Chapter: Preface Up Main page Chapter 2: System Requirements 

1 What Is Clustering?

Clustering is a way for several servers to act as one. In a typical InterWorx cluster, one machine acts as the Cluster Manager (Master), and the other machines are set as Nodes (Slaves). The nodes mount the /home partition of the CM, so when someone comes to a site on the CM it will be load balanced across all of the boxes. All data continues to be stored on the CM server.

1.1 What Does Clustering Do? What Doesn’t It Do?

Clustering provides an increase in reliability and a decrease in latency for your users. A cluster will allow you to service more requests, do so more flexibly, and share load across many machines instead of focusing all of your customers’ needs onto one machine. If any node in the cluster goes down, customers won’t notice anything except for a slight drop in response time, depending on your traffic volume.
Clustering isn’t a “magic bullet” that keeps your files up to date on a large number of machines. It doesn’t do data replication or distributed filing. Clustering is not a replacement for a backup system.

1.2 Why Cluster?

The best reason to use an InterWorx cluster is load balancing. Storage is centralized, but service isn’t - While you only have one “live” storage pool to draw from, adding a node is a very simple and straightforward process, able to be done with an extremely light piece of hardware.

1.3 What About Redundancy or High-Availability Clustering?

Unfortunately, as yet we don’t have a deployable, real-world “ready for primetime” High-Availability Clustering solution. Our software engineering team is working hard on the feature, but as yet there is no ETA. There are still solutions for storage reliability improvement, despite the lack of High Availability.
 Chapter: Preface Up Main page Chapter 2: System Requirements 

(C) 2019 by InterWorx LLC