Why you want a Witness to Disaster Recovery

How to protect data and be prepared to survive a disaster is a big topic to cover. Obviously data protection has a lot of components—ranging from hardware capabilities, algorithms used by the system to protect data and ensure integrity, and concerns like backup and replication strategies. In this note let’s mercifully contain this discussion to what is arguably one of the most important topics – surviving a disaster with minimal disruption to your business.

Disaster Recovery, DR, is a necessary consideration when designing robust enterprise solutions. By definition, DR always implies physical distance between the systems involved in the design. Isilon supports data replication over distance and the components of an Isilon DR solution include:

  1. Application Servers
  2. LAN connecting the application servers to the Isilon storage
  3. Isilon cluster at the production data center
  4. Isilon cluster at the DR data center
  5. A WAN connecting the two data centers
  6. Isilon SyncIQ for data replication

Ok, now the data is automatically replicated—but what happens in the event of a DR fail-over?  It requires manual intervention or scripts to make the data on the DR site available to users and applications. What is missing is a way to orchestrate and automate the fail-over process so that the business magically recovers from the loss of the primary site.

One way to do this is with a Disaster Recovery Witness Appliance. Like the name implies, the witness appliance is watching over the Isilon clusters and will “witness” any event that may take place

The witness concept includes monitoring data replication as well as detecting changes in the configuration data outside of the cluster. Monitoring allows the witness to coordinate share, quota, and replication policies between clusters to ensure synchronization between sites.  These clusters may be in DR implementations that are configured as Active and Standby data centers or as Active and Active data centers. The witness should be separated from the clusters and come in the form of separate hardware and software running external to the Isilon storage clusters so that data access can be tested the same way users and workstations would.

Where to find a Witness?

Fortunately, such a witness exists for Isilon. It was developed by Superna to solve these exact challenges and is named Eyeglass. Superna Eyeglass software is available from EMC Select.

The diagram below shows how DR orchestration and automation can be achieved with EMC Isilon.  Superna Eyeglass performs the “Cluster Witness” function for an end to end solution that enables application servers, networks, DNS, and Isilon storage to act as a single distributed system.Disastery Recovery

The addition of the Superna Eyeglass software completely orchestrates and automates the fail-over of data on Isilon storage clusters. The software is simple to install and use, and provides reporting and alerts to any potential issues. With Superna Eyeglass and Isilon SyncIQ you can now orchestrate and automate your Disaster Recovery operations providing a seamless fail-over in the event of an outage.

For more information, check out the solution overview on Superna Eyeglass for EMC Isilon SyncIQ.

About the Author: Bob Williamsen