Important Quality of Service(QoS) Notice
Headnode QoS
Over the past several weeks we have seen an unusual increase in use We have seen from time to time loads of the headnode resulting in loads from jobs that should be directed to the compute nodes. Often this is just a user mistake, but results in lower quality of service to others trying to prepare work for the compute nodes. It is policy that all significant research codes and related work be submitted to run on the compute nodes. In order to help promote the best possible quality of service, effective 6-24-2013, any running jobs on the head node deemed load producing and excessive will be terminated without notice. This will help insure the best possible responsiveness of headnode resources for everyone. Offending users will be further notified and reminded to submit jobs to the compute nodes. Jobs are submitted with the bsub command slurm commands to the appropriate LSF queue slurm partition meeting your needs. If you any questions about this change, please send your comments to: cluster-support@tufts.edu.
...
What is the hostname of the cluster?
New Cluster: login.cluster.tufts.edu
Legacy cluster: cluster6.uit.tufts.edu
How do I login to the cluster?
...
If the resources of the Tufts cluster do not meet your scaling needs, consider the various supercomputing centers sites such as Xsede. To discuss this further please contact cluster-support@tufts.edu.