Posted: 10 May 2018 14:16 EDT Last activity: 24 Sep 2018 9:53 EDT
Pega 7.3.1 - Node Classification
I need to classify a node with NodeType "BIX" so bix related agents should run on just this one node. Where do we set this jvm argument? We are running on websphere 8.5. Do we set it as Generic JVM arguments in "Application servers > server1 > Process definition > Java Virtual Machine" or somewhere else?
After setting the node type, I think only other step is to associate the agent with node type. Is this correct or need some further configurations?
Yes to configure a Node to start with a specific Node Type, one has to start the node with, for example, -DNodeType=<BIX>
If a Node has to behave as a combination of multiple Node Types then comma separated values for Node Types can be passed, for example, -DNodeType=<BIX>,<Background>
To effectively use node classification, you must map agents and listeners to node types. You can use the Node Classification landing page, the Agent rule form, and the Listener rule form to manage the associations. On the Node Classification landing page, you can view the list of agents and listeners across the cluster with their associated node types.
If a node is not started with a node type, node classification does not apply on that node, and the node starts with the Untyped node type. Any default associations of agents to node types do not affect the behavior of agents, and the agents run based on the Data Agent Queue configuration for that node. The listeners do not have any default association to node types, but if you mapped them to a node type, they do not run if the node is started with the Untyped node type.
Thanks for the quick response!! We have 2 nodes so I started one with nodetype "WebUser" and other with "BackgroundProcessing". We can make run BIX, Search and all other agents and listeners run on "BackgroundProcessing" node..right? What is the significance of assigning multiple nodetypes to same physical node like -DNodeType=BIX,Search etc ??
Then you can go for multiple node type association.
If you want to dedicate a node to more than one purpose, you must start it with multiple node types. In this situation, the node runs a set of listeners and agents that are associated with the specified node types. To configure a node with multiple node types, set the JVM argument with node types separated by a comma. For example, to dedicate a node for search and Business Intelligence Exchange (BIX) jobs, start a node with the following JVM argument:
If the same agent is enabled on both node types, and there is a conflict in agent schedules, the agent schedule for the node type that is listed earlier takes precedence. In this example, the agent that is enabled on the search and BIX nodes runs according to the schedule that is specified for the Search node type.
I restarted the JVM with node type set but still seeing node classified as UnTyped in pega. I am attaching the jvm setting screenshot here. Please let me know if anything wrong with the setting. I have also added the pega screenshot after nodes are started with nodetype.
If you're using Tomcat as a Windows service, you can adjust the JVM options by use CATALINA_OPTS or JAVA_OPTS. You can set the jvm options in either one of these variables in a catalina.batfile according to your requirement. Read the comments in catalina.bat file about these variables. You will understand
Set the environment variable - project.home in windows environment using this link. Now, your env variable is set
Set the JAVA_OPTS in setenv.bat (you need to create this file)
set JAVA_OPTS="%JAVA_OPTS% -Dproject.home"
for safety purpose. This will prepend the existing JAVA_OPTS with the new value.