Update 6th April 4pm
All clusters back online.
Update 6th April 9am
The clusters will remain offline until lunchtime today (6th April). Some additional issues were identified during the service window which need to be addressed before the systems can be brought back online.
There will be a service downtime for all TCHPC clusters (Lonsdale, Parsons, Tcin, Kelvin) on next Tuesday (5th April).
This downtime window will be used to add some additional inter-cluster cabling to reduce possible network contention, and perform some server maintenance tasks which involves un-mounting the shared filesystems (/home and /gscratch) from all nodes.
We expect that service will resume on Wednesday morning.
Logins will not be available on the headnodes during this service window.
In the meantime, jobs can still be queued and will start up again once the reservations are released and services are resumed.