What is the difference between active/passive and active/active clustering?

1 Star2 Stars3 Stars4 Stars5 Stars
  • Product:
  • Version:
  • Document Type:
  • Revised:
  • Reviewed:


What is the difference between active/passive and active/active clustering?


The terms active/passive and active/active have different meanings in different contexts. In the strictest sense, active/active clustering, means that the applications or services running on the cluster can access the same resources at the same time. In common usage, it means that multiple nodes in the cluster can provide the same type of service.

In the the specific context of ExtremeZ-IP, active/active clustering means that you can have multiple instances of the ExtremeZ-IP running as a service within a single node or spread across multiple nodes . Using multiple instances of ExtremeZ-IP allows for better load balancing and increased flexibility. Although 2 services cannot share a disk, each service can share out some portion of the total number of disks.

Active/active clustering also allows for simpler server consolidation. When an additional standalone server is migrated into the cluster, it is a simple as creating a new cluster group, with an ExtremeZ-IP service, and assigning it the old standalone server’s name, IP address, and files. The Macintosh end users won’t even know that the files are now hosted on a cluster, other than that they see less down time.

In most cases the advantages of an active/active cluster make it the preferred solution;however, there are some situations where an active/passive solution is easier to implement. One example is if there is another service running on the cluster that must use the same resources as ExtremeZ-IP and the other service can only run in active/passive mode. Another example is if the cluster is a replacement for a single standalone server with a limited number of volumes.

When ExtremeZ-IP is running in active/passive mode a single service named “ExtremeZ-IP” will listen for connections on all IP addresses of the node it is running on, and it will store it’s Registry settings in the same location as a non-clustered version of ExtremeZ-IP.

When ExtremeZ-IP is running in active/active mode the first instance of the service (default) can be named “ExtremeZ-IP” as in the active/passive installation. Additional instances must be created using unique service names, IP addresses, and settings for each service.

Starting with ExtremeZ-IP 4.1 the different versions and serial numbers for active/passive and active/active installations of ExtremeZ-IP were eliminated. By default when you install ExtremeZ-IP 4.1 it will create a single service called “ExtremeZ-IP” allowing it to act as an Active/Passive configuration, however at any time you can add additional virtual servers and it will become an active/active solution.