A collection of scripts to diagnose common ParallelCluster issues. The diagnostics suite is meant to be executed within the cluster head node.
| Script | Description |
|---|---|
diagnose-slurm-accounting.py |
Diagnoses SLURM accounting setup |
Run deploy.sh from your local machine. It uploads the diagnostics folder to the head node and installs dependencies.
bash deploy.sh --cluster-name <cluster-name> --region <region> --ssh-key <path-to-key>At the end it prints the SSH command to log directly into the diagnostics folder on the head node.
Once logged into the head node:
./diagnose-slurm-accounting.py --help