Storage 101: High Availability, Part 3


Global Clustering
Once you have a replica copy of the data at a remote site, you need an automated method of starting the application at the remote location, pointing to the correct data volumes, and redirecting user traffic to the secondary site. Global Clustering products such as VERITAS Global Cluster Manager extend the concept of local clustering to the wide area by providing automated application and service migration from a primary data center to a geographically dispersed location. This allows your critical application to continue to function even in the event of a site outage.

Global clustering extends the local clustering environment to a wide area configuration by managing cluster application and cluster state information and broadcasting status updates, event handling and replication commands over standard TCP/IP connections to remote sites. The Global Cluster Manager interface is completely browser based, providing administrators with sufficient access privileges and a location-independent command post from which to monitor, manage and control their global enterprise clustered environments.

Principle Purposes of Global Clustering:

  • Disaster Recovery--Administrators with clusters in physically disparate areas can use global clustering to set policies for migrating applications from one location to another in the event that clusters in one geographic area become unavailable because of an unforeseen disaster. This keeps the application available in the event of a complete site outage.
  • Automating Availability--The Global Cluster Manager "one-click" migrate button automatically administers pre-defined policies for application availability. In the event of a site outage, simply click the "migrate" button to failover applications and services to the remote site; this eliminates the possibility of human error.
  • Monitoring and Management--Monitor health of all sites. Each site knows the state information for each site worldwide. Clusters of various operating systems can coexist within the same site (a virtual collection of clusters), viewed logically from a single explorer tree and controlled by common command syntax.
  • Administrative Migration--Planned downtime for upgrades and maintenance can be virtually avoided by migrating site resources to an alternate site. Global enterprises can implement "follow-the-sun" policies and ensure that their support and data infrastructures are operating in proximity to their active business requests and queries.

    Normal Conditions
    Under normal conditions, the global clustering management tool monitors applications across all sites. While data is replicated to the remote site, the global clustering management tool tracks and displays replication status and reports any lag or issues with replication activity.

    Local Application Fault and Group Failover
    The global clustering management tool monitors the local application to detect faults then continually reports and tracks any local failover within the cluster. If necessary, modification of the cluster events and triggers can be controlled directly from the global clustering interface. During this time data replication continues and client access to the application is uninterrupted.

    Complete Site Fault and Takeover
    Complete site takeover (whether planned or unexpected) is efficiently handled by the global clustering configured policies. With a single mouse click, pre-configured events and policies are administrated. Domain Name Server updates are triggered, providing seamless redirecting of client access to the remote/disaster recovery site. Data consistency has been maintained prior to the site failure via ongoing replication allowing client access with virtually no perception of moved resources. The remote site is promoted as the primary site and replication roles are reversed in order to prepare the original site for eventual failback when site issues are resolved.

    The end result is that users can continue to access critical applications, even in the event of a complete site outage.

    Summary
    In today's competitive environment many companies must have availability requirements that only an integrated replication and clustering solution can provide. Together, these technologies provide an integrated, highly available and manageable solution that ensures business without interruption.

    Part 1
    Part 2