Skip to main content
Uncategorized

High Availability Series: Mirroring

By February 12, 2013February 25th, 2014No Comments
server mirroring

server mirroringMirroring is a technology that is catching on and is an evolution of clustering. With a mirrored solution, two separate servers run the same workload on completely separate hardware, and can fail over when necessary to each other. Nothing is shared between the two servers except for workload.  This helps to eliminate operating system issues and hardware failures at the server or storage level; however if someone deletes data, the deletion is mirrored to the other side as well.

This is a more expensive yet highly effective strategy for avoiding downtime. Since both servers run in lockstep with the exact same workload, one can take over immediately if the other fails. However, clustering servers is a less expensive, functional option.

When it comes to keeping the workload running in the instance of software failure, clustering and mirroring are the better options.  A combination of the two methods can create a functional and efficient environment for maintaining high availability. Keeping the virtual hardware up doesn’t do any good if the operating system goes down, and having to dedicate purpose built servers isn’t efficient.  In recent years we’ve seen environments get more and more available by mixing and matching technologies. Popular options are running workloads on virtual machines for consolidation reasons while running clustering/mirroring on top of that virtual infrastructure, or a hybrid where half the cluster/mirror is inside the virtual infrastructure and half is outside of it.

If you’re interested in adopting a strategy to increase your server high availability, contact us today!