Skip to content

hpc-gridware/clusterscheduler

Repository files navigation

Open Cluster Scheduler

We are pleased to announce that we will continue the development of the Univa Open Core Grid Engine, which in turn is based on the open source Sun Grid Engine originally published by Sun Microsystems. Renamed "Open Cluster Scheduler", this project will remain open source under the SISSL v2 licence, with the source code available here.

In addition to maintaining the open source version, we will introduce advanced features under the name "Gridware Cluster Scheduler". This iteration will be offered with commercial support and consulting services from HPC-Gridware GmbH. The new features will be released under the Apache Licence v2.0.

Choosing the Code Base

After careful consideration, we have selected the Univa code base as our foundation. Despite the availability of alternatives like Open Grid Scheduler and Some/Son of Grid Engine, the Univa version aligns best with our goal of making significant codebase improvements. Our inaugural release — version 9.0.0 — will mark a modernization effort, aligned with the ongoing growth in HPC and AI workloads which demand faster, more flexible computing clusters.

Technological advancements in CPU and GPU design necessitate advanced decision-making algorithms for schedulers. We aim to tackle these challenges, ensuring the Cluster Scheduler codebase can effectively support modern multi-core CPUs, GPUs, NPUs, and FPGAs.

Roadmap

Our immediate focus is on laying a robust foundation for the future of cluster schedulers. The following updates will be part of the initial "Open Cluster Scheduler" release:

  • Codebase Modernization: Convert the code base to C++ and CMake.
  • Development Environment Support: Facilitate modern development environments (e.g., CLion).
  • Data and Threading: Update internal data stores and threading mechanisms.
  • Concurrency Enhancements: Improve concurrent execution within the master service for better thread parallelism. Key enhancements planned for version 9.0.0 include:
  • Resource Management: Implement RSMAPs for host-specific resource management, such as GPUs and other accelerators.
  • Hardware Integration: Integrate with the hwloc library for hardware topology and architecture analysis.
  • Architectural Support: Extend support for diverse computer architectures, including Linux on Intel/AMD64 (lx-amd64), AArch64 (lx-arm64), OpenPower (lx-ppc64le), RISC-V (lx-riscv64), Apple’s ARM-based CPUs (darwin-arm64), Solaris on Intel/AMD64 (sol-amd64), and FreeBSD for Intel/AMD64 (fbsd-amd64).
  • Usage Reporting: Enhance online usage reporting and introduce customizable accounting values.
  • Security Measures: Implement request limits to guard against denial-of-service attacks.
  • Container Builds: Facilitate container-based builds. To further improve usability and maintainability, we plan to remove outdated or rarely-used components:
  • GUI Modernization: Discontinue the old Motif-based GUI.
  • Shell Simplification: Remove qtcsh, aligning with other commercial schedulers.
  • Code Simplification: Phase out complex components like the JGDI interface and its associated services.
  • CSP Mode: Temporarily suspend CSP mode due to low user adoption.

This roadmap outlines our strategic goals to modernize, expand, and streamline the Open Cluster Scheduler, setting the stage for future innovations and enhanced performance.

Join the Conversations

We would love for you to join us in this exciting project! Please share your questions, suggestions, or interest in contributing to this project. You can reach out to us via the public users Google Group. If you are just curious about the project's progress, you can subscribe to the Git Google Group for our project.

Commercial Offering

If you're ready to test the Gridware Cluster Scheduler, which is available with extensions and for which we provide commercial support, then get in touch! We can't wait to hear from you!

Daily Build

A daily build of the Open Cluster Scheduler's master branch can be downloaded from https://nightly.link/hpc-gridware/clusterscheduler/workflows/daily_build/master/daily-build.zip.

Please be aware that this build is done from the main development branch. It is not a stable and QAed release. Use it for testing purposes but not in a productive environment.

The daily build zip archive contains

  • the binary package for lx-amd64 binaries
  • the common package containing architecture independent files
  • the doc package containing man pages and basic documentation

Other Documents

License

The original files contributed by Sun Microsystems are licensed under

Newer contributions from HPC-Gridware GmbH (and others like Univa Inc. or individual contributors) are licensed under

Distribution packages might contain support for 3rd-party tools and libraries. License Information for corresponding components is available under

There is NO WARRANTY, to the extent permitted by law.