How do I create a virtual machine (VM) instance in Google Compute Engine?

Type Your Question

How do I create a virtual machine (VM) instance in Google Compute Engine?

Tuesday, 18 February 2025

GOOGLE

Google Compute Engine (GCE) is a core part of Google Cloud Platform (GCP), offering Infrastructure as a Service (IaaS) by providing virtual machines in the cloud. This guide will walk you through the process of creating a VM instance in GCE, covering various configuration options, best practices, and troubleshooting tips to ensure a successful deployment.

Prerequisites

Before you begin, ensure you have the following:

A Google Cloud Platform (GCP) account. If you don't have one, you can sign up for a free trial.
A GCP project created.
Billing enabled for your project.
The Compute Engine API enabled.
Familiarity with basic cloud computing concepts.

Step-by-Step Guide: Creating a VM Instance

Access the Google Cloud Console
Navigate to the Google Cloud Console at https://console.cloud.google.com/. Log in using your Google account.
Select Your Project
In the console, make sure the correct project is selected from the project dropdown menu at the top. This ensures the VM instance is created within the desired project and billing is correctly applied.
Navigate to Compute Engine
In the navigation menu (the hamburger icon in the top-left corner), navigate to Compute Engine > VM instances.

Example: Screenshot of the Google Cloud Console, navigating to Compute Engine -> VM instances
Create a New VM Instance
Click the CREATE INSTANCE button at the top of the VM instances page. This will take you to the instance creation form.
Configure Your VM Instance
This is the most crucial step. You'll need to specify several configurations. Let's break down each section:
Basic Instance Information
- Name: Enter a unique name for your VM instance. Choose a descriptive name for easy identification (e.g., web-server-01).
- Region and Zone: Select the region and zone where you want your VM instance to be located. Consider proximity to your users and other GCP resources when making this selection. Different zones within a region offer redundancy and different resource availability. For example, you might choose us-central1 (region) and us-central1-a (zone).
Machine Configuration
- Machine Family: Choose the machine family based on your workload requirements. Options include:
  - General-purpose: Balanced price and performance (recommended for most workloads). Examples: E2, N1, N2, N2D, N3
  - Compute-optimized: Optimized for compute-intensive tasks. Example: C2, C3
  - Memory-optimized: Ideal for in-memory databases and analytics. Example: M1, M2, M3
  - GPU-optimized: Designed for machine learning and other GPU-accelerated workloads. Example: A2, G2, L4G
  - Accelerator-optimized: Suited for extremely demanding HPC and ML workloads. Example: TPU v4, TPU v5e
- Series: This filters the Machine Types available within the chosen Machine Family.
- Machine type: Select the specific machine type within the chosen family that best fits your CPU and memory needs (e.g., e2-medium, n1-standard-1). Pay close attention to the vCPUs and Memory offered by each machine type. Remember to balance performance with cost.
- Confidential Computing (Optional): Enable this for enhanced security by encrypting data while in use. This feature uses AMD Secure Encrypted Virtualization (SEV).
Boot Disk
- Boot Disk Image/Snapshot: Choose the operating system for your VM instance. You can select from various pre-built public images (e.g., Debian, Ubuntu, CentOS, RHEL, Windows Server, Rocky Linux) or use a custom image/snapshot. Consider the OS version and its compatibility with your applications.
  
  Click Change to browse and select a different boot disk image. You can filter images by operating system, version, and other criteria.
- Boot Disk Size: Specify the size of your boot disk in GB. Make sure to allocate sufficient space for the operating system, applications, and data. Consider future growth and application requirements when determining the disk size.
  
  The actual size is based on the chosen image, so if the disk is the "Image default size", GCP will utilize the underlying Image size as set when the OS Image was initially uploaded. This will likely change when first booting the VM due to space utilized by configuration changes made at first boot (and this first boot will require enough space to store any configuration files and scripts required, thus possibly resulting in an error and failure to start should insufficient space be provisioned initially)
- Boot Disk Type: Select the type of storage for your boot disk:
  - Standard persistent disk (pd-standard): Cost-effective for basic workloads with moderate performance needs.
  - Balanced persistent disk (pd-balanced): Good balance of price and performance for most general-purpose workloads.
  - SSD persistent disk (pd-ssd): High performance for I/O intensive applications, such as databases.
  - Extreme persistent disk (pd-extreme): Extremely high performance for the most demanding workloads. Supports the highest IOPS (Input/Output Operations Per Second).
Networking
- Networking interface: Usually preconfigured with a 'default' VPC network but custom VPC networks can be created for isolated internal private traffic for applications that required to adhere to data residency rules, for example. Configure network tags.
- Firewall Rules: Configure firewall rules to allow specific traffic to reach your VM instance. You can allow HTTP traffic (port 80) and HTTPS traffic (port 443) if you are hosting a web server. You should also configure firewall rules to allow SSH traffic (port 22) from specific IP addresses or ranges to allow secure remote access to your VM instance. Generally it's *BAD PRACTICE* to allow port 22 open to 0.0.0.0/0 because the vm is likely to be attacked by botnets trying out random credentials on that port. Instead look at enabling Identity Aware Proxy or only restricting the allowed incoming port 22 traffic to defined office network IP ranges.
Security and API access
- Service account: Associate a service account with your VM instance to grant it access to other Google Cloud services. You can choose a pre-existing service account or create a new one with specific IAM roles. Only provide access to other resources if and only if there's a concrete justification and always implement least priviledge where a specific API can only be allowed to read (for example).
- Access scopes: Access scopes define the level of access that the VM instance has to Google Cloud APIs. Google recommends using service accounts with granular IAM roles instead of relying on access scopes.
Management
- Automation: Here you can specify a startup script that will be executed when the instance starts, and any shutdown scripts upon termination of the vm.
- Availability policy: Determine what happens with a host failure. This policy defines whether Google migrates the VM, terminates the vm if failure, and determines machine settings (host failure response).
- Scheduling: This allows one to enable 'preemptibility' and the implications of scheduling policy for resource allocation.
These configurations dictate the performance, cost, and security of your VM instance. Review each section carefully and adjust the settings based on your application requirements and budget.
Create the Instance
After configuring all the settings, click the Create button at the bottom of the form. Google Compute Engine will begin provisioning your VM instance.
Monitor the Creation Process
You'll be redirected to the VM instances page. The new VM instance will be listed with a status of "Provisioning." Wait until the status changes to "Running."
Connect to Your VM Instance
Once the instance is running, you can connect to it in several ways:
- Using the Google Cloud Console:
  On the VM instances page, click the SSH button in the row of your VM instance. This will open a browser-based SSH terminal.
- Using gcloud CLI:
  If you have the Google Cloud SDK (gcloud CLI) installed and configured, you can connect using the command:
```
gcloud compute ssh [INSTANCE_NAME] --zone=[ZONE]
```
  Replace [INSTANCE_NAME] with the name of your instance and [ZONE] with the zone it's located in.
- Using a Third-Party SSH Client:
  If you've configured your instance to accept external SSH connections (which should be secured via specific source IP ranges, *NOT 0.0.0.0/0* as mentioned earlier), you can connect using your favorite SSH client, such as PuTTY or OpenSSH.
  You will need to obtain the external IP address of the instance from the VM instances page and ensure that your firewall rules allow SSH traffic from your IP address.

Best Practices

Choose the right machine type: Select a machine type that matches your workload requirements. Over-provisioning can lead to unnecessary costs, while under-provisioning can impact performance. Regularly monitor your VM instance's resource utilization and adjust the machine type as needed.
Use custom images: Create custom images from your configured VM instances to easily deploy identical environments. This speeds up deployment and ensures consistency across your infrastructure.
Automate deployments: Use infrastructure-as-code tools like Terraform or Cloud Deployment Manager to automate the creation and management of your VM instances. This ensures consistency and reproducibility.
Monitor your instances: Set up monitoring and alerting to track the health and performance of your VM instances. Use Cloud Monitoring to monitor CPU utilization, memory usage, disk I/O, and network traffic.
Implement security best practices: Secure your VM instances by following security best practices, such as regularly patching the operating system, using strong passwords, and configuring firewall rules to allow only necessary traffic. Consider using a VPN or SSH bastion host for secure remote access.
Manage Costs Effectively: Regularly review your GCE costs. Use committed use discounts or sustained use discounts for workloads that run continuously. Explore preemptible VM instances for fault-tolerant workloads. Clean up unused resources to avoid unnecessary charges.

Troubleshooting Tips

Instance fails to start: Check the instance logs for errors. You can find the logs in the Google Cloud Console under Logging or using the gcloud logging read command. Possible causes include insufficient resources, invalid configuration, or issues with the boot disk image. Review the boot disk space allocation during creation or using snapshot to reconfigure disk images for VM boot issues.
Unable to connect to the instance: Verify that your firewall rules allow traffic on the necessary ports (e.g., port 22 for SSH). Check the instance's network configuration to ensure that it is connected to a network with internet access. Verify there is an attached Service Account in IAM Roles that grants needed permissions.
Performance issues: Monitor the instance's CPU utilization, memory usage, disk I/O, and network traffic. Identify the bottleneck and adjust the machine type or disk type as needed. Also ensure applications installed on the vm aren't competing with other applications that contend with cpu and memory.
Network Connectivity: Troubleshoot DNS resolutions. Configure DNS settings using private zone. Also configure network tags.

Conclusion

Creating VM instances in Google Compute Engine is a straightforward process with a variety of configuration options. By following this comprehensive guide and adhering to best practices, you can effectively deploy and manage your virtual machines on GCP. Remember to continually monitor your instances, optimize your configurations, and implement robust security measures to ensure the long-term success of your cloud-based applications.