Building a Linux-Based High-Performance Compute Cluster

The Rocks clustering package from the University of California at San Diego makes it easy to build and maintain a high-performance compute cluster with off-the-shelf hardware.
Step 11. Configure the Disk Partitioning

The final interactive screen of the installation sequence (Figure 12) is the disk-partitioning screen. You can partition the disks automatically or manually. If you go with the automatic partitioning scheme, the installation routine sets up the first disk it discovers as follows:

Partition                         Size
/                                 16GB
/var                              4GB
swap                              Equal to RAM size on the head node
/export (aka /state/partition1)   Rest of root disk

If you have multiple disks on the head node or you want to arrange the disk in a different fashion, select Manual Partitioning. This takes you to the standard Red Hat manual partitioning screen where you can configure things any way you desire (you still need to have a 16GB / partition and an /export partition at minimum though). Clicking Next on the disk-partitioning screen begins the automatic portion of the installation (Figures 13, 14 and 15). Once installation is complete, the head node reboots, and you are greeted with your first login screen (Figure 16).

Figure 12. Disk Partitioning

Figure 13. Rocks Installation, 1

Figure 14. Rocks Installation, 2

Figure 15. Rocks Installation, 3

Figure 16. Login Screen

Step 12. Login

Log on as root, and wait for two or three minutes. This lets the remaining configuration routines finish setting up the cluster in the background. Start a terminal session (Figure 17) to begin installing the compute nodes.

Figure 17. Root Terminal

Step 13. Install a Compute Node

Now you're ready to add nodes to the cluster. The Rocks command that accomplishes this is insert-ethers. It has quite a few options, but for this example, use the main function of inserting nodes into the cluster. After you invoke insert-ethers, you are presented with the screen shown in Figure 18.

Figure 18. insert-ethers

Rocks treats everything that can be connected to the network as an appliance. If it can respond to a command over the network, it's an appliance. For this simple example with a dumb switch, the only things you need to worry about are the compute nodes themselves. Because Compute is already selected, tab to the OK button and press Enter. This brings up the empty list that will be filled with the names and MAC addresses of the nodes as they are added (Figure 19).

Figure 19. List of Installed Appliances (Empty)

Step 14. Boot First Compute Node

Now it's time to boot the first compute node. If you have wiped the disk, most systems will start a PXE boot from the network as a default action. If you have a KVM switch and can watch the console on the compute node, you should see the PXE boot begin. When the compute node asks for an address for eth0, you will see the MAC address entered in the Inserted Appliances list on the head node (Figure 20).

Figure 20. List of Inserted Appliances (First Node Added)

The insert-ethers routine displays the MAC address it has received and the node name it has assigned that node. The ( ) will be filled in with an asterisk (*) when the compute node begins downloading its image (Figure 21).

Figure 21. First Node Installing

______________________

Webinar
One Click, Universal Protection: Implementing Centralized Security Policies on Linux Systems

As Linux continues to play an ever increasing role in corporate data centers and institutions, ensuring the integrity and protection of these systems must be a priority. With 60% of the world's websites and an increasing share of organization's mission-critical workloads running on Linux, failing to stop malware and other advanced threats on Linux can increasingly impact an organization's reputation and bottom line.

Learn More

Sponsored by Bit9

Webinar
Linux Backup and Recovery Webinar

Most companies incorporate backup procedures for critical data, which can be restored quickly if a loss occurs. However, fewer companies are prepared for catastrophic system failures, in which they lose all data, the entire operating system, applications, settings, patches and more, reducing their system(s) to “bare metal.” After all, before data can be restored to a system, there must be a system to restore it to.

In this one hour webinar, learn how to enhance your existing backup strategies for better disaster recovery preparedness using Storix System Backup Administrator (SBAdmin), a highly flexible bare-metal recovery solution for UNIX and Linux systems.

Learn More

Sponsored by Storix