TCHPC will be having a Downtime Window for the week of 27th June - 5th July 2011.

Update 4th July

Due to the electrical work taking longer than anticipated, the TCHPC clusters will remain offline until the 5th of July, rather than the 1st as previously advertised.

Downtime details

This is to facilitate electrical work in the TCHPC machine room, and also to allow for filesystem re-organisation and rationalisation (which will affect all users who have files on /home -- see below for more).

Please note that in particular:

Move towards full resource allocation

Please note that some of the systems that TCHPC manages will be moving towards a full resource allocation model, please review the cluster scheduling policy. Users will be required to have project codes for job submission.

Lonsdale Cluster Queue

Please note that software updates will be applied to the Lonsdale cluster queuing software during this period.

This will mean that any jobs left queued up will be lost.

After the upgrades, the Lonsdale queue will be managed entirely by SLURM (the maui/GOLD installations are being replaced by SLURM functionality). This will result in a cleaner interface for users, with a much more responsive queue.

Filesystem Re-Organisation

Please note that TCHPC's new data policies will come into effect after the downtime window, including a major overhaul of the /home filesystem which is on lonsdale/parsons. See our data policies page for full details.