A report server farm is a collection of report server clusters sharing a common configuration and management options.
Report server farm concept
Report servers can share common configuration or report data; this sharing is defined as a report server farm. You can manage report servers individually or centrally in a topology called a server cluster. The benefits of creating report server farms are:
Easy expansion of your current deployments.
Introducing an additional resource to the farm involves adding the new report server to the device list in a NAM Console and designating it as a primary node within a new cluster, additional node in an existing cluster, or failover node for any node within the farm. The settings for the new node are acquired from the primary node based on the synchronization options.
If the cluster is reaching its capacity limits or performance is degrading, additional servers can be added to the cluster without the need to manually adjust the load balancing.
Only new data will be balanced across the additional nodes. Existing data will remain in their respective servers.
Centralized management of configuration settings for large deployments.
The configuration settings are defined on the primary server, which then shares (broadcasts) these settings, based on the synchronization options, to all report servers (nodes) of the same type within the farm.
Clearly defined report server hierarchy and responsibilities.
Each report server added to the device list in the NAM Console can be defined as:
Standalone - Indicates a report server that is operating independently of any farm or cluster configuration and is likely using its own configuration.
Primary - Indicates that this report server is performing the role of primary report server within a cluster that is part of a farm. To view the farm and the cluster, select the report server in the device list and click the Farm configuration tab.
Node - Indicates that this report server is part of a cluster that is part of a farm. The report server in a node role acts as an equal component of a cluster. Its configuration is obtained from the primary node of the cluster and is based on synchronization options defined for the cluster. The data stored in its database is an equal portion of the workload assigned to all nodes within a given cluster. For example, a report server that is part of a cluster containing 16 nodes is capable of analyzing 1/16th of the observed traffic. Report servers in the Node role act as a load balancer within the cluster. Adding more nodes will divide equally the incoming analysis tasks among all nodes listed within the cluster.
Failover - Indicates that this report server is designated as a backup server and is ready to take over the analyzing and reporting responsibilities of the node to which it is assigned. To view the report server for a failover server, select the failover device and switch to the Farm configuration tab.
When to use a farm or a cluster
Creating a farm or a cluster should be considered in deployments where:
- NAM Probe-generated data is overwhelming the report server, lowering its processing performance
Your deployment operating with one report server and receiving data from multiple data sources can be impeded by the amount of data designated for processing. To reduce the amount of processing applied to a single report server, create a farm with a load balancing cluster, set the overloaded report server as a primary node of the new cluster, and attach additional report server to load balance the data generated by the data sources between all nodes within the cluster. For more information, see Report server load balancing.
While in the Report Data Servers configuration you to specify a number of associated report servers to be used as data sources on the DMI screen, we recommend that you use the NAM farm-cluster feature to combine data processing.
Using a Report Data Servers to specify another reports server as a data source can create situation where additional report servers (added for example for load balancing reasons) will not be aware of an additional data source for the initial report server. Such configurations will generate errors on reports relating to for example, ADS operations.
- Deployment requires an automatic failover mechanism
Any node in any cluster within a farm is capable of having a failover node which shadows all operations of its parent node. If a node fails, the corresponding failover node will automatically takeover all functions of the failed node. All historical data will be available. To add a failover to a report server, create a farm (or use an existing one), and add a cluster with your report server as a primary node. Next, using the node action menu, add a failover to your report server. For more information, see Report server failover overview.
The farm hierarchy is based on the following components:
Farm - A farm is a set of clusters with one primary cluster representing the entire farm. A farm is composed of at least one report server cluster that contains at least one node (report server).
Cluster - A cluster is a set of nodes with one primary node for each node type (NAM Server or ADS) in the cluster. Each cluster is managed by the primary node of that cluster. Based on synchronization options, clusters can work independently of each other or share a configuration.
Node - A node is a single report server that shares its configuration and monitoring responsibilities with its peers within the same cluster. A node receives all configuration updates and sends the analysis results to the primary node of that cluster.
In this example, a farm has three clusters, each of which in turn has two NAM Server nodes.