Skip to main content

IT Project: Quest Scheduling Improvement

Project Status

May 1, 2019: The new scheduler is in production.

Workshops and Training Opportunities

Project Charter

Background

Northwestern IT has identified growing insufficiencies in the current resource manager and accounting processes used by Quest, the University’s high-performance computing cluster. The project team conducted an initial gathering of requirements and consulted with members of the user community to validate the requirements and establish priorities. Based on an extensive set of research and technical requirements, Slurm was chosen as the preferred scheduler solution.

The implementation of a new Quest scheduler will positively impact the research community by creating a streamlined accounting process and efficient management of computational resources, in which users can conduct their research more effectively and the resources of Quest will be utilized more efficiently.

This project is a collaboration between Northwestern IT teams: Research Computing Services and Research Computing Infrastructure.

Goals and Objectives

Approach

Researchers using Quest have been invited to participate in a pilot of the new scheduler and accounting model. During the pilot, participants are asked to provide feedback which Northwestern IT will use to identify and resolve underlying issues prior to opening a testing environment.

The testing environment will be open to all researchers using Quest for a two-month period to test and convert their batch submission scripts. During this time, Northwestern IT will closely monitor the scheduler environment to surface and remedy issues, help users convert their scripts, and optimize the allocation model. During this time, Northwestern IT will provide researchers with detailed information to help transition to the new environment through email messages, online documentation, and workshop training opportunities.

Quest will switch to the new scheduler on May 1, 2019. Research Computing Services staff will be available throughout the entire transition period for assistance.

Project Timeline

Description Target Date
Status
Pilot February 4, 2019 Complete
Test environment available to all Quest users March 1, 2019 Complete
New scheduler in production May 1, 2019 Complete
Outgoing scheduler retired May 31, 2019 In progress

Project Contacts

Name Title Role
David Carr Associate Vice President - Cyberinfrastructure Executive Sponsor
Joseph Paris Director, Research Computing Services Sponsor
Steve Kwak Director, Cyberinfrastructure - HPC Sponsor
Matthew Rich Senior Cloud Services Specialist Project Manager
Jackie Milhans Manager, Research Computing Services Key Stakeholder
Damir Krstic Manager, Cyberinfrastructure-HPC Key Stakeholder
Alex Mamach Senior Systems Administrator, Cyberinfrastructure-HPC Technical Lead

Last Updated: 9 May 2019

Get Help Back to top