Get help

Use the issue tracker for every technical problem and technical question about chispa. It is the only support channel for this server.

Report a problem or ask a question¶

Open an issue in the issue tracker. Use it for anything technical, for example:

A job that fails or behaves unexpectedly
A container image that won't build or run
A question about SLURM, Pyxis, or Enroot
A problem with your account or SSH connection

Warning

Don't email or message an admin directly for technical support. The issue tracker keeps a record that helps other users with the same problem, and it lets admins track open issues in one place.

Note

You need a GitHub account to open an issue, and you get access to the issue tracker only after an admin approves your account request. See A GitHub account. If you have a problem while your request is still pending, contact the system administrator directly.

Tool documentation¶

These are the tools you use directly on chispa. Each entry links to the tool's own documentation and to the page on this site that covers how it's set up here.

SLURM: The workload manager that queues and runs your jobs. See the official documentation for the full command reference, and Submit and monitor SLURM jobs for how it's configured on chispa.
Pyxis: The SLURM plugin that runs your job inside a container. See the Pyxis repository for its README and source, and Run containers with Pyxis for how to use it here.
Enroot: The container runtime that builds and runs your images. See the Enroot repository for its README and source, and Build your own container image with Enroot for how to use it here.
Docker: The format most container images use, including the examples in this guide. See the Docker documentation to learn about images, and Docker Hub to browse public images.
NVIDIA System Management Interface (nvidia-smi): The command-line tool that reports GPU status and usage. See the nvidia-smi documentation, and Check GPU usage of a running job for how to run it on chispa.