Messaging Solutions in a Linux Environment

Chapter 7. Implementing highly available brokers 257

Figure 7-50 Trace node properties

Properties of the Filter node

The Filter node routes a message according to the message content. You define

the route by coding a filter expression in ESQL. The filter expression for our

scenario checks the message field Resource, which can be mem, cpu, or disk.

Right-click the Filter node and select Open ESQL. A new editor opens for an

ESQL file. The ESQL functionality used in our scenario is shown in Example 7-8.

In the ESQL function, if the ResouceType field in the message contains the string

cpu, then the message will be routed to TRUE terminal. If the ResourceType in

the message contains the string mem, then the message will be routed to FALSE

terminal. Otherwise, the message will be routed to UNKNOWN terminal.

Example 7-8 ESQL for Filter node

CREATE FILTER MODULE ResourceUsageMsgflow_Filter

CREATE FUNCTION Main() RETURNS BOOLEAN

BEGIN

IF Body.ResourceType like 'cpu' THEN

RETURN TRUE;

END IF ;

IF Body.ResourceType like 'mem' THEN

RETURN FALSE;

END IF;

258 Messaging Solutions in a Linux Environment

RETURN UNKNOWN;

END;

END MODULE;

Properties of MQOutput nodes

The MQOutput node is used to write a message to an MQ queue. The message

flow uses two MQOutput nodes. One is used to write a message when the Filter

node fails. The other one is used to write a message when the Filter node returns

Unknown as the result of the filter. Set the properties of the two MQOutput nodes

as illustrated in Figure 7-51 and Figure 7-52 on page 259.

Figure 7-51 Fail MQOutput Node property

Note: The Filter node is clearly best-suited for a traditional IF-THEN-ELSE

programming construct. For the equivalent of a CASE programming construct,

use the RouteToLabel and Label nodes. In the ESQL for a Compute node, you

can populate an element called DestinationList. This list should contain the

appropriate label (or case). Label nodes are added to the message flow for

each possible value of the label. The RouteToLabel node typically follows the

Compute node and jumps to the appropriate Label node that corresponds with

the value in the element DestinationList.

Chapter 7. Implementing highly available brokers 259

Figure 7-52 Unknown MQOutput Node property

Properties of DataInsert nodes

A DataInsert node is used to insert records in a database table using parts of a

message that hits the node at its in terminal. The database should be defined as

an ODBC data source.

The DataInsert node is a specialized version of the Database node and can be

used for all kinds of database interaction. The Database node itself can be

considered a specialized version of the Compute node. In the end, everything

that you can do with a DataInsert node or with a Database node can be

performed as well using a Compute node. The only difference is that you need to

write ESQL in a Compute node, while you have graphical tools to help you

customize the Database and DataInsert nodes.

1. The first step is to set the database name in the properties of the DataInsert

node. This name is actually the name of the ODBC data source. Set the

properties of two DataInsert nodes as illustrated in Figure 7-53 on page 260

and Figure 7-54 on page 260.

260 Messaging Solutions in a Linux Environment

Figure 7-53 DatabaseInsert node properties

Figure 7-54 DatabaseInsert node properties

2. The next step is to define the mapping between the elements of the incoming

message and the columns of the database table. In this scenario, the

mapping is straightforward. We map the elements of the message to the

columns in the database. The mappings are shown in Table 7-5 on page 261.

Chapter 7. Implementing highly available brokers 261

Table 7-5 Mapping between message elements and table columns

To implement this mapping in the Message Brokers Toolkit, perform the

following steps:

a. Right-click the DataInsert node and select Open Mappings. The mapping

editor will open as illustrated in Figure 7-55.

Figure 7-55 Mapping editor

b. In the mapping editor, right-click in the top area labeled Source and select

Add Message Mapping Input. In the Add Message window, select the

ResourceUsage message definition and click Finish as illustrated in

Figure 7-56 on page 262.

Message element Column in table

Time TIME

ResourceType RESOURCE

Usage USAGE

Qmgr QMGR