Recovering from Data Center Failure

Discusses options for data center failure recovery.

Recovering from Data Center Failure

Planning for a data center failure can be complex, as it involves a large Webtrends topology as depicted in the System Requirements for Analytics and Visitor Data Mart). Here are a few things to keep in mind before beginning:

  • If the disaster recovery planning is limited to a situation where a data center is only temporarily unavailable, disaster recovery planning could focus specifically on providing access to reporting data, and not analysis.
  • The Webtrends environment uses Web log data in the analysis process, but makes no provisions to back up or restore Web log data. You must ensure that the original log data remains available.
  • The complete storage requirements for a Webtrends installation can be quite large. As a result, copying the data between data centers for fail-over can be time-consuming. Before proceeding with this scenario, ensure that the bandwidth capacity between data centers is sufficient to meet the needs of the environment.
Option 1: Data Collection and Reporting Redundancy

If your data loss is temporary, you can use a secondary environment to provide reporting. If you use this solution, no new data analysis can occur until the primary data center is recovered. Users will be able to log in to the Webtrends environment located at the secondary data center and view historical reports, but current data will not be analyzed. If you use the Webtrends SmartSource Data Collector, we recommend enabling a redundant SDC installation at the secondary data center so you can continue collecting SDC data while the primary data center is unavailable.

Figure 1. Storage Redundancy. Systems required for data collection and reporting redundancy

As an alternative, to reduce costs in the secondary data center, you could install the UI server directly on the main application server. Report performance and capacity will decrease slightly as a result, but this trade-off is often acceptable for a temporary situation.

Option 2: Full Redundancy of the Webtrends Environment

To plan for recovery after a total failure, you can duplicate the complete Webtrends system at the secondary data center to provide full capacity and full fail-over in the event of a catastrophic failure of the primary data center. In this scenario, you deploy an exact copy of each system at the secondary data center. Because Webtrends does not provide an automated method of fail-over, you would have to accomplish this with the help of your local IT resources.

Figure 2. Full Redundancy. Completely redundant deployment with secondary data center

Was this topic helpful? Send feedback.