Installing pyautogui on Ubuntu Server

The Issue

I was recently assisting a QA engineer with some #Fun™ problems on Jenkins where tests running pyautogui (running under Python 2.7) were failing because pyautogui was not installed. Every time we attempted to pip install pyautogui it would give an error similar to the following:

Downloading/unpacking pyautogui
Downloading PyAutoGUI-0.9.36.tar.gz (46kB): 46kB downloaded
Running setup.py (path:/tmp/pip-build-tbxln81q/pyautogui/setup.py) egg_info for package pyautogui
Traceback (most recent call last):
File "", line 17, in
File "/tmp/pip-build-tbxln81q/pyautogui/setup.py", line 6, in
version=import('pyautogui').version,
File "/tmp/pip-build-tbxln81q/pyautogui/pyautogui/init.py", line 115, in
from . import _pyautogui_x11 as platformModule
File "/tmp/pip-build-tbxln81q/pyautogui/pyautogui/_pyautogui_x11.py", line 160, in
_display = Display(os.environ['DISPLAY'])
File "/usr/lib/python3.4/os.py", line 633, in getitem
raise KeyError(key) from None
KeyError: 'DISPLAY'
Complete output from command python setup.py egg_info:
Traceback (most recent call last):

File "", line 17, in

File "/tmp/pip-build-tbxln81q/pyautogui/setup.py", line 6, in

The Fix

Ensure you have the following packaged installed: xserver-xorg, python-pip, python-xlib, xfce4 (you can use a different desktop env such as Gnome or KDE), and xauth.
sudo apt-get install -y xserver-xorg python-pip python-xlib xfce4 xauth
Also ensure that once you’ve install pip that you upgrade pip by running the following:
pip install --upgrade pip
Open a GUI session either through xrdp (if installed) or console. Once you are here run the following command:
echo $DISPLAY
Open an SSH connection to the server and set the $DISPLAY environment variable to match the output of the command above, in my case it was :10.0
export DISPLAY=:10.0
You can confirm this value is set by running the echo command again. Once this is in place run the following to install pyautogui
pip install pyautogui

Other Errors
If you receive any other errors around xauth or xserver try purging and reinstalling the packages above. You may also need to install a few other pip packages such as pillow and xlib.

Managing IIS With Puppet via Powershell DSC

Use Case:

Many of us in DevOps/Cloud/Site Reliability/System Engineering have long been using Puppet to manage our Apache and Nginx configurations. However it seems like every Windows deployment I come across using IIS is either not in any form of config management, or is managed through a series of VBScript and PowerShell scripts or System Center. If your organization is big enough to afford System Center this isn’t really much of an issue for you; however for those of us managing IIS web farms without SCCM, Microsoft has extended a nice olive branch with the delivery of PowerShell Desired State Configuration.

Given that most of us work with such broad toolsets, the idea of adding yet another configuration management platform into the mix seems less than ideal. However, there is an official Puppet Forge module that allows you to leverage PowerShell DSC within your Puppet code. What I found in my journey of doing this is that, while there is some decent documentation, there are pieces that clearly seem not to be well documented that required figuring it out the hard way. In this blog post I’ll cover and distill what I’ve learned in configuration of IIS through PowerShell DSC via Puppet.

Prerequisites:

This blog post will assume that you have already installed the PowerShell DSC forge module. I will link the module and its dependencies below; if you are unsure how to install Forge modules, see the documentation on the link to the module below.

This blog post will also assume that you have already installed IIS either in your AMI/Template or via Puppet. Great, now that that’s all out of the way, let’s get started.

 

Overview Example Class:

Before diving into the weeds in each section, I wanted to provide a code sample of a single application pool, website, virtual directory, and web application, complete with the DSC dependencies:

 

class webhosting::clientname::sites::dev {
 
dsc_windowsfeature{'iis':
  dsc_ensure => 'Present',
  dsc_name => 'Web-Server',
}
dsc_windowsfeature{'aspnet45':
  dsc_ensure => 'Present',
  dsc_name => 'Web-Asp-Net45',
}
dsc_windowsfeature{'iisscriptingtools':
  dsc_ensure => 'Present',
  dsc_name => 'Web-Scripting-Tools',
}
 
#App Pool Creation
dsc_xwebapppool{'dev.example.com':
  dsc_name => 'dev.example.com',
  dsc_ensure => 'Present',
  dsc_enable32bitapponwin64 => true,
  dsc_managedruntimeversion => 'v4.0',
  dsc_managedpipelinemode => 'Integrated',
  dsc_identitytype => 'ApplicationPoolIdentity',
  dsc_state => 'Started',
  require => Dsc_windowsfeature['iis'],
}
 
#Website Creation
dsc_xwebsite{'dev.example.com':
  dsc_ensure => 'Present',
  dsc_name => 'dev.example.com',
  dsc_state => 'Started',
  dsc_physicalpath => 'D:\\www\\dev.example.com\\webroot',
  dsc_applicationpool => 'dev.example.com',
  dsc_bindinginfo => [{
  ipaddress => '*',
    protocol => 'HTTP',
    port => 80,
    hostname => 'dev.example.com',
    },
    {
    ipaddress => '*',
    protocol => 'HTTPS',
    port => 443,
    hostname => 'dev.example.com',
    certificatethumbprint => '<certificatethumbprintgoeshere>',
    certificatestorename => 'WebHosting',
    }],
 require => Dsc_xwebapppool['dev.example.com'],
  }
 
#Web Applications
dsc_xwebapplication{'AwesomeApp':
  dsc_ensure => 'Present',
  dsc_name => '/AwesomeApp',
  dsc_website => 'dev.example.com',
  dsc_webapppool => 'dev.example.com',
  dsc_physicalpath => 'D:\\www\\dev.example.com\\webroot',
  require => Dsc_xwebsite['dev.example.com'],
  }
 
#Virtual directories
dsc_xwebvirtualdirectory{'AwesomeApp':
  dsc_ensure => 'Present',
  dsc_name => 'AwesomeApp',
  dsc_website => 'dev.example.com',
  dsc_webapplication => 'AwesomeApp',
  dsc_physicalpath => 'D:\\webapplicatons\\AwesomeApp',
  require => Dsc_xwebapplication['AwesomeApp'],
  }
}

DSC Windows Feature

Starting at the top and taking this a chunk at a time, we see after the opening Puppet class statement that there are several dsc_windows_feature statements. These statements install the PowerShell DSC packages on the endpoint nodes. PowerShell DSC wrapped in Puppet assumes that you are using the pull model style of using DSC. What essentially happens on the Puppet run is that your Puppet DSC code gets converted into PowerShell DSC code and run locally on the machine. DSC module can be found at https://gallery.technet.microsoft.com/scriptcenter. Note that modules beginning with an X are considered experimental. That said, most of the experimental modules in my experience have been stable.

 

Configuring The Rest

Beyond this point, the rest of the Puppet code you will see will be prefixed with dsc_. Essentially, to write DSC code in Puppet you will use the same values outlined in the DSC documentation, just prefixed with the dsc_ and using the Ruby hash rockets => instead of =. See the example below:

DSC Example:

    xWebsite DefaultSite
        {
            Ensure          = "Present"
            Name            = "Default Web Site"
            State           = "Stopped"
            PhysicalPath    = "C:\inetpub\wwwroot"
        }

Puppet Example:

     dsc_xWebsite{'DefaultSite':
            dsc_ensure       => 'Present',
            dsc_name         => 'Default Web Site',
            dsc_state        => 'Stopped',
            dsc_physicalpath => 'C:\\inetpub\\wwwroot',
        }

At this point, you can configure your applications specific to your environments. Much like any other Puppet code there are obvious dependencies (can’t create a website that depends on an application pool unless the pool already exists). For this you can use the normal Puppet meta parameters such as before and require statements to resolve dependencies. For IIS-specific DSC parameters see the readme for Microsoft’s XWebAdministration module here:  https://github.com/PowerShell/xWebAdministration

 

Don’t forget to prefix the parameters in the readme with dsc_ when converting to Puppet code. Happy configuring and Puppeting! Until next time, may your servers always be up and your coffee mugs never empty!

VS Code!

Ever since graduating from a Systems Admin role to DevOps one of the most frequently used tools in my toolbag is the text/code editor. Like many people I started on Notepad++ which by its own right is still a legitimately good editor. However working back and forth between Mac, Linux, and Windows desktop environments I gravitated towards Sublime Text and stayed there for a long time. I’ve recently begun studying for the MCSA Linux on Azure certification and have begun poking at Visual Studio more and in doing so discovered VS Code. VS code is a beautiful open source editor with a ton of plugins and a nice UI with excellent choices in syntax highlighting. Here’s some of the selling points that caused me to jump over to it:

 

  • Dark Themes
  • Cross Platform Support (Linux, Mac, and Windows)
  • options to add “code” to your path
  • automatic detection of files in a git repo and change tracking
  • useful diff screens
  • ability to launch terminal from within the editor window
  • Chef Support
  • Puppet Support
  • Powershell Support
  • Folder tree management
  • Syntax highlighting that’s helpful without being color vomit
  • Integrations with Azure WebApps PaaS
  • Key bindings for Vim, Sublime and other familiar editors

 

Here’s a Link to download VS Code: https://code.visualstudio.com/