Resolving Common Issues
Dear users of HPC cluster
The answers to many common issues/questions are available on our website. Here is a list of helpful steps we encourage you to take:
Getting Started
Account Issues
How do I apply/renew an account?
Note: applying for a new account or renewing an existing account require connecting to the VPN
Login issues
Check that you are using the VPN or the HPC Gateway Server (see Accessing HPC and SSH Tunneling pages for more info)
Make sure you did not go over quota (both storage size, and inodes number) in home directory
Software Requests / Packages Installation
Please read the comprehensive page for supported software and the follow steps
For Singularity environments: use “:rw” while installing packages, “:ro” when otherwise running containers (walkthrough here)
Use the --nv flag to pass NVIDIA drivers to singularity containers (more details here)
Data Management, Disk Space Questions, sharing data, transferring data, etc
Read the Best Practices for Storage page for tips on managing files and quotas
General information can be found on the Data Management on the NYU HPC Clusters page
Issues with Jobs, e.g. Long Queue Times or Jobs Killed
As a reminder, don’t run jobs on login node! This may result in your job being killed. Please submit a Slurm job.
Look at the efficiency of your jobs