Norfolk.js Turns One Year Old

Norfolk.js just celebrated its first birthday in November!

Norfolk.js is Hampton Roads’ JavaScript developer user group. It was founded in Norfolk, Virginia, by small group developers and has grown to nearly 200 members are from all over the area. The goal of this user group is to create the most accessible way possible for local developers to meet to discuss all things related to JavaScript. In 2015, we hope to continue the momentum of the group by focusing on topics of particular interest to our members.

Norfolk.js has been a tremendous experience and a bigger success than I, or any of the other organizers, could have imagined. The success of of Norfolk.js isn’t just due to the dedication of any one of us, but because of support of our entire local development community. The financial and logistical support from local sponsors (like CodeBetter.com) also allows us to deliver a quality experience that rivals groups in larger cities. We’ve been able to learn from  existing user groups like 757.rb and HRNUG, and get feedback from our peers about what works for them in a developer user group. We open-sourced our website and our organizing discussions, and use Github to facilitate this valuable feedback and to make Norfolk.js as inclusive as possible.

Over the last year we’ve hosted some of the best developers in the local area and have had visits from some amazing out-of-town JavaScripters, like Chris Williams and Mike McNeil. We’ve also hosted special events like International NodeBots Day and Worldwide Meteor Day. We typically meet on the third Monday of each month to eat, drink, socialize, and listen to an interactive talk. Every four months, we have a social meetup, Beer.js, at a local pub or brewery. This month, we welcomed Hiro Asari from Chapel Hill, NC to discuss unit testing and continuous integration with Travis-CI. In February, we will be partnering with the Chrysler Museum for the first time to host our meetup in their museum space and get a private viewing of their “Art of Video Games” exhibit.

Norfolk.js has provided a way for our members to meet new people and try new JavaScript-related technologies that might have otherwise been unavailable in this area.

We are always happy to welcome new speakers, sponsors, and members. Contact one of our organizers if you’d like to get involved. To get a list of our upcoming events, please see our events page.

Posted in Uncategorized | Leave a comment

Now Witness the Power of a Fully Operational CodeBetter Scholarship!

The CodeBetter Endowed Scholarship was a $25,000 commitment over 5 years to help provide financial support for female undergraduate students in computer science.  It’s one of our efforts to try to chip away at the current gender disparity in our field.  I’m happy to announce  that we’ve reached our goal! I just received this email from the director of donor affairs at Old Dominion University, where we established the CodeBetter endowed scholarship.

With the $500 payment just received, the endowment is fully funded. The endowment will produce an award for the 2016-2017 academic year.

The way this whole process works has been really interesting to me.  Once the scholarship is funded, the return on its investments produce an award.  In this way, it will outlive all of us – that’s a powerful idea to me.

It took us a few years of continuous funding. In fact, we’ve been sending a check for $500/month over the past 5 years.  This is roughly 70-80% of our total ad revenue. We’ve also been putting in extra every month so that we can support a student while we have been funding the endowment. And I’m happy to report that our first recipient is about to graduate this month with her BS in computer science.  Congratulations to Banita!

So, now that that is done, I’ve been making some more local commitments so that CodeBetter can continue to support developers.  One is to provide some monthly funding for the local .js group here in VA (in fact, one of the group’s leaders, Linda Nichols, is soon to start blogging with us).  Another is to provide support for the local Code for America brigade.  We are going to continue to support our community…  Have a need?  Get in touch and pitch it to us.

Posted in Uncategorized | Leave a comment

Using Grunt/Karma/Jasmine-Ajax inside your Jasmine Tests

My current client is trying to add JavaScript specs to its code base and we have decided to implement a Grunt/Karma/Jasmine solution.

Like many (if not most) JavaScript code bases our code will make remote calls via jQuery to the server to Get or Post data. I wanted to be write tests which mocked the over the wire call and returned me some stubbed data. Accomplishing this took very little effort and was quite simple once I figured it out.

Here is how I accomplished this
1) Add the Mock-Ajax.js file to your test suite. I did this manually as the NPM package used a very old version of the .js file and it caused me issues.

2) Add a reference to the mock-ajax.js file to your Karma configuration

    files: [
        ...
        'js.tests/mock-ajax.js',
        ...
    ],

3) Install the mocks inside the beforeEach() method inside your spec

    beforeEach(function() {
        jasmine.Ajax.install();
    });

4) Create your test response you are mocking inside your spec

var testResponses = {
    results: {
        success: {
            status: 200,
            responseText: { json goes here }
        }
    }
};

Your test response needs to follow this pattern because the mock library is going to inspect the response for success/failure status.

5) Run your JavaScript code as normal, no need to setup spy’s or mocks as the Ajax library is going to intercept all HTTP calls automatically.

6) Grab the most recent request and shove in the mock results

        var request = jasmine.Ajax.requests.mostRecent();
        request.respondWith(testResponses.results.success);

You can add the code below AFTER the code that makes the HTTP call, it is ok. You should notice that I am pushing in the success result inside of the respondWith() call, this will tell the Ajax library how to respond inside your code.

7) setup your asserts as expected based on your mock results.

Hope this helps.

Till next time,

Posted in Grunt, Jasmine, Testing | Leave a comment

How To Avoid Common Windows Gotchas with Puppet

Originally published at http://puppetlabs.com/blog/how-avoid-common-windows-gotchas-puppet, reposted with permission and adjustments.

I’ve touched on Puppet before – Making Windows Awesome Since 2011Getting Started with Puppet on Windows and Managing Permissions with Access Control Lists.

I’m going to touch on the common things you can do, then show you how to avoid the common gotchas. Throughout the post, I’ll direct you to resources where you can dive in more deeply. When it comes to managing Puppet on Windows, a great resource is our Windows Documentation.

Common Windows Tasks You Can Automate With Puppet

Using Puppet, you can manage:

  • users /groups
  • files and folders
  • scheduled tasks (cron for Windows)
  • installation/upgrade/removal of software
  • permissions
  • windows features
  • management of services
  • IIS
  • management of databases
  • etc

The built-in types are listed on Windows Core Resource Types, along with specific examples. Much of the Windows ecosystem comes from several high-quality modules on the Puppet Forge that you can add in to Puppet to enhance your experience on Windows. Here are just a few specific modules:

There are of course many other modules for Windows, and I encourage you to explore the Forge.

As you can see, you can use Puppet to manage practically anything on Windows.

Common Windows Gotchas

We are always adding common gotchas and their resolutions to the Windows Troubleshooting page. Be sure to check that out if you don’t find what you are looking for here.

Puppet May Run as a 32-bit Application

Before Puppet 3.7.0, there was no 64-bit version of Puppet. This means you would be subject to File System Redirection and Registry Redirection from Windows because Puppet would run under 32-bit Ruby. The next two sections only apply if you are using a 32-bit version of Puppet on an x64 System.

File System Redirection (only when not using Puppet x64 on x64 systems)

Windows reserves the C:\Windows\System32 directory for 64-bit applications. In its great knowledge, Windows will redirect calls to System32 with SysWOW64, which contains the 32-bit equivalents of what is in the native system32 directory. Most of the time with Puppet, users actually want the physical system32 directory — the native system directory. In cases like these, what is wanted is to use the SysNative alias to bypass file system redirection.

To better understand what I mean about redirection, let’s take a simple manifest and execute it:

exec { 'powershell_system32':
  command => 'c:\windows\system32\WindowsPowerShell\v1.0\powershell.exe -NonInteractive -NoProfile -ExecutionPolicy Bypass -Command "get-process powershell* | %{ Write-Host $($_.Path)}"',
  logoutput => true,
} ->
exec { 'powershell_sysnative':
  command => 'c:\windows\sysnative\WindowsPowerShell\v1.0\powershell.exe -NonInteractive -NoProfile -ExecutionPolicy Bypass -Command "get-process powershell* | %{ Write-Host $($_.Path)}"',
  logoutput => true,
}  

The output?

Notice: /Stage[main]/Main/Exec[powershell_system32]/returns: c:\windows\SysWOW64\WindowsPowerShell\v1.0\powershell.exe
Notice: /Stage[main]/Main/Exec[powershell_sysnative]/returns: c:\windows\system32\WindowsPowerShell\v1.0\powershell.exe


In this example, we are executing PowerShell and then looking at the path of any running powershell.exe processes (the same process we are running). We told the first exec (named powershell_system32) we wanted c:\windows\system32, but we got c:\windows\SysWOW64. This is how file system redirection works. If we want c:\windows\system32, we need to ask for c:\windows\sysnative, like we did with the second exec (powershell_sysnative).

For more information, see File System Redirection in 64-bit Windows Versions. You should also take a look at the MSDN reference for the File System Redirector.

NOTE: In the unlikely case where you are running Windows Server 2003 x64 (ONLY non-R2), you must apply the KB942589 hotfix to be able bypass file system redirection. And there is no x64 version of Ruby available for Windows Server 2003, so there is no x64 version of Puppet available either.

Registry Redirection (only when not using Puppet x64 on x64 systems)

In most cases, you will want to use the registry module, as it will provide you with the correct calls. In cases where you cannot, like when writing custom facts for Facter, you will want to do something similar to the following:

Facter.add(:myfact) do
  confine :kernel => :windows
  setcode do
    require 'win32/registry'

    value = nil
    hive = Win32::Registry::HKEY_CLASSES_ROOT
    hive.open('SOFTWARE\Somewhere\SomeValue',  Win32::Registry::KEY_READ | 0x100) do |reg|
      value = reg['SomeValue']
    end
    value
  end
end    

 

The addition of | 0×100 ensures the registry is opened without redirection so you can access the proper keys that you expect to access. For more information, see the MSDN Reference.

UAC (User Account Control) aka Security Context

UAC can bring you into some interesting places with Puppet. The Puppet agent (service) by default runs under an elevated LocalSystem account, which requests a certificate for the machine. If you run Puppet later from the command line in a non-elevated security context, it will attempt to request a second SSL certificate, which usually causes errors. See our documentation on security context.

Case Is Important

Most items on Windows are case-insensitive. Puppet, however, is not, especially when it comes to establishing relationships between resources in the manifests. See our documentation about case insensitivity.

File Paths

The general rule of thumb is to always use forward slash (/) except when specific Windows programs, like msiexec.exe, require backslashes in file paths. In those cases you should use the backslash (\) or a double backslash (\\) when the string is double quoted. The following paths are valid:

file { 'c:/path/to/file.txt': }
file { 'c:\path\to\file.txt': }
file { "c:\\path\\to\\file.txt": }

This is not a valid path due to double quoted single backslashes, which are interpreted as escape sequences:

file { "c:\path\to\file.txt": }

You will find more information here: http://docs.puppetlabs.com/windows/writing.html#file-paths-on-windows and http://docs.puppetlabs.com/windows/troubleshooting.html#path-separator.

Reboots

Yes, that’s the thing you have to do sometimes on Windows….it’s crazy, but it works. When you need to reboot, there’s a supported module for that. See Rebooting Windows with Puppet Enterprise.

Parting Thoughts

Now you’ve seen a few common gotchas to avoid and you’ve been introduced to some some stellar modules for enhancing the functionality of Puppet on Windows. Armed with this information, it is my hope that you:

  • Have a better understanding of Puppet and Windows
  • Run into fewer issues when you work with Puppet on Windows
  • Know where to find more Windows resources
Posted in howto, puppet | Tagged , | Leave a comment

Puppet–Managing Permissions on Windows with Access Control Lists

acl type
Originally published at http://puppetlabs.com/blog/managing-permissions-windows-access-control-lists, reposted with permission and with adjustments. I wrote the original post and am also the author of the ACL module.

While there is a simple interface for setting permissions on Windows, managing and maintaining permissions has never been simple. Puppet has a module to make working with permissions almost enjoyable! You can work directly with ACLs (Access Control Lists) and to a degree, security descriptors, through the puppetlabs-acl module. The ACL module serves to make that process easy for you while satisfying your needs for very advanced permissions.

Let’s talk about a couple of real-world scenarios the ACL module can solve:

  • You need to set permissions appropriately for something like IIS/Apache.
  • You have a directory or file that you need to lock down to just admins.

These are very possible with the ACL module and just a few short lines of Puppet code. I’ll show you that in the Examples section below, but first let’s talk about ACLs and ACEs for those who want a refresher.

ACLs and ACEs, Oh My!

ACLs (also called Discretionary Access Control Lists) typically contain a list of access control entries (ACEs). An ACE is a defined trustee (identity) with a set of rights, and information about how those rights are passed to (and inherited by) child objects — for example, files and folders. For each ACE, the ACL contains an allowed/denied status, as well as the ACE’s propagation strategy. You cannot specify inherited ACEs in a manifest; you can only specify whether to allow upstream inheritance to flow into the managed target location (the location where you are applying the ACL).

ACL Features

Here are some features of Puppet’s ACL you should know:

  • Puppet can manage the complete set of ACEs or ensure that some Puppet-specified entries are present, while leaving existing entries alone.
  • You can lock down a path to only the specified permissions.
  • Identities can be users and/or groups, domain users/groups, and/or SIDs.
  • You can point multiple ACL resources at the same target path. Let me say that again in case there was confusion: You can point multiple ACL resources at the same path.
  • Propagation and inheritance of ACEs is set to Windows defaults, but can be specified per ACE.

Here’s what sets Puppet’s ACL module apart from other configuration management tools:

  • ACE order. The order of your ACEs matters. If they are incorrectly ordered, it can cause issues. So we apply ACEs in the order you’ve specified in the manifest. And we insert them in the correct location when merging with unmanaged ACEs.
  • SID (Security ID) support. We support specifying identities as SIDs.
  • Very, very granular permissions. Do you need to apply read attributes (RA)? Yes, we can do that.

Getting Started

Let’s take a look at what a typical ACL resource looks like:

acl { 'c:/temp':
  permissions => [
   { identity => 'Administrator', rights => ['full'] },
   { identity => 'Users', rights => ['read','execute'] }
  ],
}

If you were to run the above on a system that had the module installed, you would be giving the Administrator account full access to the temp folder, and giving the Users group access to read and execute (and list for folders). All the other options are set to Windows defaults, but if we need to get to them, we can. Let’s look at that same ACL resource with all the options specified:

acl { 'c:/temp':
  target      => 'c:/temp',
  target_type => 'file',
  purge       => 'false',
  permissions => [
   { identity => 'Administrator', rights => ['full'], type=> 'allow', child_types => 'all', affects => 'all' },
   { identity => 'Users', rights => ['read','execute'], type=> 'allow', child_types => 'all', affects => 'all' }
  ],
  owner       => 'Administrators',
  group       => 'Users',
  inherit_parent_permissions => 'true',
}

We have just specified the resource with all parameters and properties specified; what you are seeing is how the defaults line up. The only exception here are owner and group, which by default are not managed unless specified. The defaults for these depend on the user that created the target (the folder) and could be different based on who the user has as their default group and owner.

With both of the above examples we have done the following:

  • We’ve given Administrator full access to c:\temp.
  • We’ve given Users read/execute access to c:\temp.
  • We are by default not removing explicit unspecified permissions (purge => ‘false’).
  • We are by default inheriting the permissions from the parent folder (inherit_parent_permissions => ‘true’).
  • For each ACE, we are allowing the permission by default (type => ‘allow’).
  • For each ACE, we apply the permission to all child types by default, both folders and files (child_types => ‘all’).
  • For each ACE, we propagate the permission to self, direct children and all those further down by default (affects => ‘all’). Note that all types below child types are called grandchildren, no matter what the level of depth.

By default, if a user is not granted access through an ACE (whether individually or as a member of a group), then Windows will deny access. So if a user is not the Administrator account and not a member of the Users group, they will not have access to c:\temp. With the ACL module, this also means that access could be granted outside of Puppet (when purge => ‘false’) and/or it is an inherited ACE (when inherit_parent_permissions => ‘true’).

ACL Type Structure

Now let’s take a look at the ACL type and all it has to offer.

Parameters

The parameters are (bold means the parameter is required):

  • name – The name of the ACL resource, used as target if target is not set explicitly (see the target parameter).
  • purge – Determines whether unmanaged explicit access control entries should be removed. Combine purge => ‘true’, inherit_parent_permissions => ‘false’ to really lock down a folder. Supports true, false and listed_permissions. Defaults to false.
  • target – The location; defaults to the same value as name.
  • target_type – Currently supports only file; defaults to file.

Properties

The properties are (bold means the property is required):

owner – User/Group/SID that owns the ACL. If not specified, the provider will not manage this value.
group – User/Group/SID that has some level of access. If not specified, the provider will not manage this value. Group is not commonly used on Windows.
inherit_parent_permissions – Whether we inherit permissions from parent ACLs or not. Default is true.
permissions – The list of ACEs as an array. This should be in the order you want them applied. We cover permissions in more detail below.

Permissions Property

The permissions property could be considered the most important part of the ACL resource, because it contains each ACE in the order specified for the ACL. The available elements for each ACE hash are identity, rights, type, child_types, affects, and mask.

The elements are (bold means the key is required):

  • identity – This is the user/group/SID.
  • rights – An array with the following values: full, modify, mask_specific, write, read, and execute. The full, modify, and mask_specific values are mutually exclusive, and when any of these values are used, they must be the only value specified in rights. The full value indicates all rights. The modify value is cumulative, implying write, read, execute and DELETE all in one. If you specify mask_specific, you must also specify the mask element in the permissions hash. The write, read, and execute values can be combined however you want.
  • type – Whether to allow or deny the access. Defaults to allow.
  • child_types – Which types of children are allowed to inherit this permission, whether objects, containers, all or none. Defaults to all.
  • affects – How inheritance is propagated. Valid values are all, self_only, children_only, self_and_direct_children_only, or direct_children_only. Defaults to all.
  • mask – An integer representing access mask, passed as a string. Mask should be used only when paired with rights => [‘mask_specific’]. For more information on mask, see the granular permissions example below.

For more specific details and up-to-date information on the ACL type, see the usage documentation.

Examples

In the following examples we are going to show you how to lock down a folder, set appropriate permissions for a website, and set very granular permissions.

Locking down a folder for sensitive data

Here’s what you need to do when you keep sensitive data in a specific folder, and need to limit access to administrators only.

acl { 'c:/sensitive_data':
  purge                       => true,
  inherit_parent_permissions  => false,
  permissions => [
   { identity => 'Administrators', rights => ['full'] }
  ],
}

We’ve done the following:

  • Since permissions are inherited from parent ACLs by default, we set inherit_parent_permissions => false so no permissions are inherited from the parent ACL.
  • ACLs will also allow unmanaged ACEs to coexist with Puppet-managed permissions, so we need to specify purge => true to ensure that all permissions other than those we have specified are removed.
  • We’ve given the Administrators group full permission to the directory.
  • If an ACE has not granted permission for a user, Windows will by default deny access. So in this case, only users that are part of the Administrators group will be able to access this folder.

That was pretty simple! And very self-documenting as well. It’s worth mentioning that one goal of Puppet is for anyone to be able to read and understand the intent of the code, and the ACL module does not disappoint. Now let’s try something a little more involved.

Website setup with ACLs

Let’s take a look at setting up an IIS site and locking down permissions.

$website_location = 'C:\sites\thestuff'
$website_name = 'the.stuff'
$website_port = '80'

# add windows features
windowsfeature { 'Web-WebServer':
  installmanagementtools => true,
} ->
windowsfeature { 'Web-Asp-Net45':
} ->

# remove default web site
iis::manage_site { 'Default Web Site':
  ensure        => absent,
  site_path     => 'any',
  app_pool      => 'DefaultAppPool',
} ->

# application in iis
iis::manage_app_pool { "${website_name}":
  enable_32_bit           => true,
  managed_runtime_version => 'v4.0',
} ->
iis::manage_site { "${website_name}":
  site_path     => $website_location,
  port          => "${website_port}",
  ip_address    => '*',
  app_pool      => "${website_name}",
} ->

# lock down web directory
acl { "${website_location}":
  purge                       => true,
  inherit_parent_permissions  => false,
  permissions => [
   { identity => 'Administrators', rights => ['full'] },
   { identity => 'IIS_IUSRS', rights => ['read'] },
   { identity => 'IUSR', rights => ['read'] },
   { identity => "IIS APPPOOL\\${website_name}", rights => ['read'] }
  ],
} ->
acl { "${website_location}/App_Data":
  permissions => [
   { identity => "IIS APPPOOL\\${website_name}", rights => ['modify'] },
   { identity => 'IIS_IUSRS', rights => ['modify'] }
  ],
}

The script above does the following:

  1. Ensures IIS is installed and ASP.NET is set up.
  2. Ensures the default web site is removed.
  3. Ensures our site is set up correctly.
  4. Uses ACL to ensure IIS users have read access.
  5. Uses ACL to ensure the processes that run the website can modify the App_Data directory.

All of that, done in just a few lines of code! You can add more for getting the files there in the first place, using Puppet, but I have left that as an exercise for the reader :)

Granular Permissions

In our last example, let’s get into very granular permissions.

acl { 'c:/granular_permissions':
  permissions => [
   { identity => 'Administrators', rights => ['full'] },
   { identity => 'Bob', rights => ['mask_specific'], mask => '1507839' }
  ],
}

We’ve done the following:

So you see 1507839 and it looks like a magic number. How does one arrive at this number? I wrote a blog post on this once that shows you how to add up the numbers. Now we have a much simpler way to add those permissions: We have created and made available a worksheet to add up the ACL rights mask! This will allow you to add up the rights and will give you a heads up if you should use named rights from the module (like read or execute or a combination of read, execute).

ACL Module: You’re gonna love my permissions!

ACLs on Windows give you a lot of control over permissions. The ACL module harnesses that ability and makes it easy — even fun — to control permissions on Windows. While the ACL module makes it easy to get started, it allows advanced users to manage permissions in granular detail. And did I mention it’s a supported module? That means you get the same support for this module that you get with a Puppet Enterprise license. Give the ACL module a look today — if you manage permissions on Windows, I’m pretty sure you are going to love it!

References

Posted in howto, puppet | Tagged | 1 Comment