PRESS RELEASE

BeeGFS 2015.03-r1

A new major release of the parallel file system BeeGFS has been announced by Fraunhofer ITWM and ThinkParQ. After several months of public beta testing, the new version 2015.03-r1 has been officially released, with the introduction of high availability for storage servers as an enterprise feature. Based on data replication across different storage servers, automatic failovers will happen in case of storage server failures and self-healing will resynchronise the server when it comes back.

Additionally, in this release BeeOND (BeeGFS on demand) was added as a stand-alone package. BeeOND provides a tool that can create complete BeeGFS instances as and when needed with a single command. While this functionality was basically already usable in previous releases, usability was considerably improved now. Within the BeeOND package, users can not only find tools to create and destruct a file system, but also tools to perform a parallel copy of data between file systems (e.g. a global cluster storage and a per-job BeeOND instance).

Furthermore, BeeGFS now supports extended attributes, as well as access control lists (ACLs). In addition to the optional GUI-based setup of BeeGFS, a general usability improvement is the introduction of new setup tools, which allow command-line based setup of the BeeGFS services without the need to edit configuration files. To optimise performance of multi-target storage servers under high load, the storage server worker threads are now grouped and dedicated to the individual storage targets, resulting in better balance and fairness of parallel client request handling.

Other performance improvements include a new aggressive low-latency mode, metadata access optimisations and enhancements for many-core servers. As before, the file system client native kernel module is still compatible with a wide range of Linux kernel versions, starting with 2.6.18, while the new version adds support for the recent 4.0 kernel.

Feature

Robert Roe reports on developments in AI that are helping to shape the future of high performance computing technology at the International Supercomputing Conference

Feature

James Reinders is a parallel programming and HPC expert with more than 27 years’ experience working for Intel until his retirement in 2017. In this article Reinders gives his take on the use of roofline estimation as a tool for code optimisation in HPC

Feature

Sophia Ktori concludes her two-part series exploring the use of laboratory informatics software in regulated industries.

Feature

As storage technology adapts to changing HPC workloads, Robert Roe looks at the technologies that could help to enhance performance and accessibility of
storage in HPC