This article is part of an Essential Guide, our editor-selected collection of our best articles, videos and other content on this topic. Explore more in this guide:
1. - Designing a high availability and disaster avoidance plan: Read more in this section
- The case for OS-level high availability
- Balancing your approach to high availability
- Does resiliency dictate hypervisor performance?
Explore other sections in this guide:
- 2. - Virtual recovery and backup planning strategies
- 3. - Using snapshots as part of your virtual recovery and backup plan
Does your high availability architecture protect the operating system? Even with high availability at the virtualization layer, OS-level high availability offers additional benefits.
High availability at the virtualization layer protects the availability of virtual machines. The HA stack monitors the VM's status, and acts accordingly to maintain availability. This approach, however, does not guarantee the continuous availability of all services the VM offers. For example, if the entire operating system fails, it won't reply to anything. From the virtualization layer, it seems the VM is still using its physical resources, meaning the hypervisors sees it as available and will take no action to bring the OS back online.
Additional high availability architecture best practices
Using HA servers and systems
Low-cost high availability solutions
High availability trends and technologies
Operating system high availability, however, monitors the VM regardless of operating system, which makes it easy to implement and relatively inexpensive. In some cases, no additional protection is necessary, especially if the VM does not provide time-sensitive services.
If your VM offers a critical application, however, it could benefit from OS-level high availability. This setup monitors the availability of critical network resources, and acts the moment it detects one resource has become unavailable. OS-level high availability, however, monitors the availability of the application, a service that virtualization-layer high availability does not normally offer. But because operating system HA is OS-specific, you need a high availability stack on top of all environments (Linux and Windows, for example) running vital applications.
So what should you do if you already have a high availability architecture and you'd like to guarantee the availability of services within a VM? Should you implement operating system high availability within virtualization-layer high availability? This may seem like too much protection, but in some cases it is necessary. Many of today's data center servers in cloud environments are VMs that exist outside an admin's management scope. You need to treat these servers differently than physical servers. If network resources are vital to your business, you should ensure they are always available and add operating system high availability.