NCI Help

Page tree

On this page

Q3 Maintenance Downtime

NCI systems will be offline for scheduled maintenance from 6:00 am to 6:00 pm  AEST Thursday 1 August 2024. 

Q3 scheduled maintenance will include updates to:

  • Gadi compute nodes
  • Lustre file systems: includes /scratch and all /gdata file systems
  • Cloud
  • ARE and data services
  • Network infrastructure
  • Support ticketing services

User support will not be available during the maintenance downtime. NCI recommends that users watch the Gadi message of the day for timely status information during the maintenance downtime.

Gadi - How to prepare

Gadi users can do the following to prepare for this scheduled downtime:

  • Run any significant Gadi jobs needed to support milestones or deadlines as soon as possible during April.
  • Back up or move any important data in /scratch directories to appropriate longer-term storage locations, e.g. project directories on /gdata before 31 July.
  • Pause any automated scripts used for routine tasks involving Gadi, /scratch, /gdata or cloud or data services, and develop a plan to restart these after completion of maintenance. 
  • Plan for interruption to any services which rely on Lustre file systems: /scratch, /gdata. 

Gadi job queues will be drained in advance of the scheduled downtime. Queued jobs that would normally be scheduled to run during the maintenance period will be held until Gadi resumes normal operation. 

Nirin Cloud - How to prepare

All virtual machines running on the NCI Nirin Cloud service will be shut down no later than 6:00 am AEST on 1 August. 

If your project VMs have application dependencies, for example, a database service which should be shutdown before a presentation service, the NCI Cloud team requests that you perform a controlled shut down of all VM resources before 6:00 am AEST Thursday 1 August. All projects will need to plan to restart all VM services via the Nirin Dashboard (https://cloud.nci.org.au) after completion of the maintenance downtime. (Note that automated restart via metadata tags is no longer supported for Nirin cloud.)

Before the scheduled downtime starts, any cloud instances with gdata NFS mounts should be shut down. Instances still running while the gdata file systems are unavailable will need to be rebooted when the maintenance outage is completed.

Instances with volumes attached should be shut down before the scheduled downtime. Any instances with volumes which are still running on the morning of the outage will beed to be rebooted after completion of maintenance.

Instance owners are responsible for bringing their instances back up after completion of maintenance. Instances will not restart automatically. The NCI Cloud Team will not restart user instances. 

ARE - How to prepare

All ARE services will be shut down prior to the maintenance downtime at 6:00 am AEST Thursday 1 August. ARE services will be automatically restored at completion of the maintenance downtime.

User Support

Contact NCI user support if you have specific questions about how to prepare for the scheduled maintenance downtime: help@nci.org.au .



  • No labels