Release of cluster fs lustre 2.15

published the release of the cluster file system lustre 2.15 , used for the most part the largest linux-claster containing tens of thousands of nodes. The key components of Lustre are the processing and storage servers of metadata (MDS), control servers (MGS), object storage servers (OSS), object storage (OST, work on over EXT4 and ZFS) and customers are supported. Project code is distributed under the license gplv2.


The main innovations :

  • implemented mode
    Client Directory Encryption, which allows you to encrypt the names of files and catalogs on the client side, at the stage before transmitting data on the network and before saving in storage facilities (OST) and metadata (MDT).
  • added UDSP (user defined Selection Policy), which allows users to determine the rules for choosing network interfaces for data transfer. For example, if there is a connection through the Networks o2ib and TCP you can configure the Lustre traffic transmission only after one of them, and the second is used For other needs.
  • Support for a server for a package with a core from RHEL 8.5 (4.18.0-348.2.1.el8), and customers for unmodified nuclei RHEL 8.5 (4.18.0-348.2.1.el8), Sles15 SP3 (5.3. 18-59.27) and Ubuntu 04.04 (5.4.0-40).
/Media reports.