qiita.com/matsumo2019/items/0713dd70bcdc91e2690f Prepare by referring to this page.

• Preparation items (super omitted)

• Python installation

• Install paramiko

• Script to create (Refer to the above article. The script is put at the end.)

• get-cvm.py: Main processing

• my_command.py: Classified paramiko processing

• common.json (configuration file. CVM IP, ID, password)

• command_list.txt (write a list of commands to be executed) * Basically edit only this file

• By the way, I had never touched Python, but I could make it by looking at this book.

• Learn the basics in just one day! Introduction to Python https://a.r10.to/hbHMiv

Arrangement of various scripts

• This time I checked it on Windows for some reason.
• Follow the steps below
• Python installation
• Install paramiko with pip
• Placement of various scripts (located in C: \ scripts)
• For Windows, make sure you can run Powershell remotely.

X-Play settings (Prism Central)

• Setting flow

• From Prism Central, select "Menu"> "Operations"> "Playbook"

• "Create Playbook"

• Trigger selection / setting

• Action selection / setting

• Repeat adding actions

• This is a specific setting screen.

• "Create Playbook"

• Some are created by default, so you can refer to them.

• "Trigger selection"

• Select "Alerts Matching Criteria".

• "Trigger selection"

• Select "Alerts Matching Criteria" and select the alert condition.

• "Select Action"

• Select "VM Powershell".

• Set the following.

• Target VM: The server that executes the script * Username * Password

• Path to Script: Specify the script on the server to be executed in the format including the path

• Save and save with a playbook name.

• Enable status if you want to enable it immediately

• This completes the X-Play settings.

The result of calling Powershell from X-Play

• The command result was obtained with the alert as a trigger.

• Execution result (Prism Central side)

• You can check the execution result and error of the playbook.

• You can also see the error. I'm having trouble connecting in this case (it was Windows Firewall ...)

• Execution result (script server side)

• When an event occurs like this, the execution result is automatically saved.

• Contents of the execution result file (as expected)

Summary

• Using X-Play of Prism Pro, you can create an automation mechanism without coding. (It's much easier than what I wrote here)
• If you combine it with your own script, the processing will expand infinitely.
• Do you not issue commands manually every time? I don't have Prism Pro.

bonus. Actual script

• command_list.txt
• You can get the results of various commands by editing this text.

id
uname -a
/usr/local/nutanix/cluster/bin/hostssh date
/usr/local/nutanix/cluster/bin/hostssh uptime
/home/nutanix/prism/cli/ncli cluster info
/home/nutanix/prism/cli/ncli host ls
/home/nutanix/prism/cli/ncli cluster get-domain-fault-tolerance-status type=node
/home/nutanix/prism/cli/ncli alert history duration=14
/home/nutanix/prism/cli/ncli license get-license
/home/nutanix/prism/cli/ncli license get-allowances
/usr/local/nutanix/cluster/bin/cluster status
/usr/local/nutanix/bin/zeus_config_printer

• cvm_info.json
• The connection information to CVM is written in JSON format. Oops ID/password direct writing

{
	"cluster_address": "10.55.87.31",
	"user_name": "nutanix",
	"cvm_password": "nutanix/4u"
}

• get_cvm_cmd.py
• A child of the main process. Please call this child. Please read from the comments.

import json
import my_command

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

#Read command list
fin = open('command_list.txt', 'r')
command = fin.readlines()
fin.close()

#Read server information
sv_info = json.load(json_open)
cluster_ip = sv_info['cluster_address']
user_name = sv_info['user_name']
password = sv_info['cvm_password']

##Command execution (main processing)
def main():
	for name in command:
		print('===========<', name)
		remote_obj = my_command.RemoteCommand()
		ret = remote_obj.get_df(cluster_ip, user_name, password, name)
		print(ret)

##Main processing (magic)
if __name__ == '__main__':
  main()

• my_command.py
• A class of processing of paramiko (Python library that sshes).

import paramiko

class RemoteCommand:
	def get_df(self, ip, user, password, command):
  		client = paramiko.SSHClient()
  		client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
  		client.connect(ip, username=user, password=password, timeout=3.0)
  		(stdin, stdout, stderr) = client.exec_command(command, get_pty=True)
  		output = stdout.read().decode()
  		client.close()
   	
   	    return output

• get_cvm_cmd.ps1
• I can't read Python directly from X-Play, so I tried to call it with powershell.
• Is Powershell and I didn't understand the date operation well, so I read more bat.

cmd.exe /c "c:\scripts\get_cvm_cmd.bat"

• get_cvm_cmd.bat
• I wanted to write it to a unique log file
• After all, I just want to type the last command.

@echo off
set yyyy=%date:~0,4%
set mm=%date:~5,2%
set dd=%date:~8,2%
 
set time2=%time: =0%
 
set hh=%time2:~0,2%
set mn=%time2:~3,2%
set ss=%time2:~6,2%
 
set filename=%yyyy%-%mm%%dd%-%hh%%mn%%ss%
 
cmd.exe /c "c:\scripts\get_cvm_cmd.py > logs\log_%filename%.txt

• Location

• Just put it in the same place.

• Reflection

• I couldn't set the environment variables, so I'm embarrassed that I couldn't write with the full path or use allssh.

• Around Powershell, I cut my hand ... Shame

• Tell someone