Custom distro hammers hardware to detect system failures

Advanced Clustering Technologies releases Breakin 3.20

One of the things I love about Linux? You can do so many interesting things with it. Take, for example, Breakin — a stress-testing and hardware diagnostic tool that will put hardware through its paces to make sure it's production-ready. The 3.2 release was announced today with the addition of several new utilities and improvements to its user interface.

Breakin is a distribution to run diagnostics on systems to find hardware issues or component failures. Put together by Advanced Clustering Technologies, Breakin is a Linux-based live CD that tests memory, the CPU, hard drives, (supported) temperature sensors, and looks for any Machine Check Exception (MCE) errors generated during the tests.

Hard to use? Nope. The default mode doesn't require any feedback. Pop it in, and let it run. It runs from CD or DVD, USB keys, or if you do a lot of hardware testing you can provide the image over the network for hardware that supports PXE booting. They recommend letting it hammer the hardware for at least 24 hours, but longer is better. You can even peek in on the testing results remotely via SSH — so this is a nice little tool to have if you're renting dedicated servers in a remote facility.

It's not all about hammering hardware, though. Breakin also includes a number of utilities so it can be used as a rescue environment as well. If you're using Breakin to diagnose hardware problems on systems in production, this can come in fairly handy.

Why is Advanced Clustering Technologies developing Breakin? ACT president and CEO says that the company had looked around for a tool that would "adequately" test hardware before they ship it to customers — but they struck out. "Most commercial software you find for this is very inadequate — many are still based on DOS, others are Windows-based and require you to install an OS first before running the test. Our needs were something that could be automated into our production environment, and something we could re-distribute to our customers in case servers starting having problems once deployed."

Breakin is (almost) fully open source and can be downloaded as a pre-built image, or grab the source using git and modify and build it yourself. (It uses the Intel MKL library and AMD's Core Math Library, which are not open source.) The company is hoping for feedback from the community on its performance on a wider range of hardware. It's tested on ACT's servers and workstations, but should also work on a wide range of Intel/AMD servers.


Copyright © 2011 IDG Communications, Inc.

The 10 most powerful companies in enterprise networking 2022