How-to guides

Detailed steps for key operations and common tasks when working with Charmed HPC.

Initialize your environment

Install dependencies and initialize the backing cloud for your cluster.

• Initialize cloud environment

Deploy

Deploy and configure the core components of your cluster.

• Deploy Slurm

• Deploy a shared filesystem

• Deploy an identity provider

Integrate with other tools

Connect your cluster to observability platforms and workload tools.

• Integrate with Apptainer

• Integrate with Canonical Observability Stack

• Integrate with InfluxDB

• Integrate with a mail server for job notifications

Manage your cluster

• Customize job notification sender name

• Migrate a single slurmctld unit to high availability

• Rotate the Slurm authentication key

• Rotate the JWT authentication key

• Apply custom configuration to specific compute nodes

• Modify the default state and reason of new compute nodes

• Modify the state of compute nodes

• Scale partitions

Run workloads

Submit jobs and run containerized workloads on your cluster.

• Use Apptainer

Clean up resources

Remove previously deployed components and free cloud resources when they are no longer needed.

• Clean up Slurm

• Clean up cloud resources