To ensure our research computing platforms remain secure and stable, we are establishing a regular weekly maintenance and update window starting from Tuesday 3rd November.
Moving to a predictable schedule allows our systems team to apply essential patches regularly. This reduces the need for short-notice, emergency interventions.
What is changing?
An at-risk maintenance window will take place every Tuesday morning between 07:00 and 09:00.
During this slot, we will deploy operating system updates, apply security patches and implement general platform improvements. We have aligned this timing with the standard Tuesday at-risk period widely adopted across the Jisc network. This ensures any potential disruption is confined to hours of historically low demand.
What does “at risk” mean?
Services operating within an at-risk window remain active, but they may experience brief interruptions:
- Systems will not automatically go offline every Tuesday. Most of the time, you will notice no disruption.
- Occasionally, applying a specific software patch will require a service restart, a configuration change or a short outage.
Why this change is necessary
Modern infrastructure demands continuous attention to protect against new security threats.
Formalising this window delivers several practical benefits:
- Faster deployments: Security updates and vendor fixes reach the platform without unnecessary delay.
- Fewer emergencies: Scheduled upkeep reduces the likelihood of disruptive, out-of-hours interventions.
- Better predictability: Researchers and technical staff gain a dependable timetable to plan their workloads around.
- Long-term stability: Routine maintenance prevents complex software compatibility issues from building up across core systems.
Which services are covered?
This window applies to all BEAR services provided by Advanced Research Computing services.
This encompasses compute nodes, storage arrays, access points and the underlying infrastructure.
Some of the services covered by this maintenance window include:
- Research Data Store (RDS)
- BlueBEAR
- CLIMB
- BEAR Archive
- BEAR Data Transfer
- BEAR Cloud
- BEAR Portal
- BEAR GitLab
If a planned upgrade is ever expected to cause substantial, prolonged downtime beyond a brief restart, we will continue to issue specific advance warnings via our standard communication channels.
We appreciate your cooperation and support as we implement this standard to protect our services.