Qlustar is a full-fledged Linux distribution designed for the purpose of an all-in-one Cluster Operating System. It’s core is based on the most open Linux platform: Debian/Ubuntu. Qlustar can drive all of your clustering environments with efficiency, ease and reliability. Many innovative features make it a unique solution in the field. It is real-life proven, being the software engine for a large number of Linux HPC and Storage clusters running in industry and academia.

When choosing Qlustar, you can look forward to a platform that is capable of serving all compute requirements from basic infrastructure services to storage farms and high-performance computing. Easy to setup, extend, operate, monitor and update. Reliable, performant, secure, open and based on standards.

The technology behind Qlustar is constantly improved, while assuring that smooth updates allow moving to the most recent release when desired. If we are convinced, they add value to the overall solution, powerful new clustering and general Linux features are integrated as soon as they are ready for productive use.

The basis of any Qlustar system is the Qlustar Cluster OS. Depending on your use-case, you have the option to add functionality from the large portfolio of supplementary Qlustar components.

The Qlustar Platform

Platform

Qlustar consists of a single Core and multiple Edge platforms. By platform, we mean a specific Linux distribution, possibly for multiple CPU architectures.

The core platform provides all Qlustar components/packages including the installer. It is currently (and will be in the foreseeable future) based on Ubuntu LTS releases. On the other hand, edge platforms are not meant to be run on head-nodes and lack certain software components or packages (head- / service-node functionality).

Read more …

The Qlustar Cluster OS

Platform

Qlustar’s core is a lightweight, modular Compute/Storage-Node OS kernel that is generated using a 3rd-generation image generator. The latter is designed such that OS images contain the exact minimum of programs/files needed for the desired functionality.

Qlustar OS images are modular, composed of the core module needed by any compute or storage node and optional modules for additional functionality (e.g. to integrate into a particular batch queuing system or to activate support for Infiniband, parallel filesystems, High-Availability or GPU computing, etc.).

Read more …

The Qlustar HPC Stack

Platform

The Qlustar HPC Stack, optionally based on Ubuntu or CentOS, contains all components necessary, to render a networked server farm based on off-the-shelf hardware into a high-performance compute cluster. It provides it’s own software repositories with 150+ additional up-to-date software packages relevant to HPC (e.g. different MPI variants in different versions for various compilers, monitoring tools, etc.). The CentOS edge platform also integrates OpenHPC, providing a wealth of up-to-date ready-to-use HPC packages.

Qlustar’s workload management provides the premier open-source software Slurm to achieve optimal use of your compute resources. To complement Slurm’s superb feature set, QluMan provides a powerful GUI management front-end to Slurm.

Read more …

The Qlustar Data Stack

Platform

Fast storage is vital for most clusters these days. Qlustar supports ZFS/Lustre and BeeGFS, the most popular open-source parallel file-systems, out of the box - including high-availability. Just create an image containing the Lustre or BeeGFS module, assign it to your storage nodes, format your disk devices and you’re done.

For scalable and highly available block or object storage needed by your private cloud, Qlustar provides the proven Ceph Storage System. It delivers fast access to block device images being striped and replicated across a sub-cluster of specifically assigned storage nodes.

Read more …

The Qlustar HA Stack

Platform

Any cluster should have a maximum utilization. Failing head or storage nodes can make it useless without a high-availability (HA) setup. With redundant hardware, Qlustar is capable to provide it’s services fully protected from SPOFs (single point of failure) by setting up proven HA configurations.

Read more …