Massed Compute

When your Ubuntu VM is running but services are failing, ports aren’t responding, or deployments break silently, you need visibility into what’s actually happening. This guide provides a systematic approach to audit systemd units, journal logs, and service states without restarting anything.

Use this when systemd units fail, services exit unexpectedly, or expected ports go silent after a deployment. The outcome is a structured audit that correlates failed units with journal errors and listening ports.

Step-by-Step Deployment

# Test SSH connection
ssh -i /path/to/your/private-key -p 22 ubuntu@YOUR_VM_IP 'echo "Connection OK"'

# Should return: Connection OK

# List all failed units
sudo systemctl --failed

# Get system status overview
sudo systemctl status

# Count failed units
sudo systemctl --failed --no-legend | wc -l

# Check logs for a specific failed unit
sudo journalctl -u SERVICE_NAME -n 25 --no-pager

# Get error-priority logs from the last hour
sudo journalctl -p err -S "1 hour ago" --no-pager

# Show critical and alert messages
sudo journalctl -p crit..alert --no-pager

# Show what depends on a service
sudo systemctl list-dependencies SERVICE_NAME

# Show what a service depends on
sudo systemctl list-dependencies SERVICE_NAME --reverse

# Check if dependencies are running
sudo systemctl is-active SERVICE_NAME

# Show all listening ports
sudo netstat -tlnp

# Alternative using ss
sudo ss -tlnp

# Check specific port
sudo netstat -tlnp | grep :80
sudo netstat -tlnp | grep :443

# Create audit summary
echo "=== System Audit $(date) ===" > /tmp/audit-report.txt
echo "Failed Units:" >> /tmp/audit-report.txt
sudo systemctl --failed --no-legend >> /tmp/audit-report.txt
echo "" >> /tmp/audit-report.txt
echo "Recent Errors:" >> /tmp/audit-report.txt
sudo journalctl -p err -S "1 hour ago" --no-pager -o short >> /tmp/audit-report.txt

# View the report
cat /tmp/audit-report.txt

Troubleshooting

SSH Connection Issues

If you can’t connect to the VM:

• Verify the VM is in “running” state via your cloud console
• Check security groups allow SSH (port 22) from your IP
• Remove old host keys: ssh-keygen -R YOUR_VM_IP
• Verify SSH key permissions: chmod 600 /path/to/private-key

Permission Denied Errors

If you get permission errors when running commands:

• Ensure you’re using sudo for system commands
• Verify your user is in the sudo group: groups $USER
• Check if the journal is readable: ls -la /var/log/journal

Too Many Failed Units

If you have many failed units:

• Focus on services critical to your application first
• Look for common error patterns across multiple units
• Check if a single dependency failure is causing cascading failures
• Use systemctl list-dependencies --all to understand relationships

Skip All of This: Deploy with an AI Agent

This guide exists as a tested, machine-readable recipe in the Massed Compute MCP. Instead of running commands manually, connect an AI agent to audit your VM automatically.

{
  "mcpServers": {
    "massed-compute": {
      "type": "http",
      "url": "https://vm.massedcompute.com/api/mcp",
      "headers": { "Authorization": "Bearer MC_TOKEN" }
    }
  }
}

Then say:

The agent matches your request against the recipe catalog, connects to your VM, runs the audit commands above, and returns a structured report with failed units, journal excerpts, and listener correlation. It stops on any SSH or permission issues rather than proceeding with incomplete data. Recipe tested June 10, 2026.

Quick Setup Guide

For teams that prefer a checklist approach:

1. Connect: SSH to your Ubuntu VM
2. Check: sudo systemctl --failed for failed units
3. Investigate: sudo journalctl -u SERVICE_NAME -n 25 for each failed service
4. Correlate: sudo netstat -tlnp to match listeners with services
5. Document: Save failed unit names and key error messages
6. Plan: Prioritize fixes based on service dependencies

FAQ