There is a lot of interest, and still a lot to be learned, about how storage replication will work with failover clusters when Microsoft offers up its newest (and still unnamed) version of Microsoft Server. Microsoft has provided us with some basic information about what we can expect.
Until the release of Windows Server 2012, Microsoft required failover clusters to make use of shared storage. The shared storage requirement for cluster nodes has since been removed, but cluster nodes must still have access to a cluster volume. One way of meeting this requirement will be through the use of storage replication.
According to Microsoft, the storage replication feature will allow failover clusters to be stretched to metropolitan distances -- essentially from one side of a city to another. The reason for the limitation on the distance over which a cluster can be stretched has to do with the way that the replication process works.
Storage replication is based on the SMB 3.0 protocol and the storage replication process is designed to be synchronous. Synchronous replication allows a server volume to be replicated without any data loss occurring. In contrast, asynchronous replication occurs on a scheduled basis, so if something happens to the primary server, the replica might not be in perfect sync with the primary server, which would mean that an unplanned failover could potentially result in data loss.
Microsoft has indicated that it will be possible to manage storage replicas for individual nodes and clusters through the Microsoft Failover Cluster Manager. This is the same tool that is currently used for managing Microsoft failover clusters. As such, administrators will be able to continue to manage failover clusters in familiar ways; the tool is simply being extended to expose the new functionality. In addition, storage replication will also be able to be scripted through PowerShell.
Microsoft addresses the Active Directory issues with Windows Failover Clusters
Dig Deeper on Storage pooling
Related Q&A from Brien Posey
Disaster recovery may not be the most prominent use case for blockchain technology, but the idea of using it for data protection and recovery is ... Continue Reading
Appliance-based hyper-convergence integrates software with hardware, while software-only HCI enables you to use the hardware of your choice. Each ... Continue Reading
When failing over VMs to the cloud or a remote data center, slow data transfer speeds can spell disaster. Don't let failed replications and partial ... Continue Reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.