Service Level Expectations
The mission of ARCH Research Computing is to support researchers across Johns Hopkins by providing reliable, performant, and well-supported high-performance computing resources. Our team is committed to partnering with you in your academic, scientific, and computational work by maintaining infrastructure, responding to issues, and enabling productivity through clear expectations and open communication.
To foster a strong working relationship, we’ve outlined the following service expectations:
Support and Communication Channels
For help with the Rockfish Cluster: help@rockfish.jhu.edu
For general questions, ARCH-wide issues, or DSAI support: help@arch.jhu.edu
Hours of Operation
Business Hours: Monday-Friday, 8:00AM – 4:30PM (Eastern)
Response Time: Submitted Tickets are often responded to quickly, but may take up to 1 business day. This includes acknowledgement of your request, a status update, and an estimated time for resolution or next contact. Requests submitted towards the end of the day or on holidays will be addressed the next business day.
Emergency Issues: May be handled outside business hours when possible, particularly for major outages or widespread service disruptions.
Outages & Cluster Status
Some outages – whether scheduled or unexpected – are unfortunately unavoidable. We aim to notify users by email as early as possible. For real-time updates on service status for the Rockfish cluster, please refer to https://status.arch.jhu.edu. Please note that this link only monitors outward network connectivity, not necessarily the health of the node as it relates to responsiveness, or filesystems.
Expectations of Users:
To ensure prompt and effective support:
- Use Slurm for all job submissions. Do not run compute-intensive tasks on login nodes. This ensures system responsiveness for all users.
- Provide as much detail as possible when opening a support ticket, including error messages, job IDs, relevant environment information, or screenshots.
- Respond promptly to follow-up questions from support staff to avoid delays.
- Submit requests for major changes (e.g., software installations, environment support, or workflow troubleshooting) with appropriate lead time, especially before deadlines or teaching schedules.
- Understand that ticket priority is influenced by severity, impact, and time sensitivity. While all requests are important, some may take precedence on system-wide demand.
We also ask that users are transparent and proactive when requesting help related to a time-sensitive deadline or resource constraint. If something is urgent, please communicate the context clearly. While we do our best to assist, we may not always have the time or staff available to resolve last-minute requests without advance notice.
Out of Scope Support
While we strive to help users be productive, some types of requests fall outside the scope of what our team can directly support. These may include, but are not limited to:
- Client-side software troubleshooting (e.g., local IDEs, VPN software, SSH clients)
- Network configuration on personal machines or remote networks
- Unsupported or unapproved software packages not available on the cluster
- Custom code debugging, unless it’s to confirm cluster-related issues or job submissions (e.g., environment variables, module conflicts)
In these cases, we will do our best to guide you with general advice or refer you to broader IT support, but our ability to resolve these issues may be limited.
Commitment to Support
We take ownership of your request from the moment we acknowledge it. The staff member who responds will guide the issue through to resolution – whether it involves internal troubleshooting, scheduling changes, or coordination with broader IT teams.
When our team members are away (e.g., vacation, illness), you’ll be directed to another point of contact to avoid delays.
Our goal is to be transparent, responsive, and collaborative – helping you make the most of the computing resources available to you.