MB toolkit

Top

• Integration Toolkit Version 9.0.0.1 Interim Fix 001
• toolkit needs to connect to remote brokers, so it installs MQ java client libraries

The Singleton solution is a step half-way between the Aggregation pattern and WPS. Singleton is lighter-weight than WPS but more intelligent than Aggregation.

En principio no hay ejecucin en paralelo dentro del flujo. Sacar dos conexiones de un nodo no significa paralelismo. La ejecucin es secuencial pero adems no tienes garanta sobre cual va a ser la conexin que se va a ejecutar en primer lugar (salvo que uses el nodo flow order).

Para lo de la ejecucin en paralelo lo nico que se me ocurre es que B1 en lugar de ser un subflujo, lo hagas un flujo independiente (por ejemplo usando MQ) y tendras que B1 no tendra ninguna dependencia transacional y se ejecutara en el hilo de ejecucin que atiende a dicho flujo, distinto del flujo ppal

How to have a node out of the UOW :

Top

1. AggregateControl node
2. AggregateReply node
3. AggregateRequest node
4. Check node
5. Collector node
6. Compute node
7. Database node
8. DatabaseRetrieve node
9. DatabaseRoute node
10. DataDelete node
11. DataInsert node
12. DataUpdate node
13. EmailOutput node
14. EndpointLookup node
15. Extract node
16. FileInput node - retrieves data from a local or remote file system
17. FileOutput node - outputs data to a local or remote file system
18. Filter node
19. FlowOrder node
20. HttpInput node
21. HttpReply node
22. HttpRequest node
23. Input node
24. Java Compute node
25. JMSInput node
26. JMSMQTransform node
27. JMSOutput node
28. JMSReply node
29. Label node
30. Mapping node
31. MQeInput node
32. MQeOutput node
33. MQGet node
34. MQInput node
35. MQJMSTransform node
36. MQOptimizedFlow node
37. MQOutput node
38. MQReply node
39. Output node
40. Passthrough node
41. PeopleSoftInput node
42. PeopleSoftRequest node
43. Publication node
44. Real-timeInput node
45. Real-timeOptimizedFlow node
46. RegistryLookup node
47. ResetContentDescriptor node
48. Route node
49. RouteToLabel node
50. SAPInput node
51. SAPRequest node
52. SCADAInput node
53. SCADAOutput node
54. SiebelInput node
55. SiebelRequest node
56. SOAPAsyncRequest node
57. SOAPAsyncResponse node
58. SOAPEnvelope node
59. SOAPExtract node
60. SOAPInput node
61. SOAPReply node
62. SOAPRequest node
63. Throw node
64. TimeoutControl node
65. TimeoutNotification node
66. Trace node
67. TryCatch node
68. TwineballInput node
69. TwineballRequest node
70. Validate node
71. Warehouse node
72. XSLTransform node

New Web services nodes :

• Provider nodes (SOAPInput, SOAPReply)
• Consumer nodes (SOAPRequest, SOAPAsyncRequest, SOAPAsyncResponse)
• Helper nodes (SOAPExtract, SOAPEnvelope)

WebSphere Service Registry and Repository :

• EndpointLookup
• RegistryLookup

Routing :

• DatabaseRetrieve -- to ensure that information in the message is up-to-date.
• DatabaseRoute and Route -- to direct messages down different paths of a message flow based on certain criteria.
• Collector -- to create a message collection from one or more sources based on configurable criteria.

WebSphere Adapter nodes :

• PeopleSoftInput
• PeopleSoftRequest
• SAPInput
• SAPRequest
• SiebelInput
• SiebelRequest

File nodes :

• FileInput
• FileOutput

Transport nodes

• JMSReply
• EmailOutput

New and enhanced nodes in MB V6.1

Top

Use the Check node to compare the structure of a message arriving on its input terminal with a message structure definition that you supply when you configure the Check node.

messagebroker Message Flows, page 302 [312 of 556]

Use the Filter node to route a message according to message content.

messagebroker Message Flows, page 326 [336 of 556]

A Compute node takes as input a message tree (called InputRoot) and outputs a different message tree (called OutputRoot). Notice that InputRoot is read-only and therefore cannot be modified in place.

HTTPin & HTTPout nodes allow the CGI world to talk to "c" programs.

SOAPin & SOAPreply nodes allow to talk to the WebServices world - mind NameSpace(s) and Encoding

Trace node : where to send its output ( diferent at Development and Production ). When the BAR file is generated, select the "Configure" bottom flap and enter the "File Path" property. If empty, does it mean "no trace" ?

Other nodes :

• SendMailPlugIn - support pack IA07.
  A xml tag with the error message is added to the original message.
• z/OS File Adapter - support pack IA11.

WS MB Reference + Built-in nodes

local tkt help

Top

The type of input node that is used in a message flow depends on the type of application that sent the message to the broker:

• If the message flow gets the messages from the business application by means of a WebSphere MQ queue, use an MQInput node.

  El único queue manager que el MQInput node puede usar es el que usa el Broker. Por eso no se puede especificar otro. Gracias, Emir !

• If the messages are sent to broker by an MQe application, use an MQeInput node.
• If the messages are sent by a telemetry device, use a SCADAInput node.
• If the messages are sent by a Web services client, use an HTTPInput node.
• If the messages are sent by a JMS or multicast application, use a Real-timeInput node or a Real-timeOptimizedFlow node.
• If the message flow is to be reused as part of one or more other message flows, use an Input node.
•  
• MQInput
• JMSInput
• HTTPInput
• SOAPInput
• SAPInput
• Timer node

WBI-MB Basics, SG24-7090, page 125 [139 of 262]

If the incoming message has an MQRFH2 header, you do not have to set values for the Default properties because the values can be derived from the <mcd> folder in the MQRFH2 header.
If you set values, and those values differ from those in the MQRFH2 header, the MQRFH2 header values take precedence.

messageBroker_Message_Flows, page 359 [369 of 556]

MQInput Node Properties.

• Message Domain
• Message Set
• Message Type
• Message Format
• Topic

Each input node in a message flow causes the broker to start a separate thread of execution. Including more than one input node might improve the message flow performance.

Message Flows, page 68 [76/1020]

In Message Domain, select the name of the parser that you are using from the drop-down list.

• MRM
• XML
• XMLNS
• XMLNSC
• JMSMap
• JMSStream
• MIME
• BLOB (same as blank)
• IDOC
• DFDL {mb v8}

Body is equivalent to Root followed by the parser name (for example, Root.XML), which you can use if you prefer.

If there is a "domain" versus "received message format" mismatch, you can see in trace:

Que es ? Que proporciona al fluxe ?

A Message Set corresponds to CICS Copy Book or (Siebel) Web-Service WSDL

In older versions (WMB V7 and below), the primary way general text and binary messages were modeled and parsed was through a container called a message set and associated 'MRM' parser. Since WMB V8, and including IBM Integration Bus, such messages are modeled and parsed using a new open technology called DFDL from the Open Grid Forum. This is IBM's strategic technology for modeling and parsing general text and binary data. The MRM parser and message sets remain a fully supported part of the product; in order to use message sets, a developer must enable them as they are disabled by default to encourage the adoption of the DFDL technology.

wiki

IA96: WebSphere BI Message Broker - Message Set Tutorial (Article). This SupportPac shows how to develop and exercise message sets in WebSphere Business Integration Message Broker version 5.

Quan en un fluxe, al que hem entrat mitjançant un node MQInput, hi ha un problema (as THROW), es torna al node d'entrada, es recupera el missatge inicial, i es surt pel pin Catch, mantenint l'entorn, on s'ha enmagatzemat la raó del error. Es posa un "Trace" node i després

• MQOutput amb càlcul del destí via Oracle + DestinationList
• MQOutput a una cua fixe de "problemes".
• Publish del problema, i que l'agafi qui vulgui

Top

The output node that is used in a message flow depends on the type of application that receives the message from the broker:

• If the message flow puts the message to the application by means of a WebSphere MQ queue, use an MQOutput node.
• If you want the message flow to put different messages on different WebSphere MQ queues, depending on the queue name in the Reply-to queue field of the message header WebSphere MQ, use an MQReply node. Using the MQReply node means that the name of the output queue is not hardcoded in the message flow.
• If the messages are received by a telemetry device, use a SCADAOutput node.
• If the messages are received by a Web services client, use an HTTPReply node.
• If the message flow interacts with a Web services client, use an HTTPRequest node.
• If the messages are received by a JMS or multicast application, use a Real-timeOptimizedFlow node. This node performs as both an input and an output node.
• If the message flow is to be reused as part of one or more other message flows, use an Output node.
• If you want to distribute the messages, using Publish/Subscribe, to many unknown applications that have subscribed to receive the data, use a Publication node.

WBI-MB Basics, SG24-7090, page 126 [140 of 262]

Petit dubte :
Les propietats del nus de sortida tenen dos camps : queue manager name i queue name. Que tinc que definir al qmgr local si vull que el missatge se'n vagi a un qmgr remot (a una cua especifica) ?
O a l'inrevés : si vull que vagi a un qmgr remot, puc especificar aqui una local "remota" ?

Solució :

1. podem no-posar queue manager i posar com a cua una QUEUE REMOTE
2. podem posar queue manager (remot) i nom de cua (remota), pero llavors haurem de definir al queue manager local una cua de transmissió que es digui com el queue manager remot i que estigui assignada a un canal que porti cap al queue manager remot on estigui la esmentada cua. [Emir, 26-07-2006]

Select the Destination Mode from the drop-down list. This identifies the queues to which the output message is put.

• Queue Name. The message is sent to the queue named in the Queue Name property. The properties Queue Manager Name and Queue Name (on the Basic tab) are mandatory if you select this option. This is the default.
• Reply To Queue. The message is sent to the queue named in the ReplyToQ field in the MQMD.
• Destination List. The message is sent to the list of queues named in the LocalEnvironment (also known as DestinationList) associated with the message.

Si al nodo MQoutput (Properties, Advanced) le ponemos "Destination mode := Destination List", entonces podemos hacer este cálculo :

Compte : el node "Compute" ha de tenir a les seves propietats Compute mode = LocalEnvironment ! Otherwise veurem aquest missatge a la llista de Events:

Top

• To verify that the structure of a message is correct for subsequent processing, use the Check node.
• To perform a Boolean filter on a message so that if it matches the rule the message takes one route through the message flow, and if it does not match the rule the message takes another route through the message flow, use the Filter node.
• To force messages to be processed in one way and then processed in a second way, use the FlowOrder node.
• To define a number of alternative destinations to which messages can be sent, depending on the message header, use a Label node for each destination, and a RouteToLabel node to determine which Label node a message must be sent to.
• To change the message properties before the message is parsed again by nodes downstream in the message flow, use the ResetContentDescriptor node.

WBI-MB Basics, SG24-7090, page 129 [143 of 262]

A key component of the broker architecture will be the basis on which routing decisions are made.

B2B Integration using MQ and MQSI, SG24-6010, page 210 [234 of 372]

Para comparar en hexadecimal se usa:

Pero hay que decir que escrito así el flow no funciona en EBCDIC. El código anterior (CAST con CCSID) asegura la portabilidad del código.

1. in the Compute node, code :
   CREATE COMPUTE MODULE DecideOnQuery CREATE FUNCTION Main() RETURNS BOOLEAN BEGIN SET OutputRoot = InputRoot; IF InputRoot.XML.PassengerQuery.ReservationNumber <> '' THEN SET OutputLocalEnvironment.Destination.RouterList.DestinationData[1].labelname = 'SinglePassenger'; ELSE SET OutputLocalEnvironment.Destination.RouterList.DestinationData[1].labelname = 'AllReservations'; END IF; RETURN TRUE; END; END MODULE;
2. Compute.Out -> RouteToLabel node In
3. RouteToLabel is EMPTY
4. Label node 'Label Name := SinglePassenger ;'
5. Label node 'Label Name := AllReservations ;'

URL

When you use RouteToLabel and Label, you must include a Compute node that determines, using some combination of message content, database content, and ESQL logic, how messages are to be processed next. Configure the Compute node to create a destination list (within the DestinationList folder in the LocalEnvironment subtree) that contains the destination for each message, specified as the LabelName of a Label node. The Compute node passes the message to the RouteToLabel node, which reads the destination list and propagates the message to those destinations.

"Message Flows", page 71 [79/1020]

Top

Before you design and implement user-defined extensions, you should familiarize yourself with the core components and ensure you understand the basic WebSphere Business Integration Message Broker runtime architecture.
You should ensure that you are familiar with the following runtime components and concepts:

• Runtime environment
• Broker domain
• Configuration Manager
• Broker
• Execution group
• Execution model on page 57

You should also make sure you understand the following concepts:

• message flows
• the logical message tree
• how the message tree is populated

messagebroker_User_Defined_Extensions.pdf, page 56 [64 of 212]

url

Top

(short) Introduction to DFDL

El parser construye un "syntax tree". Ejemplo simple con XML - si el mensaje que llega al MQ Input node es:

El parser le permite al ESQL usar estas referencias:

Y si el message flow ejecuta esto:

El mensaje de salida sera:

El parser hace dos cosas:

• construye el "syntax tree" (entrada)
• reconstruye el mensaje a partir del "syntax tree" (salida)

developerWorks :

• message modeling : Creating a DFDL CSV message model, also Creating a DFDL Text (TDS) message model
• Creating and testing message models for C and COBOL using DFDL

DFDL is not a data format - it is a language for describing data formats

DFDL is an open standard, hosted by the Open Grid Forum, OGF

Introduction to DFDL

Quan escullo "DFDL" com a "Message Domain", em demana un "Message" - "select a message from a DFDL schema file in a library"

Solució : "New" + "Message Model" - Message model sirve tanto para xsd como para dfdl.

MRM - Message Repository Manager (Supported only in Message Broker)
DFDL - Data Format Description Language (Introduced in MB 8.0 as a new Parser and Domain)

• Standard:
  MRM- Proprietary standard
  DFDL- Open standard (like WSDL)
• Usage:
  MRM - Used only in message broker
  DFDL - Can be used any product as it is Open standard and Message Broker is the 1st software product that supports DFDL in WebSphere family.
• Testing message model:
  MRM - Can't test the modeling without deploying message set and flow that uses it
  DFDL - Can be tested without deploying the messaging model (WMB8 toolkit has DFDL tester)
• Performance:
  Performance of DFDL parser is much better than MRM
• Migration:
  MRM - Can't migrate or share message modeling to any other software products.
  DFDL - Can migrate or share message modeling to any other software products.

url

This tree is always present in the input message: an empty LocalEnvironment tree is created when a message is received by the input node.

Variables
This subtree is optional. If you create local environment variables, you are recommended to store them in a subtree called Variables. It provides a work area that you can use to pass information between nodes. This subtree is never inspected or modified by any supplied node.
Variables in the local environment can be changed by any subsequent message processing node, and persist until the message flow goes out of scope and the node that created it has completed its work and returns control to the previous node.
The variables in this subtree are only persistent within a single instance of a message flow. If you have multiple instances of a message passing through the message flow, and need to pass information between them, you must use an external database.

Destination

Written destination

Para poder modificar el Local Environment, hay que copiarlo (de Input -read only - a Output).

When the message flow processing is complete, the LocalEnvironment tree is discarded.

The Environment tree is a part of the logical message tree in which you can store information while the message passes through the message flow.

This tree is always present in the input message; an empty Environment tree is created when a message is received and parsed by the input node. You can use this tree as you choose, and create both its content and structure.

The Environment tree differs from the LocalEnvironment tree in that a single instance of it is maintained throughout the message flow.
No hace falta copiarlo. Pero se recibe un Environment vacio con cada mensaje.

LocalEnvironment => MQ and MB.

Environment => user.

Both are message-related !

A user-defined property (UDP) is a variable that is defined in your ESQL program by specifying the EXTERNAL keyword on a DECLARE statement. You can define external variables only at the module and schema level.

Before you deploy the message flow that uses the UDP, you can change the value of the UDP by using the Broker Archive editor. A deployment failure occurs if you try to deploy a message flow that contains a UDP that has had no value assigned to it.

At run time, after the UDP has been declared its value can be queried by subsequent program statements but it cannot be modified.

Using UDPs, configuration data can be set easily and used just like normal constants.

Java : To access broker user-defined properties from a JavaCompute node, use the getUserDefinedAttribute(name) method, where name is the name of the property that you are accessing.

"Message Flows", page 50 [58/1020].

Shared variables can be used to implement an in-memory cache in the message flow (see Optimizing message flow response times). Shared variables have a long lifetime and are visible to multiple messages passing through a flow (see Long-lived variables). They exist for the lifetime of the execution group process, the lifetime of the flow or node, or the lifetime of the node's SQL that declares the variable, whichever is the shortest. They are initialized when the first message passes through the flow or node after each broker startup.

A logical message tree is a structure that is created by one or more parsers when an input message bit stream is received by a message flow.

url

The message tree is initially populated by the input node of the message flow.

url or messagebroker_Message_Flows.pdf, page 251 [261 of 556]

A correlation name is a field reference that identifies a well-defined starting point in the logical message tree and is used in field references to describe a standard part of the tree format.

When you access data in any of the four trees (message, environment, local environment, or exception list), the correlation names that you can use depend on the node for which you create ESQL or mappings, and whether the node creates an output message. For example, a Trace node does not alter the content of the message as it passes through the node, but a Compute node can construct a new output message.

You can introduce new correlation names with SELECT expressions, quantified predicates, and FOR statements. You can create non-correlation names in a node by using reference variables.

url

You cannot manipulate the contents of a BLOB message, because it has no predefined structure. However, you can refer to its contents using its known position within the bit stream, and process the message with a minimum of knowledge about its contents.

The BLOB message body parser does not create a tree structure in the same way that other message body parsers do. It has a root element BLOB, that has a child element, also called BLOB, that contains the data.

You can refer to message content using substrings if you know the location of a particular piece of information within the BLOB data. For example, the following expression identifies the tenth byte of the message body:

url

You need to refer to the BLOB data as 'Root.BLOB.BLOB' or you could use 'Body.BLOB'.

url

Emir : Si pones 'XML' en el Domain, el mensaje deja de ser BLOB. El Filter node deberia usar Root.XMLNSC.xxx

Mezcla : Si usas BLOB, y despues del filtro necesitas ver el mensaje como XML, usa un "Reset Contents Descriptor" node (despues del filter node), que pone XML en el Domain y tiene un TICK en "Reset Domain".

Top

A Reset Content Descriptor (RCD) node has been inserted after the Catch node. This prevents an exception being thrown by the Trace node in the case where a message stored by a TimeoutControl node cannot be parsed by the TimeoutNotification node and wouldn't be able to be parsed by the Trace node either. The RCD node resets the domain to BLOB.

Schutz

Top

Select Destination as File.
Set File Path to "e:\toolkit_trace\mb_trc.txt"
Set Pattern to

My selection is :

A more complete one :

Body sample :

url

Testing and debugging message flow applications url or Message Flows, page 489 [499/556]

It Trace Node has "User Trace" selected, then you have to process the usertrace using mqsireadlog and mqsiformatlog commands to get the UserTrace information.

Without having two flows, obviously.

Possible values for Fields :

• Destination : "None", "User Trace", "File", "Local Error log".
• File Path : "c:\tmp\mb\traza.txt"
• Pattern :
• Message Catalog :
• Message Number : 3051.

Those values end up in file cyclon.cmf.

Three ways :

1. <null> File Path for the Trace Node properties.
2. promote property
   Right-click on the Trace node and select "Promote Property". Drag-and-Drop "Destination". It will appear it the flow properties from now on. Right-click on any part of the white space and open Properties to modify it.

   url

3. Use a filter before the trace which checks the broker name. If the name is the same as your production broker, dont invoke the tracenode. If it isn't, go trough the node and then the same path.

Have no trace nodes in the flow and use mqsichangetrace at runtime to trace the flow. I've never had to use trace nodes in a flow to determine what it is doing. I simply have a batch file (startTrace) to enable a trace and another (stopTrace) to stop the tracing. Then I read the trace, everything you want to know is there.

Start :

Stop:

You can fine tune the script to trace just a flow as opposed to every flow in the EG but you get the idea.

This is the com.ibm.etools.mft.util plug-in.

If you select Errors to file, the default file to which trace is written is install_dir\eclipse\workspace\cmptraceerror.log

The default is None.
If you select Stored in a file, the default file to which trace is written is install_dir/eclipse/workspace/.metadata/.plugins/com.ibm.etools.mft.admin/mbda.log

Troubleshooting, pages 116-127 [124-135/188]

Nice command (from "db2cmd") : DB2 LIST SYSTEM ODBC DATA SOURCES

Muestra la lista de bases de datos registradas con ODBC.

Top

ESQL Reference v 6.0 v 6.1

Writing ESQL

The message body is always the last child of root, and is identified by its parser name, for example XML or MRM.

If you are referring to the input message to interrogate its contents in a Database or Filter node, use correlation name Body to refer to the start of the message. Body is equivalent to Root followed by the parser name, for example, Root.XML, which you can use if you prefer.

Broker properties {from InfoCenter} that are accessible from ESQL and Java {mqJeff} :
The only broker-defined properties that can be used in a Trace node are those in the "General broker properties" group. See trace node

Top

El Compute Node tiene varias salidas :

• Out (de salida)
• Out1
• Out2
• Out3 - the third alternate output terminal to which the transformed message might be routed by a PROPAGATE statement.
• Out4

Compute Node : Message Flows, page 492, [500/1020].

For example, if you want to create three copies of the input message received by the Compute node, and send one to the standard "Out" terminal of the Compute node, one to the first alternate "out1" terminal of the Compute node, and one to the Label node "ThirdCopy", code the following ESQL:

In the above example, the content of OutputRoot is reset before each PROPAGATE, because by default the node clears the output message buffer and reclaims the memory when the PROPAGATE statement completes. An alternative method is to instruct the node not to clear the output message on the first two PROPAGATE statements, so that the message is available for routing to the next destination. The code to do this is:

url

Also, to generate three copies of the input message, and not perform any further action, include this code immediately before the RETURN TRUE statement:

Alternatively, you can modify the default behavior of the node by changing RETURN TRUE to RETURN FALSE:

url

Si queremos que el Mensaje ( OutputRoot ) se propague al nodo siguiente, debemos seleccionar modo "Message".

Si queremos que el LocalEnvironment se propague al nodo siguiente, debemos seleccionar modo "LocalEnvironment".

Valores posibles en "Compute Node properties -> Basic -> Compute Mode":

• Message
• LocalEnvironment
• LocalEnvironment and Message
• Exception
• Exception and Message
• Exception and LocalEnvironment
• All

Per fer que un missatge no expiri mai, hi posarem :

Per fer que el fitxer de sortida de un node FileOutput porti un TimeStamp com la part final del nom del fitxer, posarem aquest ESQL:

Top

The terminals of the MQInput node are:

• Out - the output terminal to which the message is routed if it is successfully retrieved from the WebSphere MQ queue.
• Failure - the output terminal to which the message is routed if an error occurs. Even if the Validation property is set, messages propagated to this terminal are not validated.
• Catch - the output terminal to which the message is routed if an exception is thrown downstream and caught by this node.

Quina diferencia hi ha entre "an error occurs" i "an exception is thrown" ?

• Yes - means that the input node completes its own operation under sync point, and the default behavior in the message flow is for actions to be taken under sync point.
• No - means that the input node completes its own operation out of sync point, and the default behavior in the message flow is for actions to be taken out of sync point.
• Automatic - means that the node gets the message under sync point if the message is persistent, and out of sync point if it is non-persistent.

Top

The terminals of the MQOutput node are:

• Out

  The output terminal to which the message is routed if it has been successfully put to the output queue.

• Failure

  The output terminal to which the message is routed if a failure is detected when the message is put to the output queue.

Top

The default port numbers for the embedded execution group listener are 7800 for HTTP and 7843 for HTTPS.
You can change these port numbers, and port ranges used by the execution group listeners, by using the mqsichangeproperties command.
To display actual values, use mqsireportproperties command :

The process biphttplistener.exe shall be listening there (use netstat plus tasklist)

The HTTPRequest node determines the URL for the web service to which it sends a request. Select one of the following three options; the node checks these in the order shown (that is, the first always overrides the second, the second overrides the third):

• X-Original-HTTP-URL in the HTTPRequest header in the input message
• LocalEnvironment.Destination.HTTP.RequestURL in the input message
• The Web service URL property

Top

To trigger a web service, all we need is (in a Compute node preceeding a SOAP Request):

SOAP Requets node has a Operation Mode checkbox.

• invoke a specific web service defined by a WSDL interface
• invoke a generic web service

Gateway mode allows you to configure your flow to handle generic SOAP request/response and one-way messages, or to act as a facade between multiple web services clients and multiple back-end web services providers. The node can onvoke any binding operation.

Gateway opetarion mode for SOAP nodes :

Here soap12 is the prefix used in the outbound message.

Use this ESQL to create the namespace prefix definition for ns1:

developerWorks

Changing the default URL for a SOAPRequest node or a SOAPAsyncRequest node request

You can also override the Operation property :

My first webservice requires

DECLARE statement

Here is the answer, due to magnificent Nacho:

• (a) classic way, using WSDL
• (b) generic - requires 3 parameters
  • namespace
  • operation
  • url

Top

Required classes, as mongo-java-driver-2.4.jar, have to go into c:\MQSI\9.0.0.0\classes !

Top

Use the FileInput node to process messages that are read from files. One or more messages can be read from a single file, and each message is propagated as a separate flow transaction. The part of a file that generates one message flow transaction is called a record. A file can be a single record, or a series of records. Properties on the node specify how the FileInput node determines the records in a file. When the FileInput node propagates a message, it stores information about it in the LocalEnvironment.File message tree. If the input file is empty, an empty message is propagated (assuming that it is valid).

Its Basic properties are:

• Input directory
• File name or pattern
  Your input file name will be in InputLocalEnvironment.File.Name. You could store in environment to use it for your output files creation.
• Action on successful processing
  • Move to Archive subdirectory (mqsiarchive)
  • Add Timestamp and Move to Archive subdirectory (mqsiarchive)
  • Delete
• Replace duplicate archive files: Y/N

On the Input Message Parsing tab, set values for the properties that the node uses to determine how to parse the incoming message.

Use the Records and Elements tab to specify how each file is interpreted as records, as "Record detection := Whole File ; " - specifies that the whole file is a single record

En el nodo FILEINPUT de la versión 6.1 del broker, una de las propiedades del nodo es la ruta del fichero de donde queremos leer los ficheros.
Es posible cargar este valor dinámicamente de las variables de sistema operativo de alguna variable de entorno ? O cargarlo de alguna variable de un fichero properties o similar?
Es importante porque las rutas en los distintos entornos (desarrollo, preproducción y producción) son distintas.

Lo más fácil es cambiar el atributo en el fichero BAR.
También se puede usar LocalEnvironment:

Para más detalles, Help -> Search: ac55410

If we code:

• FileInput
• Trace
• MQOutput

We get this trace:

Top

File Name : the name of the output directory and the names of the output files are determined by the node properties that you specify and by elements of the message that is being processed.

The file name to be used is determined as follows:

• If the file name contains no wildcard, the value of this property is the name of the file created.
• If the file name contains a single wildcard, the value of the element LocalEnvironment.Wildcard.WildcardMatch in the current message replaces the wildcard character, and the resulting value is the name of the file created.

The output directory path to be used can be overridden by values in the current message. The name of the file can be overridden by values in the current message. See the information relating to the Request tab for details about how to do this, using $LocalEnvironment/Destination/File/Directory and $LocalEnvironment/Destination/File/Name

EOF : in tab "Records and Elements", set "Record definition := Record is Whole File ;"

If the output is to a remote filesystem, accessed using SFTP, we have to specify a "security identity" in the node properties (here "sftp_id"), that has been defined using the mqsisetdbparms command. The user identifier and password that are to be used to log on to the FTP or SFTP server are obtained from this definition, the name of which must have the prefix ftp:: :

Use ssh keys instead. First generate them for the broker id (using ssh-keygen) then supply the broker id's public key (.ssh/id_rsa.pub) to the sftp server (to manually add it to their .ssh/authorized_keys file as a one time task). Then set up the setdbparms to reference the brokers private key file.
Code:

Top

Two built-in nodes:

• TimeoutControl
• TimeoutNotification

Note naming convention of adding UID (TIME) to node name. Since the purpose of the TimeoutControl node is to dynamically set the characteristics of a TimeoutNotification node, the UID of the TimeoutNotification node should be the same as the UID of the controlling TimeoutControl node. Multiple TimeoutControl nodes can control the same TimeoutNotification node by specifying the same UID.

The purpose of this node is to act as an alarm clock, initiating a message flow when the timer goes off. The TimeoutNotification node would be the first node in a timer-controlled message flow. It has no In terminal.

As a standalone node, TimeoutNotification nodes support simple, fixed interval timer events. For example, you can have a message flow executed every 30 seconds, every hour, or even every 17 years.

As always, wire the Catch terminal of the TimeoutNotification node so that you avoid any unpleasant retry conditions.

The tree coming out of the node contains only a Properties and LocalEnvironment subtree.

The purpose of a TimeoutControl node is to let you set of the parameters of a TimeoutNotification node.

Including one or more TimeoutControl nodes in either the same or a different message flow (or even different execution groups) makes the patterns of generated timer events much more flexible.

The TimeoutNotification node's Operation Mode properties must be changed from Automatic to Controlled for the TimeoutControl node to have any effect.

developerWorks Schutz. See "Programming patterns"

The TimeoutNotification node in "Automatic" mode does not produce a message tree suitable for sending to an MQOuput node (for example, there is no MQMD subtree). The following ESQL code can be used to create an appropriate message (and would be coded in the "BuildTimeoutMessage" node above).

Top

IA9Z: WebSphere Message Broker - Job Execution Node : download, PDF

How about using a JCN with Runtime.exec() instead of the Job Execution Node?

Top

Debug using ExceptionList !

tkt help

Top

Allows message flows to cal IMS transactions and handle responses. Typical scenarios include Web Service -> IMS, File -> IMS, SAP -> IMS. Complements MQ IMS Bridge and IMS Web Services currently available.

info center

Top

Create required objects using CMD files in "\\MQ\CMDS\T400\MB7"

Lets start with "New(Message Flow Project)" := 'ICS', with no referenced projects. On the "ICS flow -> Flows", create a "New(Message Flow)" := 'first flow' ;

Place 3 nodes on the canvas :

• a MQ input (ICS_IN) - select "Message Domain := XMLNSC" under "Input Message Parsing" ;
• a Trace node - destination := "file", path := "c:\temp\mb\mytrace.txt", pattern := ${BODY}
• and a MQ output (ICS_OUT), without destination queue manager

Join Trace-Out pin with MQ_Output-In and Trace-In pin to MQ_Input-Out

Identify the version by clicking on the canvas of the "msgflow" and filling the "version" property at bottom. You can see this value in the message flow deployed in the execution group, under the "Properties" tab.

Save it using "Ctrl+S".

To create the BAR file, right-click on the ICS project, and select "New(Message Broker Archive)" := 'mybar.bar' ; On the right, select the "first_flow.msgflow" resource and click "Build broker archive" at bottom.

To deploy the BAR file, right-click on the BAR file, and select "Deploy", and select Execution Group "EG1".

If there is an error in the deployment, fix it, and then right click on BAR file, and select "Build and Save BAR" to re-create the BAR file - icon disapears. Deploy again.

If the deployment ends ok (and the Input Queue exists), there must be a thread having opened the input queue for "input" - use MQ Explorer. And placing a message in the intput queue shall trigger the writing into the trace file ; also, a message is written into the output queue.

By example, this message goes fine thru the flow, and leaves this trace:

Top

We use these nodes :

• MQInput - Input Message Parsing, Message Domain := BLOB ;
• Compute - Compute Mode := All ;
• Trace - ${LocalEnvironment}
• FileOutput - Directory := "c:\temp\mb\fileout" ; File Name or Pattern := "fo*" ;

In Compute Node we code:

The result is a MQ message written into a file with timestamp in its name:

And this trace:

Top

The MQ_Input node had defined Message Domain := "XML" ;

El missatge de entrada era :

El filtre i la transformació foren :

Els resultats obtinguts son :

Input: mensaje XML. Por ejemplo:

Compute node:

Output (en caracter):

Output (en hexadecimal):

Notarás que se convirtió la eñe de codepage 437 (ASCII) a 1252 (Unicode).

Tengo un flujo con un nodo MQ-IN, un nodo "COMPUTE" en medio y un nodo MQ-OUT a la salida.

Con un mensaje de entrada como

Con un "compute node" con este codigo, obtenemos estos mensajes de salida

Top

Shared variables (defined with the SHARED keyword) can be used to implement an in-memory cache in the message flow; see Optimizing message flow response times. Shared variables have a long lifetime and are visible to multiple messages that pass through the flow; see Long-lived variables. They exist for the lifetime of the integration server process, the lifetime of the flow or node, or the lifetime of the node's SQL that declares the variable (whichever is the shortest). They are initialized when the first message passes through the flow or node after each broker startup.

See also the ATOMIC option of the BEGIN ... END statement. The BEGIN ATOMIC construct is useful when a number of changes have to be made to a shared variable and when it is important to prevent other instances seeing the intermediate states of the data.

{pend} I want to have a flow that receives messages and inserts a sequence number in each one of them, keeping the value as a global variable in cache. {sag, 20131030}

Top

"Global Cache allows a number of Brokers to be integrated for workload balancing. Global Cache provides the ability to have a cache storing information about the requester which would later be used to correlate the replies correctly. Using Global cache gives the flexibility for different brokers to handle the request and reply parts of an application. Using a global cache allows you to scale up the number of clients, while maintaining predictable response time for each client."

1. start DB2 CLP
2. create savings database : "db2 create database SAVG_ACC"
3. create the table and procedure in the savings database: "db2 -vf setupSavingsAccount.sql"
4. create loan database: "db2 create database LOAN_ACC"
5. create the table and procedure in the loan database: "db2 -vf setupLoanAccount.sql"
6. create queue manager: "crtmqm QM"
7. start queue manager: "strmqm QM"
8. create queue manager objects: "runmqsc QM < queues.defs"
9. start MBCL
10. create broker: "mqsicreatebroker BROKER -i myuser -a mypwd -q QM"
11. start the broker: "mqsistart BROKER"
12. start toolkit
13. import project interchange "DebitCredit.zip" file
14. create execution group "DebitCredit"
15. open DebitCreditFlow_deploy.mbtest, click "Send message", select execution group "DebitCredit" and click "Finish".
16. open DebitCreditFlow_test.mbtest, select Enqueue and then click "Send message"; select Dequeue and then click "Get message".

accessing DB2 from WebSphere Message Broker using stored procedures {\\MB\ToolKit\DB2}

Top

The THROW esql statement generates a user exception. It produces a rollback to the MQInput node, where the Catch pin will be used.

MESSAGE is an optional clause; if you omit it, it defaults to the first message number of the block of messages provided for using THROW statements in WebSphere Message Broker catalog (2951). If you enter a message number in the THROW statement, you can use message numbers 2951 to 2999.

ESQL, MB version 6, page 252

Si usas el pin "CATCH" deberías tener el Environment intacto. Es en "FAILURE" donde recibes un Environment "nuevo".
Es muy común poner cosas en el Environment antes de un THROW, para pasarle datos al CATCH. Por ejemplo:

Top

Permite código C (o Java) escrito por el usuario.

Se despliega en un fichero ".LIL" en el broker.

Top

A transaction describes a set of updates that are made by an application program, which must be managed together.

Configuring global coordination of transactions (two-phase commit)
Configuring transactionality for message flows

You can configure the nodes in your message flow to determine how the work taken by each node participates in the message flow transaction. Most nodes for which transactionality is relevant have one or more properties that you can configure to dictate behavior. Therefore, you can decide for each individual node whether it participates in the message flow transaction, or operates independently. Typically, these properties include an option of Automatic, so that subsequent nodes in the flow assume the characteristics set by the input node.

If the resource can participate in a transaction, you can configure it so that the work it does is committed or rolled back only when the message flow completes, or when the node completes. Databases and WebSphere MQ queues are examples of resources that you can use in this way. If the resource does not have transactional behavior, all the work that it does is committed immediately. For example, files and HTTP connections do not support transactions.

Updates that are made by a message flow are committed when the flow processes the input message successfully. The updates are rolled back if both of the following conditions are met:

• a node in the flow throws an exception that is not caught by a node other than the input node (for example, the node itself, or a TryCatch node)
• the Catch terminal of the input node is not connected

Read message flow transactions to understand how the broker handles transactions

Start here :

1. setup DB, DB2 or Oracle
2. import the sample and create the queues
3. set up the sample - Important item: configuring ODBC connections for globally coordinated transactions, as DB2
4. run the sample

No em funciona un fluxe que escriu en 2 nodes de sortida. Si el segon node esta ple i no te "Failure" cablejat, el missatge sempre va a "Catch" del node de entrada, sense importar si es transaccional o no.

Top

Open toolkit and select "start from patterns" in Quick Start wizards list. Select "Service Enablement" on left and then MQ request-response.

The solution is to implement a message flow that provides a service entry point. When a service request is received, the reply identifier is stored on an internal queue and the request is forwarded to the provider application. When the response is received from the provider application, the reply identifier is recovered and a web services reply, which contains the provider response, is returned to the requesting application.

Top

If we place a MB flow between App_Origin and App_Server, and the reply message has to be also modified by MB, it is very usual to use a temporary queue in the flow to keep the original Reply-To parameters until the reply message comes

If we want to be state-less, so reply can use another (identical) flow, then the state (msg in tmp q) must go in the message header, guess RFH2.

No - If you're using v7 you shouldn't be using an MQRFH2 but message properties
If working with V7 you can set the message properties and have the system transform them into an RFH2. See the property control parameter on the queue / channel.

An 18 minute Powerpoint presentation about message properties, including how it relates to RFH2 headers: url

Top

First, let's set flow debug preferences.

1. Window + Preferences - start Preferences wizard
2. select Flow Debug - opens the Flow Debugger preferences page
3. make own selections

Use the "Flow Debug" perspective for debugging. In its default mode, this perspective shows 4 windows :

1. Flow Debug view
2. Flow Breakpoints view (shared pane)
3. Flow Debug Message view (shared pane)
4. Message Flow editor

Before you can debug a message flow, you must attach the flow debugger to the flow engine (execution group) where your flow is deployed.

1. switch to the Flow Debug perspective (escarbat vermell que mira cap a la dreta del menu lateral esquerre)
2. click the Attach to Flow Runtime icon on the Toolbar (escarbat vermell a dalt)
3. in the Hosts list box, select the host computer that the flow engine is running on.
4. in the listbox, select select the flow engine containing the flow you want to debug.
5. in the Flow Debug view, double-click the message flow you want to debug.
6. in the Message Flow editor, add a breakpoint to a connection that leads out of the input node. The breakpoint appears.
7. when the next message comes into the flow and arrives at the breakpoint, the flow pauses and the breakpoint icon changes.

Message Flows, page 490 - url

The Message Flow editor has different functions in different perspectives:

• In the Broker Application Development perspective, the Message Flow editor is used to create, graphically display, and edit flows.
• In the Flow Debug perspective, the Message Flow editor is used only to graphically display and debug flows, as described in this topic.

Important workspace setting :
Solapa superior, segunda por la derecha, "Ventana" + "Preferencias" + "Flow Debug" :
Debug Message Broker CSD 3 and beyond

Gracias, Lourdes !

1. install RAC ( without security )
2. start RAC at Port 10002 (RAService.exe @ Guindows)
3. add Breakpoint(s)
4. open perspective Debug
5. "New Configuration" :
   • escollir Flow Project
   • escollir Host
   • escollir Flow Engine
6. send messages and debug flow
7. Right-click on New Configuration + Terminate & Remove

Gracias, Mario !

Top

If we connect the "Exception" pin of the MQ Input node directly to a MQ Output node and there is an Exception in the flow, we shall find a message like this one in the queue :

Top

Qui hi ha escoltant:

Problemes :

• msgflow is not configured to be published
• file korreu.esql can not be added to archive as it is undeployable

Top

Rational Agent Controller (RAC) supports message flow debugging. If you want to use the Debugger, you must install RAC on each system on which you start a broker that hosts message flows. The RAC component that you install on the broker is a server that communicates with the RAC client which is embedded in the Message Brokers Toolkit.

V 6 Installation Guide, page 19 [37/176]

RAC is not required after MQ v 6.1 [javier]

Top

Pre-requisites. Before you can publish (deploy) a configuration using rapid application development (RAD) you need to :

• create a Configuration Manager
• the MQ listener must be running
• create a domain
• add a broker to that domain
• create an execution group within that broker

Top

• the installation samples
• korreus
• praktikes Lourdes
• ping responder

First, we have to configure the product, this is, we need to create a broker domain, that represents the broker runtime environment. A broker domain contains a single Configuration Manager with associated brokers. It may also contain a User Name Server if authentication is used for publish/subscribe messaging.

To quickly create a default configuration, run the Getting Started Wizard :
Start -> Programs -> WebSphere Studio -> MB Toolkit

The queues to use will be :

• PAGER
• SURFWATCH
• SURFWATCH_FAIL
• TEXTMESSENGER
• TEXTMESSENGER_FAIL

Input message

Output message

ESQL

"Getting started with WBI MB" Redbook, 2004, page 28 de 74. "Getting started with the Broker Exercises" RedBook, page 27 of 386.

Top

MB v7 samples in publib. MB samples in "Toolkit samples" in on-line documentation v7 : import, deploy, run.

You can run samples only when you use the information center that is integrated with the WebSphere Message Broker Toolkit : Help -> Samples and Tutorials -> Product = MB tkt - MB

The samples are categorized as either Application or Technology samples.

• airline reservations
• Coordinated Request Reply
• DatabaseInput Node
• data warehouse
• error handler
• large messaging
• message routing
• pager
• scribble
• Solar Pattern Authoring
• user-defined extension
• video rental
• WBI JDBC Adapter to DatabaseInput Node Migration
• Web Service Aggregation

The Technology samples are small Message Broker applications that each show a specific feature of WebSphere Message Broker.

• Controlling and Routing samples
  Aggregation, collector node, simplified database routing, timeout processing
• File Processing samples
  batch processing, file output, managed file transfer, wildcard match.
• Industry samples
  healthcare, TLOG processor.
• Message Formats samples :
  CSV, EDIFACT, FIX, SWIFT, X12, XMLNSC namespaces, XMLNSC validation.
• Message Transformation samples
  JavaCompute node, message map, PHPCompute node, XSL transform.
• Monitoring samples
  WebSphere Business Monitor
• Security samples
  security identity propagation, security policy enforcement point (PEP).
• Transports and Connectivity samples
  Browsing MQ queues, CICS transaction server for z/OS connectivity, CICS Transaction Server for z/OS channel connectivity, CORBA nodes, eMail, HTTPHeader node, IMS Synchronous request, JD Edwards connectivity, JMS codes, JMSHeader node, MQHeader node, SAP callout to a synchronous system, SAP callout to an asynchronous system, SAP connectivity, SCA nodes, TCPIP client nodes, TCPIP handshake, Twineball example EIS adapter.
• Web Service samples
  address book, asynchronous consumer, RESTful Web Service using JSON, SOAP nodes, web services using HTTP nodes, web service gateway, websphere service registry ans repository connctivity.

Amb el Wizard fer la configuracio per defecte :

Get it here.

Once deployed, accessing this url http://9.137.147.65:7080/httpHealthCheck produces {search for biphttplistener at TCPview to find the HTTP Lister port} :

This sample shows how to use the FileInput and FileOutput nodes to read different input files and append them to one output file. It also shows you how to read a file "as is" from one local input directory and write the file to a different local output directory.

Links : File processing -> batch -> run the sample.

This sample shows how to use the SOAPInput, SOAPReply and SOAPRequest nodes to provide and consume a web service. Two sets of example input messages are provided: one set to call the consumer flow which in turn calls the provide flow, and one set to ocall the provider flow directly. This sample can also be extended too show how to set up WS-Security for existing messages flows for both a provider and a consumer.

This sample shows how to use MB to front an existing appplication as a web service - local information center {T400}, T430+MB v9

The Web Services using HTTP nodes sample has two parts:

1. the Web Service Host sample, which demonstrates how to use a message broker to provide an interface to a legacy application as a Web service.
2. the Web Service Client sample, which demonstrates how to call a Web service, and how to process the reply from the Web service.

From WebSphere Message Broker v6.1 onwards, SOAP nodes are provided to implement Web services.

Click install and deploy the sample

Summary of tasks: WebServices using HTTP nodes sample

1. Create the following projects in the workspace: WSHOST_LEGACY1 WSHOST_MSP1 WSHOST_MFP1 WSCLIENT_MSP1 WSCLIENT_MFP1
2. Add the following WebSphere MQ queues: WSHOST1_IN1 WSHOST1_OUT1 WSHOST1_STATE1 WSCLIENT1_IN1 WSCLIENT1_OUT1
3. Deploy the samples into the following new execution groups: WSHOST

Now we have

Now we can run the sample and also resolve problems when running samples

When you run the Web Service Host sample, the broker receives and sends back a SOAP/XML message, over HTTP protocol. The broker is acting as a Web service host.

The broker takes data from this SOAP message, and generates a new message mapped to a C header file. This new message is sent to a legacy application by using WebSphere MQ. The legacy application responds, and data from this response is put into the HTTP reply that is generated by the broker. The broker has fronted a legacy application as a Web service.

To run the Web Service Host sample:

1. start the legacy application

   Run "legacyservice.exe" from "runLegacyApp.bat" to provide params, as qname = WSHOST1_OUT1
   ( at C:\Documents and Settings\Administrator\IBM\wmbt70\workspace\WSHOST_LEGACY1 )
   Source code is in "legacyservice.c"

2. put a SOAP message to the URL that is specified in the WSHOST_WSHOST1.msgflow HTTPInput node, using any freeware tools available for testing web services and sending SOAP messages - SOAP tools, SOAP msg builder - use NetTool (XML only), or CURL, or SOAP UI (with WSDL) or "XML Spy" (30 day trial, complete)

   The sample message is in the WSHOST1IN.xml file
   ( at c:\Documents and Settings\Administrator\IBM\wmbt70\workspace\WSHOST_MFP1 )
   and the URL is http://localhost:7080/samplebrokerwshost

   Using wget:

   set FN=WSHOST1IN.xml set URL=http://localhost:7080/samplebrokerwshost "c:\Program Files (x86)\GnuWin32\bin\wget.exe" "%URL%" --post-file="%FN%" --header="Content-Type:text/xml"

   Where have we said that our web-service has a bphttplistener at port 7800 ?

When using toolkit built-in help, you can install sample objects by clicking on a link that will issue a command like this:

It would be good to know how a MQ "runmqsc" command is issued from within a such a link

• "samples.WBMonitor.res" is a directory under "c:\Program Files (x86)\IBM\SDPShared\plugins"
• there we can find files of type BAR, XML and ESQL, but no CMD's

Top

XP is not supported in IIB 9. Windows 7 is the lowest supported client version (32 or 64 bit).

We start with this file :

Start install by launching installToolkit.bat.
Shared directory (1,5 GB) = C:\Program Files (x86)\IBM\SDPShared ;
Installation Manager v 1.6.2 = C:\Program Files (x86)\IBM\Installation Manager\eclipse ;
Product (262 MB) = C:\IntegrationToolkit90 ;

"Default Configuration" setup:

When you add a message flow to a BAR file in preparation for deploying it to a broker, you can set additional properties that influence its run time operation. These properties are available for review and update when you select the Manage and Configure tab for the broker archive file.

It is very important to be sure you are running the flow version you have in mind, not a previous one.

To assign a flow version number and to verify you have it deployed, we can do:

1. on any flow, under "User Defined Properties", assign Properties -> Description -> Version
2. on the Execution Group, verify you have the correct flow version deployed

Or

1. use $MQSI Author=Sebastia.Net MQSI$ and $MQSI Subflow 1 Version=v1.3.2 MQSI$ in "Long Description" field at the "Properties" of the ".msgflow"
2. then use command
   c:\> mqsireadbar -b %TKT%\workspace\BARfiles\cnt_bar.bar -r BIP1051I: Reading Bar file using toolkit mqsireadbar... cnt_app.appzip (9/01/14 19:55): cnt_flow.msgflow (9/01/14 19:55): Author = Sebastia.Net Subflow 1 Version = v1.3.2

Keywords are extracted from the compiled message flow (the .cmf file) rather than the message flow source (the .msgflow file). Not all the source properties are added to the compiled file. Therefore, add message flow keywords in only these places:

• the label property of a Passthrough node
• ESQL comments or string literals
• the Long Description property of the message flow

Message flow version and keywords

This property specifies whether all additional instance threads are to start when the flow starts. Both flow and node-scope instances will start if this option is set to true. The default value is false.

To set it, select the BAR file on the left pane, expand the "Manage" tab and select the message flow. Select "Workload management" to display and set the values.

For WebSphere MQ messages, this property specifies how many input messages are processed by a message flow before a sync point is taken (by issuing an MQCMIT command). The default value of 1 is also the minimum permitted value. Change this property to avoid frequent MQCMIT calls when messages are being processed quickly and the lack of an immediate commit can be tolerated by the receiving application.
Use the Commit Interval to ensure that a commit is performed periodically when not enough messages are received to fulfill the Commit Count.
This property has no effect if the message flow does not process WebSphere MQ messages.

Toolkit writes in {$workspace}/.metadata/.log

.broker file syntax (expor from MQExplorer):

Top

doc

For information about current fixes, follow the link Recommended fixes in the Download section of the IBM Integration Bus support web page. Click the link for Version 9.0, and select the updates you want to apply.

Fix : d:\fonts\websphere\mbtkt\mb7000_7002\repository.config

URLs :
updates
extensions

Top

Revision control sw list, IBM.

Concurrent Versions System or Concurrent Versioning System, is a client-server free software revision control system in the field of software development. A version control system keeps track of all work and all changes in a set of files, and allows several developers (potentially widely separated in space and time) to collaborate.

From Toolkit, on a project, we do

1. right click on a project name
2. Team -> share project
3. repos := cvs ;
4. location := ip/port ;
5. cvs -> flow : update or check-out ;
6. flow -> cvs : commit ;

Both Java and ESQL in normal circumstances perform broadly the same. There are some things which can be done in ESQL which can't in Java, and may perform better, such as SELECTs against the message tree. If however the customer's skills are predominantly Java then this would be preferred to save learning a new language.

Top

Flow troubleShooting :

• IH03: Message Broker V7 - message display, test & performance utilities; as American Express: "don't leave home without it"!
• IS03 - MB Message Flow Statistics Visualiser
• start accounting :
  • mqsichangeflowstats <broker> -s -e <EGname> -f <message flow> -c active -t basic -n advanced -o usertrace
  • mqsireadlog <broker> -u -e <EGname> -f -o trace.xml
  • mqsiformatlog -i trace.xml -o trace.log
  
  Stop accounting : mqsichangeflowstats <broker> -s -e <EGname> -f <message flow> -c inactive -t basic -n advanced -o usertrace

The following properties are reported (command: mqsireportproperties ? "mqsichangeflowstats" ?)

Resource Statistics can be found here (MB v8)
(MBX -> expand Brokers -> window -> show view -> resource statistics)
(MBX -> expand EGs -> right-click on EG -> [start or stop or view] [msg flow or resource] statistics)

• InitialMemoryInMB
  The initial amount of memory (in megabytes) that the JavaT virtual machine requests from the operating system for memory management during startup.
  Its value may be undefined which is indicated by a value of -1.
• UsedMemoryInMB
  The amount of memory currently used (in megabytes).
• CommittedMemoryInMB
  The amount of memory (in megabytes) that is guaranteed to be available for use by the Java virtual machine.
• MaxMemoryInMB
  The maximum amount of memory (in megabytes) that can be used for memory management.

Interpretation

• how much memory is the JVM using ?
  • CommittedMemoryInMB - how much memory is currently allocated to the JVM
  • MaxMemoryInMB - how big this may grow.
• how often am I doing garbage collection ?
  • CumulativeNumberOfGCCollections - Frequent increase in this indicates the garbage collection (GC) may be excessive. Excessive GC can impact performance.
  • CumulativeGCTimeInSeconds - If this is increasing at more than 2 seconds per 20 second stats interval then consider increasing the JVM Max Heap size
  • Inspect any Java Plugin nodes or Java Compute nodes in your message flows to ensure that you are not creating and deleting many objects that could be reused.
• should I change my min or max heap size ?
  • If your CumulativeGCTimeInSeconds is increasing by more than 2 seconds per stats interval then look to increase the Max Heap Size to reduce this.
  • If your UsedMemoryInMB is never close to your InitialMemoryInMB then this indicates that you are pre-allocating more memory for the heap than necessary and could reduce the JVM Min Heap Size value for you execution group to a value closer to the UsedMemoryInMB value.

WSTE MB v7 Administration and Troubleshooting Made Easy - Shah.pdf [T430\\WS-BI-MB\TroubleShooting]

An Execution Group is a process. If we are under AIX, use nmon to display memory usage. Specifically, verify there is no paging to file system.

If there is paging, everything becomes slow and indicates a problem of capacity, not performance. So, unless we have a very ineficcient flow, the only solution is to add physical memory.

• use references for fields of 3 or more levels, as "InputRoot.XMLNSC.msg.a"
• avoid using "SET OutputRoot = InputRoot" if message does not change.
  To propagate InputRoot and InputLocalEnvironment, change Compute Mode to "ExceptionList", selecting "RETURN TRUE"
• dont use subscripts

Sample (1) :

Sample (3):

• use TDS, not CWF
• if v8 or v9, use DFDL

Top

• an internal error occurred during: "Cleaning up symbolic strings in the WebSphere Integration Developer index"
• conectar MB a Reuters mediante protocolo "TOF", del cual se conoce la API.
  Sabiendo la API, lo mejor es implementar el protocolo con Java Compute nodes.
• hi ha alguna comanda per veure la versió d'un BAR file ? Solucio

Top

• Infocenter
• IP04 : designing for performance
• Q04 : MB troubleshooting and common pitfalls (debug, etc)
• lista exhaustiva de adaptadores : url {Emir, 12/2010}
• HTTP transport nodes in MB v6 - url.
• Integrating with TCP/IP using MB : url, MATIP & BATAP (airline) & ISO8583 (banking).
• MB forum at developerWorks
• SOAP SSL example flows (int)
• SWIFT or ISO 15022
• nice product on top of MB : FTM (Francisco, 2013)
• Samples {*****}
• ESQL vs Java
• Error Handling in WMB : (1) (2).
• MQ admin from MB using admin agent