• No results found

Adapter Framework Administration

In document BIT480 - PI – Operations(Col74) (Page 72-82)

The Adapter Framework is part of every SAP NetWeaver PI installation. In addition, another Adapter Framework can be installed on a separate host, and be configured and monitored centrally. The AF is also part of the Partner Connectivity Kit, where it brings its own configuration and monitoring tool.

Figure 44: SAP PI 7.1 Architecture

The Adapter Framework is based on the SAP J2EE Engine and the J2EE Connector Architecture (JCA). It provides interfaces for configuring, managing, and monitoring adapters.

You can find the Services in the SAP NetWeaver Administrator under: Operation Management→ Systems → Start & Stop → Java EE Services. Search for XPI using the filter function.

Figure 45: Services of the Adapter Framework in the NWA

2008 © 2008 SAP AG. All rights reserved.

65

For configuration purposes, a restart of the services may become necessary. If you want to stop the adapter service, click on the service name and choose Stop from the context menu that opens on the bottom of the screen. An inactive service can be started with Start from the context menu.

Figure 46: Start and Stop the Entire XPI Service

If there are any problems with a certain channel communication, then turning the dedicated communication channel off is the better choice than stopping the whole service.

Figure 47: Activate/Inactivate the Communication Channel in the Integration Directory

Afterwards ensure that the changes recorded in a changelist are finally activated.

As an alternative the communication channels can be administered using the Communication Channel Monitoring accessible from the RWB→ Component Monitoring and selection the Adapter Engine. Select in the Status tab the Communication Channel Monitoring. Search for the dedicated channel and click on the details to Start or Stop the channel.

2008 © 2008 SAP AG. All rights reserved.

67

Figure 48: Use Start/Stop in Communication Channel Monitoring

XPI Service CPA Cache Facts:

• Cache for Directory & Repository data (CPA Cache)

• Cache is filled during activation within directory

• Contains objects from directory Services

Parties

Bindings (inbound/outbound; sender/receiver agreements) Channels

Adapter metadata Module configuration

• Used in Adapter Framework during runtime

You can access these parameters through the NWA: Configuration Management → Infrastructure → Java System

Proper-ties.Underneath the separator Details choose Services.

Select XPI Service: CPA Cache.

Extended Details contains the parameters for the CPA Cache Service.

Figure 49: Properties of CPA Cache Example: SLD.selfregistration.hostName

Parameters for the CPA Cache Service:

cacheType:

DIRECTORY for Adapter Engine, PCK for Partner Connectivity Kit. This property can only be modified offline, using the configtool.

directoryPwd, directoryURL, directoryUser:

These connection parameters are taken for the cache refresh,

when SLDAccess is set to false. Otherwise, the settings are taken from the ExchangeProfile.

SLDAccess:

This parameter is set to true for the Adapter Engine and set to false for Partner Connectivity Kit (usually no SLD available). If the parameter is set to true: the SLD address is taken from the central Exchange Profile.

2008 © 2008 SAP AG. All rights reserved.

69

Figure 50: Parameters for the CPA Cache Service

The Parameter SLD.selfregistration.hostName is used normally in HA setups to force the CPA cache service to register with the host name of the SAP WebDispatcher (or any other used loadbalancer). It can also be used in cases when network communication requires for the use of fully qualified host names (FQDNs), because local host names are not always resolved properly.

You can perform a CPA Cache refresh manually using the URL

http://host:port/CPACache/refresh?mode=full or via the menu with PIDIRUSER.

You can reach the menu by navigating to PI Tools startpage→ Administration → (Repository tab/Directory tab) Data Cache Overview.

You may use the cache refresh functions only if configuration requires you to reset the content of the PI caches and for testing purposes, because a working cache refresh is crucial for a properly configured system.

Figure 51: Manual CPA Cache Refresh

Audit Log Configuration in the Adapter Engine:

During message processing on the AAE the AuditLog information is per default only kept in memory and only up to 2000 entries per server node.

AuditLog relevant properties of 'XPI Service: Messaging System'

Property Description

messaging.auditLogEnabled

Value type: Boolean Default value: true

This property enables or disables the audit log. You can modify this property online.

messaging.auditLog.memoryCache

Value type: Boolean Default value: true

This property enables or disables the audit log caching. You can modify this property online.

2008 © 2008 SAP AG. All rights reserved.

71

The first decision you take is to enable or disable the AuditLog in general. Usually in developement and test systems you would like to see what is going on, but on a productive system you might want to have better performance with the drawback if less analysis information. Especially for the AAE with local processing mode, writing audit logs is relatively expensive and depending on needs the disabling of the audit log entries can be reasonable on a productive system.

The next decision is how many AuditLog data and where you want to keep the AuditLog data. If you decide for the new Memory Cache functionality than only 2000 entries are kept per J2EE Engine Server Node for successfully processed messages. For failing message the AuditLog data is always stored and persisted in the database. Only if you want to record also for successfully processed messages the AuditLog data, you will have to switch the value of the messaging.auditLog.memoryCache property to false.

Important differente between J2EE Engine 7.0 and 7.1 in regards to service parametrization:

Figure 52: JEE Properties Maintenance Option - 7.1 versus 7.0

Installing JDBC and JMS drivers:

For the JDBC and JMS adapters, you need external drivers that have to be provided as Java archives (jars) by the provider of the database or messaging system, respectively. For SAP NetWeaver 7.1 it is mandatory to use JEE5 compliant external drivers.

You have to deploy these drivers to enable them to be used by the adapters in the SAP J2EE Server. The Adapter Engine installation provides the archive aii_af_jmsproviderlib.sda, to which you have to add the required driver(s).

Figure 53: Vendor Specific Driver (JMS, JDBC)

• Add driver(s) (jar) to aii_af_jmsproviderlib.sda

• Unpack, modify and pack provider.xml from/to aii_af_jmsproviderlib.sda;

for packing: use the same path/server

• Deploy aii_af_jmsproviderlib.sda (ignore version)

• JMS only:

Remove the javax/jms classes from the driver jar first – this is because the javax/jms classes belong to the J2EE server base classes and confuse the class loader in the server when deployed more than once

2008 © 2008 SAP AG. All rights reserved.

73

Figure 54: Example: provider.xml with IBM WebSphere MQ

In the picture above you can see an example of IBM WebSphere MQ Java libraries to be added to aii_af_jmsproviderlib.sda:

• com.ibm.mq.jar

• com.ibm.mqbind.jar

• com.ibm.mqjms.jar

• connector.jar

After you have carried out these steps, deploy the resulting SDA file using the JSPM. You can check the deployment in the SAP NetWeaver Administrator.

For additional information, refer to SAP Notes on the Service Marketplace:

1138877: PI 7.1 : How to Deploy External Drivers JDBC/JMS Adapters

831162: FAQ: XI 3.0 / PI 7.0 / PI 7.1 JDBC Adapter

In document BIT480 - PI – Operations(Col74) (Page 72-82)