- Introduction
- Technology Options for Disaster Recovery Solutions
- Quick Checklist for Deployments
- Campus Cluster Maximum Distances
- Campus Cluster Topologies and Components
- Campus Cluster Configurations
- Performance in a Campus Cluster Environment
- Management Aspects of Campus Clusters
- Glossary
- Related Resources
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
Quick Checklist for Deployments
Sun Cluster 3.0 software is a scalable, flexible solution that can be deployed with equal benefit to small local clusters and larger extended clusters. Before deploying a campus cluster solution, however, consider your enterprise's requirements, resources, and risks. The following checklist provides an overview of factors to consider when determining which level of solution is best for an enterprise.
General Questions
What do you want to protect against?
Certain kinds of failures are more probable than others. For example, if the data center is close to a river, flooding may be a likely risk. The potential risk might impose restrictions on the solution. For example, to protect against site outages due to a major earthquake, a remote site may need to be established 500 kilometers away.
If the planned second site is in another geography, other solutions may have to be applied. If the desired result is total protection against all disasters, a campus cluster may not be adequate.
How much revenue per hour or market credibility could your enterprise lose if your mission-critical services were not available?
Understanding the financial impacts of a disaster on business operations helps define an adequate solution. Consider at least three factors:
The cost of a disaster recovery solution
The cost and impact of a disaster, including recovery times, potential data loss, lost revenue, and loss of reputation and credibility in the market
The probability that disaster will occur
If the potential loss is less than the cost to protect against that loss, a disaster recovery solution does not make sense. Unfortunately, history has shown that enterprises without good disaster recovery solutions suffer far more setbacks than businesses with plans. Campus clusters generally provide for very cost-effective protection against the more probable disaster scenarios, such as the loss of a whole site due to fire or other natural disasters.
Do you understand that a campus cluster does not address all aspects of a disaster recovery solution?
A campus cluster provides the appropriate infrastructure against many types of disasters, however, it does not offer complete protection against all disasters. In addition, it must be accompanied by other mechanisms to help ensure business continuity, for example, to recover from a total loss of data. Finally, a product alone is not enough; a solution needs the right people and processes in place to make it complete.
Infrastructure Questions
Are you prepared to accept performance degradation because of prolonged distances?
Even traveling at the speed of light through a fiber takes time, and the time it takes for data to travel 10 kilometers is 1000 times longer than it takes to go 10 meters. Even the latency introduced by the long wires is only a small part of the overall latency. The additional components involved, such as transceivers, switches, and multiplexers, add to the latency.
Does your data center infrastructure provide two or more sites?
Quorum devices help the cluster decide which nodes may form a new cluster in case of any failure. Thus, the availability of the quorum device is key in a disaster situation. In the case of the very common two-site infrastructure, the quorum device has to be placed at one of the two sites, making the loss of that data center more catastrophic than the loss of the other. In a three-site infrastructure, the quorum device is at the third site, so that the loss of one site would not affect the majority of quorum votes. For more information, refer to "Campus Cluster Topologies and Components" on page 12.
Does your infrastructure provide for independent Fibre Channel and network lines to span the distance between the data centers?
To prevent interference from other components on the same network or storage connections, independent lines or multiplexers should be available. Refer to "Campus Cluster Configurations" on page 18 for more information.
Does your infrastructure provide for a single IP subnet across the two (or three) sites?
To maintain accessibility to high availability services, clusters failover IP addresses. To configure the same IP addresses on network interface cards (NICs) and networks in different sites, it is necessary to have a single IP subnet across the two or three sites. Refer to "Campus Cluster Configurations" on page 18 for more information.
People and Processes
Is your staff well-trained and willing to undergo continuing training and exercising?
Having well-trained and experienced administrative staff is one of the key factors for achieving high availability. Periodically testing, at least annually, the disaster prevention and recovery processes is essential for training personnel to minimize service interruptions and restore normal operations.
Are there already processes in place that deal with recovery procedures for disasters?
Defining and establishing new processes to be used during disasters is a major task. If similar procedures are already in place, adapting them to the new infrastructure and new products is less effort. Administrative staff who are familiar with the processes understand and incorporate changes faster.