Compute and Login Nodes
Interacting with Slurm compute and login nodes
Compute nodes
SUNK compute nodes are Kubernetes Pods running slurmd. They are mapped to physical Nodes in the cluster. Jobs run within Slurm cgroups on these Nodes, which are typically large, feature high memory, fast networking, and have either GPUs or significant CPU resources. Compute nodes are the "workhorses" of the Slurm cluster. They define and manage the specific hardware used to run Slurm jobs, and handle the resources used to run jobs submitted to Slurm.
In SUNK, Slurm nodes run in Kubernetes Pods. Slurm nodes are not the same as Kubernetes Nodes, the worker machines that run Kubernetes Pods. To maintain a distinction between the two in this documentation, Kubernetes Nodes are capitalized while Slurm nodes are not.
Login nodes
SUNK login nodes provide the ability to access your Slurm cluster. They are Kubernetes Pods with IP addresses exposed in order to enable connections so that users can prepare data, submit jobs, and check job statuses. They are not intended for heavy computation, so they do not typically have a GPU.
With SUNK, you can create flexible Compute node definitions to meet the specific resource requirements of your workloads.