The release of the Lustre 2.17 cluster file system, used in most of the largest Linux-clusters containing tens of thousands of nodes. Lustre's key components include metadata processing and storage servers (MDS), management servers (MGS), object storage servers (OSS), object storage (OST, which supports running on ext4 and ZFS), and clients. The project's code is distributed under the GPLv2 license.
Main innovations:
- A hybrid I/O performance optimization mechanism is enabled by default, automatically applying either buffered or direct I/O mode depending on the nature of the load and the size of the data being transferred.
- Added support for dynamic configuration of the Nodemap engine, which is responsible for mapping user identifiers (UID/GID) in multi-tenancy configurations in which different groups of clients have access to the file system.
- The dynamic configuration system of LNet NID (LNet Network Identifier) ββnow correctly processes attempts to register on server more than 32 NIDs (previously this situation led to a failure, but now NIDs over the limit will be ignored).
- Support provided Server for the kernel package from RHEL 9.7 (5.14.0-611.13.1.el9_7), and clients for unmodified kernels of RHEL 9.7 (5.14.0-611.13.1.el9_7), RHEL 10.1 (6.12.0-124.20.1.el10_1), SLES15 SP7 (6.4.0-150700.53.22.1) and Ubuntu 24.04 (6.8.0-35).
Source: opennet.ru
