Infoworks Installation on Multiple Edge Nodes
Infoworks DataFoundry supports running concurrent jobs on multiple edge nodes.
To install Infoworks DataFoundry on multiple edge nodes, you must first install Infoworks DF on the main edge node. This operates as the master edge node. While running the installation script, configure services such as MongoDB, Governor, and Orchestrator, to run these services on the secondary edge nodes.
For more details on installation of the master edge node, see On-premise Installation.
Perform the following steps after installing the master edge node:
- Install Infoworks DataFoundry on all the secondary edge nodes. Each node can run upto a maximum of 20 concurrent jobs . For example, to run 100 concurrent jobs, a master node with 4 secondary edge nodes are required.
- Copy the
$IW_HOME/conffolder from the master edge node to all the secondary edge nodes. - Ensure to configure the
$IW_HOME/logs/jobfolder on all the edge nodes to a single mount location in the network file system. This enables you to view the job logs running in all the edge nodes in the master edge node UI. - Ensure that the folders corresponding to the keys
user_scripts_base_pathanduser_extensions_base_pathin the $IW_HOME/conf/conf.properties file on all nodes (master node and secondary nodes) that are mounted on the same network file system. This ensures that the post hook script and custom transformation jars are available to all the edge nodes. - Start all the services on the master edge node by running the following scripts:
$IW_HOME/bin/start.sh mongo
$IW_HOME/bin/start.sh all
- In the master edge node, go to Admin > System Configuration page and add a configuration with key as cluster_mode and value as true.
- In the master node, restart Hangman by running the following scripts:
$IW_HOME/bin/stop.sh hangman
$IW_HOME/bin/start.sh hangman
- Start the Hangman services on all the secondary edge nodes by running the following the script:
$IW_HOME/bin/start.sh hangman