HA-OSCAR: the Birth of Highly Available OSCAR

As clusters reach thousands of nodes, eliminating single points of failure becomes critical. The Open Cluster Group's HA-OSCAR is a solution.

The seeds for the Open Cluster Group (OCG) and the resulting Open Source Cluster Application Resources (OSCAR) Project were planted when a number of like-minded individuals had the good fortune to sit together for dinner at a meeting sponsored by the Department of Energy on February 17, 2000. Over dinner, this group discussed the effort involved in deploying the software necessary to build a Beowulf high-performance computing cluster. Although this group agreed that it was simple to wire together commodity computers to build the cluster, they went on to agree that the amount of effort required to install and configure the requisite software stack on a Beowulf cluster was inordinately high. Challenging to some and tedious to all was the consensus description. Thus, the idea to simplify the process was born.

Further fertilization of this idea took place in Oak Ridge, Tennessee, in April that year at a meeting that included industry, academia and research labs. It was at this first formal meeting that the OCG was formed and work was begun on OSCAR, specifically the Beowulf cluster software stack and associated installation process. At this meeting the group agreed on three core principles:

  1. The adoption of clusters for mainstream, high-performance computing is inhibited by a lack of well-accepted software stacks that are robust and easy to use by the general user.

  2. The OCG embraces the open-source model. As a result, the OSCAR distribution must contain only freely redistributable codes, with a preference for the inclusion of source code under a Berkeley-style open-source license.

  3. The OCG can accomplish its goals through the use of best-practices codes currently available.

Further details about the beginning of OCG and OSCAR can be found in an article by Richard Ferri in the June 2002 issue of Linux Journal titled “The OSCAR Revolution”.

Throughout OSCAR's brief history, the group has managed to adhere to these three principles while allowing OSCAR to encompass other forms of software. For example, although the OSCAR distribution itself contains only freely redistributable codes, others are deploying an OSCAR package that, while not a part of the formal OSCAR distribution, may be installed or dropped into an existing OSCAR distribution for installation.

The OSCAR Project continues to contain a mixture of industry and academic/research members. The overall project is directed by a steering committee elected every two years from the current core organizations. This core list is composed of those actively contributing to project development. The 2003 core organizations include: Bald Guy Software (BGS), Dell, IBM, Intel, MSC.Software, Indiana University, the National Center for Supercomputing Applications (NCSA), Oak Ridge National Laboratory (ORNL) and Université de Sherbrooke (UdS).

In the past year, additional OCG working groups have been created to address other cluster environments. These new groups are working to leverage the technology provided by OSCAR when producing their cluster distributions. The two groups working today are Thin-OSCAR and HA-OSCAR. Thin-OSCAR is headed by the Université de Sherbrooke in Canada and is dedicated to delivering a diskless variant of OSCAR. The HA-OSCAR group is led by the authors of this article and is focused on providing a high-availability version of OSCAR.

HA-OSCAR: Mission, Goals and People

In a July 2001 meeting at Ericsson Research Canada, Ibrahim Haddad made the case for high availability in cluster computing. Initially, the discussion centered on the necessity of high-availability computing for the telecom industry. As the discussions progressed, it became clear that with the anticipated tens of thousands of nodes in high-performance computing (HPC) clusters, high-availability techniques can provide some level of the fault tolerance desired by the HPC community.

Ibrahim's group at Ericsson Research worked primarily alone on the high-availability effort until the recent addition of Dr Chokchai Leangsuksun and his team at Louisiana Tech University and the continued interest in HA-OSCAR by Stephen Scott at ORNL. In 2002, the HA-OSCAR effort was recognized officially by OCG as another working group. The primary goal of the group is to leverage existing OSCAR technology and provide for new high-availability capabilities in OSCAR clusters. The anticipated customers of this technology include the telecom industry and HPC sites.

HA-OSCAR Improvements over OSCAR

HA-OSCAR introduces several enhancements and new features to OSCAR, mainly in areas of availability, scalability and security. Most of these features can be mapped to ITU (International Telecommunication Union), TMN (Telecommunication Management Network) and FCAPS (Fault-management, Configuration, Accounting, Performance and Security). These concepts are widely adopted in the telecom industry to manage its network elements.

______________________

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