HPC Governance and Coordination Meeting

Friday 5 Jul 2024, 10:00 → 11:00 Europe/London

Agenda

Invited

Jianguo Rao, Kuda Nguruwe, Krishan Purahoo, Chris Morgan, Frederik Ferner, Kingsly Collie, Dan Short, Callum Harrower

Attendees

JR, NG, CM, CH, KP

Topics

Discussion on Governance

Shared governance between ISS, EIS and DSU.

EIS hardware and OS, ISS software, DSU Storage, Batch System, configuration management will be join 

For GPU, All request to JR

Current Status of Hardware

Current Status of Batch System

Current Status of CEPH

• KN has tested CEPH on login 3, including CEPHFS and S3
• KN to resend test results
• KP mentioned currently reserving 100TB of usable space for openstack, SdW suggested moving this to an erasure encoded pool
  • ND to try and set up EC pool

Currrent Status of Software Environment and OS

• Requests for new modules in a microsoft tracker: https://tasks.office.com/ukaea.uk/en-US/Home/Planner#/plantaskboard?groupId=604be0a4-3916-4060-a276-207f088c2185&planId=-MMgxU4z80K6OAry6gpWFpYAEpJz
• Currently deployment of MATLAB delayed by access to license server (ticket in system)
• Discussion on whether we should allow remote job submissions thru MATLAB Parallel Cluster
  • Generally not a good idea as will compete with SLURM, and not immediately needed
  • JR to train Joshua Mitchell on SLURM and include example MATLAB job submission scripts in documentation
  • SdW to look at whether we can get parallel cluster to talk to SLURM
• As there are many things in the request queue, agreed to have a separate discussion on priortisation
  • SdW to set up meeting
  • CH to look at adding request dates to existing and future requests.

Tools for moving data between Scratch and Persistent Storage

Cost Modelling and Service Levels

Ongoing Work

 

 

 

10:00 → 11:00 Agenda

Meeting Agenda and Minutes