When you configure a dataset profile to run on an Insight Server cluster, all of the machines in the cluster share all of the dataset configuration files for that profile.
Therefore, the entries for the parameters in these files must be applicable to all Insight Servers in the cluster. For example, the locations of the log files to be read, the lookup files to be used by Insight Server, and the location of the data output by Insight Server must be the same on all machines in the cluster.
You perform all configuration tasks on the cluster’s master Insight Server, which is the Insight Server you use to edit your configuration files. All saved configuration file changes made on the master Insight Server are synchronized automatically to the files on the processing Insight Servers in the cluster.
To run a dataset profile on an Insight Server cluster, you must perform the following processes in the order listed:
It is not required that all Insight Servers in the cluster process event data. You can designate one Insight Server in the cluster as a File Server Unit that stores the source files (VSL and log files) and serves the files to all of the Data Processing Units (processing servers) in the cluster. This setup provides the benefit of a single event data repository and leverages the processing power of all the processing servers in the cluster. The processing servers divide the data files among them and guarantee that the same file is not processed more than once.
For more information about designating an Insight Server to run as a File Server Unit, see the Log Processing Configuration File chapter of the Dataset Configuration Guide.
If you decide to store source data files on each of the processing servers rather than on a single File Server Unit, you must divide the files equally among the processing servers. Do not store all of the dataset’s source files on each of the processing servers. If multiple copies of the same file are available to multiple processing servers, the data is read multiple times (once by each machine) and skews your data.
For help determining which Insight Servers should process log files, please contact Adobe Consulting.
In the profile.cfg file, specify the processing servers that process the data for the profile.
To access the profile.cfg file
You access the profile configuration file using the Profile Manager in Insight.
To modify the dataset configuration files
If you need to make changes to the dataset configuration files (Log Processing.cfg, Transformation.cfg, dataset include files, Log Processing Mode.cfg, and so forth), do so only on the master Insight Server.
For instructions to access the files, see the Dataset Configuration Guide.
An Insight Server cluster permits the centralized storage of .vsl log files (from Sensor) on a single Insight Server machine called a File Server Unit (FSU). For information about installing an FSU, see Installation Procedures for an Insight Server FSU. For information about configuring an FSU, see the Dataset Configuration Guide.