Our users may have noticed that the hpc cluster dashboard is reflecting some infrastructure changes. Please note that this post refers to the older hpc cluster, not hex. The 200 series are going to be decommissioned soon, and this is a good opportunity to remind users that there is a data centre shut-down from Saturday 21 June to Sunday 22 June. Your jobs need to be terminated by the evening of Friday 20 June.
Right now we are deploying a test series, Series500 linked to a new queue, AQ. We have also deleted several older, unused queues, batch, UCTshort and GPUQ as the gpu servers moved from hpc to hex. We have enabled queue affinity linking UCTlong to Series400 and AQ to Series500. ACL groups are used to control user permissions. The default queue and series for all new users on hpc, is UCTlong and Series400. If your scripts do not reference a queue that's fine, your jobs will run automatically on UCTlong and series400.
When will the Series500 be available? In a few weeks we hope, keep watching this blog.
What will distinguish it from Series400? The worker nodes will be running the latest version of Scientific Linux, 6.5, and will hopefully be linked to a large pool of fast storage disk.
Why are we sticking with SL6.5? To provide OS support for applications that run on Red HatCentos as SL is a fork of RH.
Wait, what large, fast storage? Can't say yet, sorry. Watch this space.