Building a Linux-Based High-Performance Compute Cluster
Select one machine to be the head node for the cluster. The rest of the machines will be the compute nodes in your new cluster. As it installs the compute nodes, Rocks numbers the machines as compute-x-y, where the x is the rack number, and y is the number of the machine within a rack. Say you spread the 32 nodes over four racks. If you want to follow the Rocks naming convention, you would set things up as follows: rack 0 contains the head node, so the numbering of the nodes would be compute-0-0 to compute-0-6. Rack 1 would come out as compute-1-0 to compute-1-7. Rack 2 would contain compute-2-0 to compute-2-7. Rack 3 would follow suit.
Alternatively, you simply could pretend that all the machines are in a single rack: compute-0-0 to compute-0-30. Either way works, so use whatever is comfortable for you.
First, you need to get a copy of the Rocks package that will be appropriate for your cluster's hardware. Navigate to the Rocks Web site, and select the Download tab at the top of the home page to access the various versions of the package. The 5.1 version is the latest at the time of this writing. Click the link to get a listing of the components of the package. For this exercise, I selected the x86-64 Jumbo DVD image, downloaded it and burned it onto an empty DVD. While you are at the site, download the documentation. If nothing else, it will give you something to read while the software loads.
Boot the head node from the newly minted Rocks DVD. If everything is working as it should, you will be greeted with the welcome screen shown in Figure 2.
Enter build at the boot: prompt to start the installation sequence. The system boots in the normal Linux fashion and eventually presents the user with the initial Rocks configuration screen shown in Figure 3.
A Note about Rolls
The Rocks package is composed of a series of rolls. Some of the rolls are central to the cluster system itself (the Base Roll, the OS Roll, the Kernel Roll and the Web Server Roll). Some contain cluster functionality (the SGE Roll, the Java Roll, the HPC Roll and the Ganglia Roll). And finally, some rolls contain application software (the Bio Roll). Each roll is documented, so you can decide for yourself whether you need it in your installation. The commercial version of the Rocks package, Rocks+ from Clustercorp, adds additional packages that include commercial compilers from Absoft, Intel and the Portland Group as well as the TotalView debugger.
Because you have all of the software components on the Jumbo DVD, you will do your installation from CD/DVD-Based Rolls. Select CD/DVD-Based Roll. This brings up a screen listing all the individual components you can select from the DVD (Figure 4).
For the purposes of this installation, I selected everything except the Bio Roll and the Virtualization Roll. You probably will select a different set of components. At absolute minimum, you need to select the Base, Web Server, Kernel and OS Rolls. Once you have made your selections, click Submit to continue the installation.
The installation now repeats the first screen, showing your selections (Figure 5). If you are satisfied with your selections, click Next to continue to the first of the administrative screens in the installation. If you want to make a change, click CD/DVD-Based Roll to go back to the component selection screen.
Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.
Sponsored by AMD
If you already use virtualized infrastructure, you are well on your way to leveraging the power of the cloud. Virtualization offers the promise of limitless resources, but how do you manage that scalability when your DevOps team doesn’t scale? In today’s hypercompetitive markets, fast results can make a difference between leading the pack vs. obsolescence. Organizations need more benefits from cloud computing than just raw resources. They need agility, flexibility, convenience, ROI, and control.
Stackato private Platform-as-a-Service technology from ActiveState extends your private cloud infrastructure by creating a private PaaS to provide on-demand availability, flexibility, control, and ultimately, faster time-to-market for your enterprise.
Sponsored by ActiveState
| Speed Up Your Web Site with Varnish | Jun 19, 2013 |
| Non-Linux FOSS: libnotify, OS X Style | Jun 18, 2013 |
| Containers—Not Virtual Machines—Are the Future Cloud | Jun 17, 2013 |
| Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer | Jun 12, 2013 |
| Weechat, Irssi's Little Brother | Jun 11, 2013 |
| One Tail Just Isn't Enough | Jun 07, 2013 |
- Speed Up Your Web Site with Varnish
- Containers—Not Virtual Machines—Are the Future Cloud
- Linux Systems Administrator
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- RSS Feeds
- Senior Perl Developer
- Technical Support Rep
- Non-Linux FOSS: libnotify, OS X Style
- UX Designer
- Web & UI Developer (JavaScript & j Query)










5 min 54 sec ago
10 min 26 sec ago
2 hours 56 min ago
3 hours 13 min ago
4 hours 29 min ago
5 hours 18 min ago
5 hours 21 min ago
5 hours 30 min ago
6 hours 1 sec ago
8 hours 26 min ago