Managing Running Instances: Start, Stop, and Reboot Operations

Introduction

Instance lifecycle management is essential for maintaining your cloud infrastructure. Whether you need to perform maintenance, apply updates, or recover from issues, understanding how to properly start, stop, and reboot your instances prevents downtime and data loss. This guide covers both the Horizon dashboard and command-line methods for managing your instance state.

Prerequisites

• An active InMotion Cloud account with at least one instance
• Horizon dashboard access or OpenStack CLI installed and configured
• Basic familiarity with your instance's operating system

Starting a Stopped Instance

Stopped instances use minimal resources but remain part of your environment. When you're ready to bring them back online, you have two methods.

Using Horizon Dashboard

1. Log in to your Horizon dashboard
2. Navigate to Project > Compute > Instances
3. Locate the instance with a "Stopped" status (indicated by a red square icon)
4. Click the instance row to select it
5. Click the Start Instance button in the Actions dropdown menu
6. The instance will transition to "Starting" and then "Running" status
7. Wait for the instance to fully boot (usually 30-60 seconds depending on your operating system)

Once running, the instance will have an assigned IP address and be accessible via SSH or RDP.

Using OpenStack CLI

List your instances to identify the one you want to start:

1openstack server list
Copy

Start the instance by name or ID:

1openstack server start my-instance-name
Copy

Check the status to confirm it's running:

1openstack server show my-instance-name
Copy

The output will show the instance state as "ACTIVE" once it's fully booted.

Shutting Down Instances (Soft Shutdown)

A soft shutdown is the graceful way to stop an instance. It sends a shutdown signal to the operating system, allowing it to close applications and processes cleanly before powering off. This prevents data loss and filesystem corruption.

Using Horizon Dashboard

1. Navigate to Project > Compute > Instances
2. Find the running instance you want to stop
3. Click the instance row to select it
4. Click the Shut Off Instance button in the Actions dropdown menu
5. Confirm the action when prompted
6. The instance will transition to "Shutting Down" and then "Stopped" status

This process typically takes 30 seconds to several minutes, depending on what services are running on the instance.

Using OpenStack CLI

Stop the instance gracefully:

1openstack server stop my-instance-name
Copy

Verify it has stopped:

1openstack server show my-instance-name
Copy

The status will show "SHUTOFF" when the instance is fully stopped.

Rebooting: Soft vs. Hard Reboot

Soft Reboot (Recommended)

A soft reboot sends a reboot signal to the operating system, which performs a graceful restart. Applications are given time to shut down properly, and the filesystem is cleanly unmounted.

When to use: For regular maintenance, applying system updates, or clearing temporary issues while keeping the instance running.

Using Horizon:

1. Navigate to Project > Compute > Instances
2. Select your instance
3. Click Soft Reboot Instance from the Actions menu
4. Confirm the action
5. The instance will show "Rebooting" status during the operation

Using OpenStack CLI:

1openstack server reboot --soft my-instance-name
Copy

A soft reboot typically completes in 2-5 minutes.

Hard Reboot (Use With Caution)

A hard reboot forces the hypervisor to restart the instance immediately, without waiting for the operating system to shut down gracefully. This is equivalent to pulling the power cable and plugging it back in.

When to use: Only when an instance is unresponsive and a soft reboot has failed, or when debugging kernel-level issues.

Using Horizon:

1. Navigate to Project > Compute > Instances
2. Select your instance
3. Click Hard Reboot Instance from the Actions menu
4. Confirm the action
5. The instance will immediately restart

Using OpenStack CLI:

1openstack server reboot --hard my-instance-name
Copy

Hard reboots complete quickly but risk data loss and filesystem corruption. Avoid using hard reboot as a routine operation.

Best Practices for Instance Operations

Before Stopping or Rebooting

• Notify users: If the instance hosts services that others depend on, announce the maintenance window in advance
• Back up data: Ensure critical data is backed up to persistent storage
• Drain connections: Allow active client connections to complete before initiating shutdown
• Check scheduled tasks: Review cron jobs and scheduled tasks that might conflict with your maintenance window

Choosing the Right Operation

| Operation | Use Case | Duration |
|-----------|----------|----------|
| Soft Reboot | System updates, service restarts, clearing temporary issues | 2-5 minutes |
| Hard Reboot | Unresponsive instance, kernel debugging | 30-60 seconds |
| Stop/Start | Extended maintenance, reducing costs, hardware troubleshooting | 5 minutes |

Monitoring After Operations

After starting, stopping, or rebooting an instance:

1. Check instance status in Horizon or CLI
2. Verify SSH/RDP connectivity
3. Confirm critical services are running: systemctl status service-name
4. Check system logs for errors: journalctl -n 50
5. Verify networking: ip addr show or ifconfig

Automating Recurring Operations

For regular maintenance windows, consider using OpenStack automation tools or infrastructure-as-code frameworks to schedule reboots during off-peak hours.

Troubleshooting Common Issues

Instance Stuck in "Shutting Down" State

Problem: Instance remains in "Shutting Down" status for more than 10 minutes.

Causes:

• Services on the instance are hung and not responding to shutdown signals
• Network connectivity issues preventing the shutdown command from reaching the instance
• Hypervisor-level issues

Solutions:

1. First, attempt a hard reboot through Horizon or CLI
2. If the hard reboot doesn't resolve the issue, contact InMotion Cloud support with your instance ID
3. Avoid repeatedly trying soft reboots as this may exacerbate the problem

Instance Fails to Boot After Stop/Start

Problem: Instance status shows "Running" but is not accessible via SSH or RDP.

Causes:

• Operating system failed to boot properly
• Network configuration issue
• Security group rules blocking access

Solutions:

1. Check security group rules allow inbound SSH (port 22) or RDP (port 3389)
2. Verify the instance has a valid IP address: openstack server show my-instance-name
3. If the instance has a console, try accessing it through the Horizon dashboard to see boot messages
4. Perform a hard reboot if the soft reboot didn't resolve the issue
5. If the instance still won't boot, contact support with the instance ID and recent error messages

Soft Reboot Takes Longer Than Expected

Problem: Soft reboot stuck in "Rebooting" status for more than 10 minutes.

Causes:

• Large processes or database services taking time to shut down cleanly
• Filesystem checks running during boot
• Custom startup scripts running slowly

Solutions:

1. Allow additional time—filesystem checks can take 5-15 minutes on large drives
2. Try stopping the instance first, then waiting 30 seconds before starting it
3. For future operations, consider scheduling reboots during planned maintenance windows
4. If hung processes are the issue, you may need to manually kill services before rebooting

Instance Becomes Unresponsive After Reboot

Problem: Instance appears to boot but services are not responding.

Causes:

• Services failed to start automatically
• Configuration file errors preventing service startup
• Insufficient disk space or memory

Solutions:

1. Connect to the instance and check service status: systemctl status
2. Review system logs: journalctl -xe
3. Check disk space: df -h
4. Check available memory: free -h
5. Manually restart failed services: systemctl start service-name
6. If configuration errors exist, restore from backup and try again

Related Resources

• OpenStack Instance Task State Documentation
• OpenStack Reboot Operations Guide
• OpenStack Compute Troubleshooting