Spring Semester HPC Downtime (April 14 – 17)

Beginning at 9:00 a.m. on Monday, April 14, 2014, the entire HPC cluster—including all head nodes, filesystems, almaak machines, and compute nodes—will be unavailable due to our spring 2014 maintenance. We anticipate releasing the entire cluster back to the user community by 9:00 a.m. on Thursday, April 17.

During this downtime, we will complete the CENTOS 6.5 upgrade to all head nodes and compute nodes; apply security patches and operating patches to all file servers; apply firmware patches to various storage devices; upgrade the resource manager TORQUE; and update the /staging filesystem.

The upgrade of TORQUE will clear all jobs from the job queue.
The TORQUE upgrade removes /usr/bin/mpiexec. You will need to source either openmpi or mpich2 to have mpiexec in your path. We recommend openmpi as it works across both clusters, Myrinet and Infiniband.
/staging will be re-created, and all data will be lost on this temporary filesystem. We recommend you migrate any data that you wish to preserve to your project directory as soon as possible.

If you have any questions or concerns about this downtime, please contact us at hpc@usc.edu.