I ran into an issue while booting Foreman earlier:
Googling didn’t bring in many results, so after a long chase down the rabbit hole, I found the ultimate problem: despite having Java 7 installed, my terminal was still using the Apple-provided Java 6:
You can verify this using the
java -version command.
To fix this error, make sure that you have Java 7 downloaded and installed. You can grab this from this page:
You’ll want to the Mac OS X x64 dmg.
This install does not replace the $PATH binaries used by bash. You can verify this using the
whereis java command:
The files located at
/usr/bin/java are symlinks to the actual java install directory (verified through
ls -lha /usr/bin/ | grep JavaVM):
The new binaries that you downloaded and installed are located in the directory
/Library/Java/JavaVirtualMachines/jdk1.7.0_51.jdk/Contents/Home/bin/. To fix this, you’ll need to replace the symlinks in
/usr/bin/. I’ve written as script to take care of this that you can grab from here:
Download and run that from the terminal to create new symlinks in
/usr/bin/, and to move the old binaries to
Thanks to this Stack Overflow thread for helping in my investigation. If you have any issues or corrections, feel free to submit a pull request to the github page or comment in the section below.
During a recent shoot for Nobody Likes Onions, I ended up recording a lot of short clips (1-2 minutes over the course of 2-3 hours). After realizing I had configured the timecode incorrectly, I then had the tedious task of having to sync those clips up with a larger, un-cut recording of the entire event. Automatic synchronization through Premiere wasn’t working for every clip due to a lot of background noise in the audio, and I wasn’t prepared to shell out money for PluralEyes.
I started searching to see if anyone had built a script to sort clips in Premiere by capture time. I couldn’t find anyone who had, so I constructed this script: VideoClipSpacer.php
It requires a couple of in-between stages, but gets the job done. How this script works:
For more information on how to use this script, check it out on Github!
If you run into any problems or have any questions, post them in the comments below!
(Note: This post is neither sponsored nor endorsed by Rackspace.)
After an article popped up on Hacker News a couple months ago about the profitability of mining Litecoin on EC2, it inspired me to take Rackspace’s cloud servers for a mining spin. Rackspace doesn’t currently offer any GPU instances, so I had no expectation that they’d mine at the same level, but I was hoping that my plethora of servers that are being gently used for other projects can have their plentiful spare CPU cycles put to a bit of mining.
My initial test only focused on Rackspace’s new performance flavors. The graph of the hash rates vs the threads are below:
For those haven’t worked with Rackspace’s cloud server offerings, there are currently two types of servers: First Gen (pre-OpenStack servers) and Next Gen (OpenStack-based servers). Under Next Gen, the server types can also be broken down to Standard Instances and Performance Flavors (instances containing SSDs and higher bandwidth NICs) (more technical information about the servers can be found here). Because of the RAM vs CPU tradeoffs of the First Gen vs Next Gen servers, I decided to give the First Gen servers a shot with mining.
However, if you flip the graph to compare the cost to run the server to the KH/s it can generate, you get the following:
The cost-to-performance ratio of the 256MB First Gen server blows away the competition. This is due to the server’s four available processors (verses the one available in the 1GB Performance flavor, two in the 2GB Performance flavor, and so on). Since Litecoin mining app minerd has little to no dependency on the system’s RAM, there is a negligible impact by increasing the amount of RAM in the First Gen servers.
These experiments aside, does this matter? At the time I started writing this, my performance servers were achieving approximately 7.5 KH/s. This came out to about 0.115 LTC per month, so with a $35/LTC exchange rate, this came out to around $4 per month. Now that the price of LTC has dropped to around $15, this is only $1.73 or so. In comparison to the ~$29 per month that it costs to run the base 1GB RAM/single core performance flavor, it is by no means worthwhile to run servers exclusively for mining (even the $11/month 256GB First Gen server won’t get you to a break-even point). However, because this is a deflationary currency, a Litecoin price jump in the future could turn these minings into profit.
If you want to try these tests out on your own server on another provider, the scripts that were used are available here:
If you want to play with the data I generated from these tests, you can download it from here:
EDIT: I realized I didn’t list off the image type I was using for these tests. All of these tests were run under Ubuntu 12.04 LTS, with updates to the latest packages.
I’ve been diving back through CDs from grade school to back up and organize old data. Along the way, I came across an SVCD Steve compiled from a 2002 father-son trip to Las Vegas, Zion National Park, Bryce National Park, and the Grand Canyon. I was able to rip the MPEG2 file from the disk and re-encode it to a more web-friendly H.264.
You can check out the full-length video here. It’s more than 15 minutes in length (and might not be interesting to anyone other than myself), but a cool artifact to find none the less.
While attempting to read through my teams pull requests, I found that I was having trouble quickly skimming and identifying the posts just by my teammate’s usernames. I decided to put together a small Chrome extension to help mitigate this problem.
Github Pull Request Team Highlighter will automatically highlight pull requests by certain users when you navigate to the /pulls page for a repository. You can see the effects of the extension below:
Download the extension on the Chrome App Store: https://chrome.google.com/webstore/detail/github-pull-team-highligh/plbgniheajfkpjdipmnncjkbigjihfac
View the source on github: https://github.com/amussey/pull-request-team-highlighter
As a side note, I didn’t anticipate how provocative the pull request symbol looked when colored with rainbows.
With the recent inclusion of tasks into gists, I started to realize how dependent I’ve grown to be on the service. The other snipped-keeping app I’ve always relied on, Simplenote, continues to be great for day-to-day jotting. Somehow, the idea of combining the two came together, and I started exploring how to back up a Simplenote account to a git repo.
After not finding much, I put together a small set of scripts to do the task. These scripts will automatically download the latest copies of your notes and commit them to a repo. The structure of the repo is designed to be completely compatible with github’s gist service.
Visit the repo to get started! https://github.com/amussey/simplenote-to-git
My little sister Emily assembled her first computer last week! Here’s a time-lapse of the process.
Miss the announcement of The Space last month? Check out the recorded pitches from some awesome teams at Blacksburg’s latest social creative thinking space.
I recently received an update notification for my Samsung Galaxy Stratosphere II. Having some concern about losing my device’s ability to tether, I avoided the update for a a couple days. I finally gave in, but not before grabbing as much OS data as possible.
A new Verizon splash screen on boot was the first indication that there was going to be trouble. Once the update was done installing, FoxFi’s root-free wireless tethering ceased operation. Here is a comparison of the baseband information before and after the update:
Before & After:
If you heavily depend on this feature for day-to-day use, avoid this update as long as possible.
According to the Softpedia article regarding this update, the two major changes in version JZO54K.I415VRBMJ2 are:
As the moment, the client can be run against a standard VNC server. It has been successfully tested against a Windows Server 2012 box (hosted on a Rackspace Performance 1 Cloud Server). Testing has also been done against a Ubuntu 12.04 server configured using these directions, but there were problems with the decoding of the image client-side.
Furthermore, after the initial discolored frame, no additional frames would be rendered. I investigated into the rfb2 library for an extended time to find the source of these problems, but unfortunately had no luck.
The Windows Server interaction did not suffer from this same issue. However, it was not without its own problems. UltraVNC hit occasional compatibility issues with Aero, causing the screen to randomly blank out. There were also occasional issues with they keyboard transmitting each keypress multiple times.
The largest problem currently comes from the remote frame buffer library. Because each frame is transmitted as raw data and rendered as a PNG, the framerate peaks at about one frame every three seconds.
If you would like to play around with these projects for yourself, you can try them out by doing the following:
I’d love your feedback or recommendations on these explorations in the comments below or directed at amussey on Twitter. If you would like to link back to this post from inside Rackspace, feel free to use http://rax.io/reach-js-vnc to reach this post directly.