The following is a logical diagram of an Intellicus report servers cluster.
Figure 1 – Logical diagram – Intellicus Report Server Cluster
App portal component is either Intellicus web portal application running on any web server or Intellicus web APIs & pages embedded in your host web application. App portal is to be configured to point to load balancer in place of pointing to report server. App portal determines the appropriate report server for the current request with help of load balancer.
Intellicus Load Balancer is an independently installed executable component, which distributes the requests coming from the App portal among the nodes in the report server cluster. It is configured to run as a daemon process/ service.
Load balancer behaviour can be configured using property settings either using web UI or by editing its property file.
You can assign weightage of each report server node based on the node’s capacity and purpose. Load balancer does weighted allocation of requests to the nodes.
Note: You can install more than one load balancer components in your cluster to support load balancer failover.
Report Server (node)
Report server node has all the software runnable installed on its own machine and the key folders pointed to a shared file system. It performs all reporting operations as a standalone report server except of all common configuration files – namely connection files, style templates, language bundles, property files and persisted storage – namely published reports folder(rpg), report data cache, etc.
All report server nodes connect to same repository database. Any report object saved/deployed on the cluster will be immediately available to all nodes of the cluster.
Report server node communicates its current load information and health to the load balancer component at a regular interval.
Shared File System
Intellicus Report Servers cluster shares a file system for common persisted storage and common configuration files. The shared file system can be made available using a NFS or a shared folder mounted on other nodes. We recommend sufficient IOPS (Input/Output Operations Per Second) be available to cater to all the report servers to read and write data to this Shared File System.
Note: Shared file System is the single point of failure in an Intellicus report server cluster. It is strongly recommended that you ensure highly available, fail-proof RAID systems supporting the Shared File System.
All nodes on a report server cluster connect to same Intellicus repository and share all report objects seamlessly.
Note: Shared repository database is the single point of failure.
It is strongly recommended that you ensure it is highly available by database replication systems.
All nodes on a report server cluster connect to same reporting databases. The database schema is configured to fetch only once and shared by all nodes.