Lustre Parallel Filesystem

Lustre Parallel Filesystem

Lustre is the most widely used parallel filesystem in HPC. It is POSIX compliant and highly scalable in capacity as well as throughput. Its architecture is based on the idea of the separation of meta- and object storage data, where each of these are stored in separate so-called targets (MDTs and OSTs). These targets may be distributed among multiple storage nodes thus achieving scalability.

Qlustar supports Lustre for more than a decade, so its users benefit from the experience of a large number of installations. The Qlustar implementation is based on the Lustre community releases with long term support (LTS).

Qlustar Lustre Features

Qlustar Lustre exclusively uses ZFS as the backend filesystem. This allows to take advantage of all its data security and protection features as well as transparent data compression.

With QluMan’s RXEngine and its command editor you have the possibility to define custom commands specifically acting on Lustre nodes. These can then be executed conveniently within the QluMan GUI. Furthermore, by defining corresponding user roles, it easily becomes possible to delegate standard maintenance tasks to less experienced admins without the need to hand out full control. Managing and operating your Lustre clusters can’t become any simpler and more controlled.

If a hardware setup is available that supports it, Qlustar Lustre can be run in high-availability (HA) mode. This is achieved by specially developed pacemaker resource types usable within a Qlustar corosync/pacemaker setup. Contact us, if you need assistance in designing, setting up or supporting proven Lustre (HA) configurations.

^{Lustre is a registered trademark of OpenSFS and EOFS.}