http://thinkingconcurrently.com

Installing VMware Tools on Fedora 19

2013-07-22T21:39:23-04:00

A freshly installed Fedora 19 guest comes preinstalled with open-vm-tools, an open-source version of the VMware tools for guest operating systems. However, these tools can't do everything the proprietary version can, at least on VMware Fusion, so here's the steps required to swap things out.

First, we need to remove open-vm-tools from the system:

sudo yum remove open-vm-tools open-vm-tools-desktop

Next we need to install the kernel header files required for compiling VMware tools:

sudo yum install -y kernel-devel-$(uname -r)

Normally this would be the last step before starting the tools installation, but Fedora 19 (and RHEL 6.4) have a version.h file that's not where the tools installer expects it. So, copy it to where it's needed:

sudo cp /usr/src/kernels/$(uname -r)/include/generated/uapi/linux/version.h \
        /lib/modules/$(uname -r)/build/include/linux/

Now, initiate the VMware tools install - on VMware Fusion this is the menu item Virtual Machine -> Reinstall VMware Tools. After the virtual CD is mounted, proceed with the install:

cd /tmp/
      cp /run/media/bbrowning/VMware\ Tools/VMwareTools-*.tar.gz .
      tar xzf VMwareTools-*.tar.gz
      sudo vmware-tools-distrib/vmware-install.pl

I just accepted all the defaults for every prompt, including allowing the installer to run vmware-config-tools.pl.

Hope that helps - having to copy the version.h file is the biggest thing that will trip up experienced VMware users.

Fedora 19 and Amazon Instant Video

2013-07-22T21:39:23-04:00

Amazon Instant Video and other DRM-laden flash websites don't work out of the box with flash on Fedora 19. Below is the list of steps I took to get this working in case it helps anyone else. The majority of these instructions were pieced together from the post and comments at http://wtogami.blogspot.com/2012/12/amazon-instant-video-on-fedora-16.html. The linked RPMs were originally taken from that post and recompiled for Fedora 19. These steps have only been tested in Firefox and 64 bit Fedora 19 - I don't think they'll work for Chrome.

First, head to http://get.adobe.com/flashplayer/ and download the "YUM for Linux (YUM)" RPM. Then:

# Install Adobe Flash and its browser plugin
      sudo yum install -y adobe-release-x86_64-1.0-1.noarch.rpm
      sudo yum install -y flash-plugin
      
      # Install a simple SELinux policy file for Flash
      sudo yum install -y policycoreutils-devel
      wget http://togami.com/~warren/archive/2012/adobedrm.te
      checkmodule -M -m -o adobedrm.mod adobedrm.te
      sudo semodule_package -o adobedrm.pp -m adobedrm.mod

Download the fakehal RPMs for Fedora 19 at http://thinkingconcurrently.com/files/f19_flash/fakehal-0.5.14-7.fc19.x86_64.rpm and http://thinkingconcurrently.com/files/f19_flash/fakehal-libs-0.5.14-7.fc19.x86_64.rpm

# Install the fakehal RPMs
      sudo yum install -y fakehal-0.5.14-7.fc19.x86_64.rpm \
        fakehal-libs-0.5.14-7.fc19.x86_64.rpm

At this point, make sure all Firefox windows are closed.

rm -rf ~/.adobe/Flash_Player/
      sudo mkdir -p /usr/share/hal/fdi/preprobe \
        /usr/share/hal/fdi/information \
        /usr/share/hal/fdi/policy/20thirdparty \
        /var/cache/hald/
      sudo ln -s /usr/share/hal /etc/hal
      sudo touch /var/cache/hald/fdi-cache
      sudo systemctl start haldaemon.service

Now start up Firefox and watch your favorite Amazon Instant Video. Let me know if it works for you!

JRubyConf Presentation

2013-07-22T21:39:23-04:00

I had a blast at JRubyConf 2011 this year, and not just because of the whisky tasting. It was fun to finally put faces to all the people I interact with over the internet and to meet a few new ones.

Video and slides from my presentation are below:

Scaling Rails Applications with TorqueBox

2013-07-22T21:39:23-04:00

Below are video and slides from my Scaling Rails Applications with TorqueBox presentation given at JUDCon:2011 Boston on May 3rd, 2011. This was my first opportunity to speak at a conference and I thoroughly enjoyed and learned a lot from it. That's one of the great things about having a video of the presentation; I see plenty of things I can improve upon for next time.

Despite my critiques, I feel the talk went well and look forward to bringing Ruby to more JBoss users and developers at next year's JUDCon!

Scaling Rails With Torquebox Presented at JUDCon:2011 Boston

Updated Deltacloud VirtualBox Driver

2013-07-22T21:39:23-04:00

The SteamCannon project uses Deltacloud to abstract away the details of specific cloud providers when launching instances. For local development, it's nice to be able to launch a cluster of virtual machines in VirtualBox instead of actually starting and stopping instances on a public cloud like Amazon EC2.

Michal Fojtik created a deltacloud-virtualbox-driver many months ago to add VirtualBox support to Deltacloud. However, Deltacloud has evolved quite a bit since it was initially released and that driver no longer works with the current implementation.

I've taken Michal's great start and updated it to work with the latest Deltacloud release, giving it a few other changes in the process.

Switched to the excellent virtualbox gem which uses the VirtualBox COM API instead of shelling out to the VBoxManage process. The COM API is substantially faster and should allow this driver to work on Windows instead of just Linux/Mac (but it hasn't been tested on Windows).
Networking settings of the launched instance are now copied from the parent image instead of being hardcoded to a specific value.
The driver now works under JRuby. There are still some optimizations to be made in the code based on which environment it's running in, but it will at least run in JRuby and MRI.

Installing the VirtualBox Driver

The driver is very new and still needs to mature a bit before I submit it to the upstream Deltacloud project. Until then, if you're comfortable living on the edge, read below for installation instructions.

Install VirtualBox 3.2.x (might work with 3.1.x but has not been tested) - http://www.virtualbox.org/wiki/Downloads
Install bbrowning-virtualbox gem (temporary fork of upstream implementing some missing features)
```
  $ gem install bbrowning-virtualbox --pre
      
```

Install bbrowning-deltacloud-core gem

  $ gem install bbrowning-deltacloud-core

Launch deltacloud-core with the VirtualBox driver
```
  $ deltacloudd -i virtualbox
      
```

If all goes well, you should be able to go to http://localhost:3001 in your browser and see the Deltacloud web interface. Clicking the Images link should display all virtual machines you have setup inside VirtualBox - if you don't see any, try creating a new virtual machine in VirtualBox and then come back to play around.

The instances launched from an image clone the hard drive of the image but any changes made in the instance are not reflected in the parent image. If you don't have a lot of free hard drive space, make sure to go to the Instances page and destroy stopped instances that you no longer need.

Enjoy!

Moving From GitHub Pages to Awestruct

2013-07-22T21:39:23-04:00

I'm switching the site over from GitHub Pages to Awestruct. I was turned on to Awestruct by Bob McWhirter and am really digging the compass / blueprint integration and ease of writing extensions.

As part of the transition, I'm also switching web hosts. It should go pretty smoothly but because of a poor IntenseDebate implementation on my old site I'm having a hard time pulling all the previous blog comments over. There weren't many to begin with, but I apologize in advance if I can't get them all to show up on the new site.

TorqueBox Vagrant Box Updated to Beta20

2013-07-22T21:39:23-04:00

Hot on the heels of the TorqueBox Beta20 release, I've pushed an updated version of the TorqueBox Vagrant box. The box is now 1.6GB in size - I'm not entirely sure why it's larger than last time but if I figure it out I'll upload a smaller version.

Grab it here:

http://bbrowning-boxes.s3.amazonaws.com/torquebox-1.0.0.Beta20.box?torrent (bittorrent)

http://bbrowning-boxes.s3.amazonaws.com/torquebox-1.0.0.Beta20.box (direct download)

Refer to my previous post for details on getting started with TorqueBox and Vagrant.

Developing Rails Apps with TorqueBox and Vagrant

2013-07-22T21:39:23-04:00

The TorqueBox project does a great job of bringing together the best parts of Ruby, Rails, JBoss, and the JVM. However, I've recently started using Vagrant for all my development work and couldn't find anyone that had used the two together. So, after hacking around a bit, I've created a Vagrant box anyone can use to easily get up and running with TorqueBox.

The rest of this post assumes a basic level of familiarity with TorqueBox and Vagrant - things will make more sense if you browse some of their documention before reading further.

What's Included in the Box

Ubuntu 8.04
OpenJDK 1.6.0_0-b11
TorqueBox 1.0.0.Beta19

Because it contains a full TorqueBox binary and a Java runtime, the box is 1.2GB in size.

Creating Your First Project

Install Vagrant if you haven't already

Add the TorqueBox Box to Vagrant

Preferred Method - Bittorrent

Use your favorite Bittorrent client to download http://bbrowning-boxes.s3.amazonaws.com/torquebox.box?torrent and then run
```
  $ vagrant box add torquebox path_to_downloaded_file/torquebox.box 
      
```
Alternate Method - Direct Download

The direct download will be faster (but cost me more money) unless several nice people help me seed the torrent.
```
  $ vagrant box add torquebox http://bbrowning-boxes.s3.amazonaws.com/torquebox.box
      
```

Create and initialize a sample application

  $ mkdir sample_app
        $ cd sample_app
        $ vagrant init torquebox

Run the box
```
  $ vagrant up
      
```
This step may take a few minutes.
Create a Rails project

To simplify things, all rails/rake commands should be run from inside the TorqueBox VM.

Log into the VM and cd to the shared vagrant folder.
```
  $ vagrant ssh
        $ cd /vagrant
      
```
Follow the steps from the TorqueBox Rails documentation for creating our Rails application with the TorqueBox template:
```
  $ rails -m $TORQUEBOX_HOME/share/rails/template.rb .
      
```
If you look on your host machine, you'll see the Rails application files were created in your sample_app directory. The /vagrant directory inside your VM is a bidirectional shared-folder with the sample_app directory on your host.

After creating a new project, you'll need to redeploy your Rails app into TorqueBox. You only have to do this once - the VM will automatically pick up on the project the next time it's started.
```
  $ rake torquebox:rails:deploy
      
```
Wait 15-30 seconds and point your browser to http://localhost:3000. You should see the familiar Rails "Welcome aboard" page.

Congratulations, you're now running a Rails application inside TorqueBox.

Next Steps

If you've made it this far, you've done the hard part. Use your favorite editor on your host machine to develop your Rails application like you always have. The only thing to remember is to run script/generate, rake db:migrate, etc from inside the Vagrant VM. If you try to run these on the host you most likely won't have the right version of Rails or JRuby installed.

Changes made to your Rails project should be visible on the next page refresh just like regular Rails development. To try this, generate a simple Posts scaffold.

$ ./script/generate scaffold post title:string body:text
      $ rake db:migrate

Point your browser to http://localhost:3000/posts and your should see the scaffolded post index.

When you've finished working on this project or need to switch to another one, issue a vagrant halt or vagrant destroy from the sample_app directory on your host. The next time you need to work on this project again, issue a vagrant up to boot everything back up. See the Vagrant docs for more details on these commands.

Let's try it.

If you're still logged into the VM via ssh, log out.

$ exit

Destroy the VM

$ vagrant destroy

Now re-create the VM and start it back up (will take several minutes)

$ vagrant up

Wait 15-30 seconds for TorqueBox to start up and then point your browser to http://localhost:3000/posts - your app is back right where we left off.

If you have any feedback, requests, or problems with the TorqueBox box, contact me on Twitter (@bbrowning). Enjoy!

Running Bespin on the Couch

2013-07-22T21:39:23-04:00

Sparked by a thread on the CouchDB Users mailing list with promises of beer and internet groupies, I've started a project to get Bespin running as a CouchApp on CouchDB. The long-term goal is to allow you to create and edit CouchApps from within Bespin, which is itself a CouchApp.

My first goal was to get Bespin's UI functioning as a CouchApp. Except for a few small things, that's already finished.

Here's a few screenshots of Bespin running as a CouchApp on my local CouchDB:

Index Page

Help Menu

Dashboard

Editor

Over the next few days I'll be focusing on porting their filesystem implementation to CouchDB and CouchApp conventions. If you'd like to get involved, fork my bespin_couch project and have at it! Also, if anyone knows how I can make the GitHub repo track upstream changes from Bespin's Mercurial repo I'd like to get that setup so I'm not just working off of a snapshot.

I haven't had time to integrate Disqus with the blog yet so for now if you have any comments please send me a message on GitHub or a reply to @bbrowning on Twitter.

Scribbish and Jekyll

2013-07-22T21:39:23-04:00

I really wanted a minimalistic, simple theme for my blog. After looking around a bit I settled on Scribbish, a popular theme in the Ruby world with packages for Typo, Mephisto, and WordPress. Unfortunately there's no package for Jekyll so I decided to try and port it myself.

Both Mephisto and Jekyll use the Liquid templating engine so the port was pretty straightforward. It's still a work-in-progress but the basics are finished.

Jekyll has an include tag to create reusable components but, unlike Mephisto's, you can't directly pass variables to the included file. This is fairly easy to workaround by using Liquid's assign tag or explicitly looping over a collection of objects instead of using Mephisto's shortcuts. For example:

Jekyll {% for post in site.posts %} {% include post.html %} {% endfor %}

Mephisto {% include 'article' with articles %}

The Jekyll include tag looks for an _includes directory at the root source directory. However, if you try to use the include tag from an already included file there's a bug where it looks for _includes/_includes/included_file.html. To workaround this for now I just created a symlink in the _includes directory to itself. cd _includes && ln -s . _includes

Take a look at this blog's source, particularly the _layouts and _includes directories, if you'd like to use the Scribbish theme for your own Jekyll-based blog.

Blogging With GitHub Pages

2013-07-22T21:39:23-04:00

I've always wanted a place to talk about the projects I'm working on but a traditional blog just didn't feel right. A few weeks ago I stumbled across GitHub Pages and my inner geek was instantly aroused. GitHub Pages is really just free static site hosting coupled with a tool called Jekyll to generate your site's contents based on layouts and Markdown or Textile formatted text. This is both simple and powerful. It removes the distractions and lets me focus on what really matters - the content.