DataFlux Data Management Studio 2.5: User Guide

Usage Notes for JMS Nodes

• Design Considerations When Using JMS Reader Nodes in Data Jobs

Design Considerations When Using JMS Reader Nodes in Data Jobs

JMS Reader nodes for data jobs can potentially run indefinitely, never returning an EOF marker to the child node. Accordingly, specify appropriate time-out values and EOF markers, or design the flow of your data job such that a JMS Reader node will return an EOF when a condition is met.

Another consideration has to do with the combination of other data job nodes with the JMS Reader nodes. The data job nodes listed below require an EOF from the parent before they will produce any output rows. Accordingly, if a data job has one or more JMS Reader nodes and needs to use one or more of the other nodes below, special consideration needs to be given to the configuration of the JMS Reader nodes or the nature of the information to be read.

• Clustering
• Cluster Update
• Cluster Diff (when "land all data" option is enabled)
• Sort
• Join
• Union
• Monitoring
• Basic Statistics
• Frequency Distribution
• Text output nodes might buffer data before writing resulting in some data being delayed ) 
• Data output nodes might not behave correctly if single transaction mode is set