Lustre 1.2 was released on March 9th with little fanfare, followed today by the minor update 1.2.1. Lustre 1.2.x is a significant improvement over the 1.0.x series, and contains dozens of bug fixes, performance improvements, and new features. Most user-visible among them are: - Integrated support for Linux 2.6 - Disk-full conditions are handled gracefully, and return the correct errors to your applications - Protocol optimizations for vastly improved stat() and read() performance while other clients are writing - Proper orphan handling to avoid leaking space after client/server crashes - Improved parallelism and lower overhead on the object storage servers allow for higher I/O bandwidths - Fixes related to running an IA-32 metadata server with IA-64 clients - Dozens of less user-visible bugs, particularly related to error handling and recovery A complete list can be found at http://www.clusterfs.com/changelog.html Lustre 1.2.x is available now to customers with a current CFS support contract, and will be made available to the general public later this year. A link to the download page can be found at http://www.clusterfs.com/lustre.html We are also pleased to announce new Lustre management tools (to be released on March 31) and a simplified support plan, to make it easy for new users to get help directly from Lustre''s designers. A low per-cluster annual agreement includes early access to the latest release, a license to the tools, a copy of the administration manual, discounts on classroom training, and 10 hours of support. More information is available at http://clusterfs.com/techsupport.html For the Lustre team, -Phil