
Hyperpod Slurm Debugger
A read-only diagnostic that walks you through Slurm scheduler and daemon problems on HyperPod clusters. It covers the usual suspects: nodes stuck in down or drain, jobs pending forever despite idle capacity, GPU counts wrong, slurmctld unresponsive, and the "Node unexpectedly rebooted" message that shows up after auto-repair. The procedure runs a bash script that checks slurmd status, disk usage, RPC layer health, and controller state, then maps findings to the official AWS troubleshooting guide without printing any remediation commands. You still have to run fixes yourself. Requires the SSM plugin and unbuffer, which is non-negotiable because start-session will silently fail otherwise. Good for teams managing Slurm on HyperPod who need structured diagnosis before escalating or running state-changing commands.
- GitHub stars
- 783
- last updated
- Jun 11, 2026
Install
$ npx -y skills add awslabs/agent-plugins --skill hyperpod-slurm-debugger --agent claude-codeInstalls into .claude/skills of the current project.
Files
HyperPod Slurm Debugger
Diagnostic-only. Identify and classify Slurm scheduler and node-daemon issues on HyperPod Slurm clusters. Do not run, recommend, or print any state-mutating command. For remediation, link to the official AWS or Slurm documentation.
When to invoke
Invoke when the user reports any of the symptoms in the decision table.
When NOT to invoke
- Cluster has
Orchestrator.Eks— invokehyperpod-node-debuggerorhyperpod-nccl. - Single-node hardware fault with healthy Slurm scheduler — invoke
hyperpod-node-debugger. - NCCL training-hang investigation — invoke
hyperpod-nccl. - Node unreachable via SSM — invoke
hyperpod-ssm.
Constraints
- Read-only. Do not run, recommend, or print state-mutating commands.
- For any remediation, link to AWS or Slurm docs. The user authorizes and executes.
- IaC-managed cluster (Terraform / CloudFormation / CDK): warn that direct mutation drifts the live state from the IaC plan.
Canonical recovery URLs: references/slurm-details.md → Authoritative recovery documentation.
Prerequisites
Sponsors
Details
- Repository
- github.com/awslabs/agent-plugins
- Categories
- Documentation, AI & Agent Building, Debugging
- License
- Apache-2.0
- First seen
- Jun 11, 2026
More from awslabs/agent-plugins
All 32- Model Selection
- Sdk Getting Started
- Aws Architecture Diagram358
- Aws Lambda260
- Deploy240
- Directory Management183
- Use Case Specification183
- Planning168
- Dataset Evaluation156
- Aws Serverless Deployment154
- Finetuning154
- Model Deployment154
- Dataset Transformation153
- Hyperpod Ssm150
- Hyperpod Version Checker149
- Hyperpod Issue Report147
- Model Evaluation140
- Amplify Workflow101
- Amazon Location Service
- Api Gateway
- Aws Lambda Durable Functions
- Aws Lambda Managed Instances
- Aws Transform
- Document Service