beautypg.com

Requirements, Parallel processing for load operations – HP Neoview Release 2.4 Software User Manual

Page 65

background image

Requirements

These are the requirements to use the Trickle Feed feature:

ActiveMQ (JMS provider), which is included with the Java client package

A separate (and dedicated) instance of the Transporter Java client

A JMS trickle feed load is similar to loads from flat files or named pipes, except for the following
differences:

Workflow

When the Java client parses the control file and determines that the source is JMS, a trickle
feed process known as a JMS Consumer is launched to consume data from the JMS queue.
The JMS Consumer waits up to

“timeout” (page 59)

for data to arrive. If no data arrives

during this time, the JMS Consumer shuts down.

JMS Queue Message Handling
— Each message received is considered a single data record.
— Messages are in JMS text format, and may be either delimited or fixed-width format.
— Messages are read from the JMS Queue asynchronously and removed automatically as

they are read.

— Transporter continues to load data in real time continuously, a record at a time, until

the JMS Consumer reads the

“endseq” (page 53)

sequence.

Message Characteristics
— Message domain is Point-To-Point – each message has a single consumer, and the

message is retained in the queue until it is consumed.

— Message format is JMS text message type.

Control File Specification

sourcename jms jmsname [options provider-options]

jmsname

The name of the queue.

provider-options

The options are:

“endseq” (page 53)

“tenacity” (page 59)

“timeout” (page 59)

“retries” (page 58)

Limitations

The Java Consumer process commits to the JMS Queue after the data (row) is committed
on the Neoview platform. However, there is a window where a row may be inserted twice.
This situation could occur if the row is committed on the Neoview platform, but the network
connection between the client and server is dropped before confirmation is received by the
client.

For information about Java, refer to these web sites:

http://java.sun.com/products/jms/index.jsp

http://java.sun.com/products/jms/tutorial/

Parallel Processing for Load Operations

If you want job entries in a load operation to process in parallel, you must set

“parallel” (page 57)

= true

. These considerations apply:

Parallel Processing for Load Operations

65