Update an IBM Websphere Application Server Installation from 7.1 to 7.2
The following steps describe how to update the Camunda artifacts on an IBM was application server in a shared process engine setting. For the entire migration procedure, refer to the migration guide. If not already done, make sure to download the Camunda 7.2 IBM was distribution.
The update procedure takes the following steps:
- Uninstall the Camunda libraries and archives
- Add the new Camunda libraries
- Install optional Camunda dependencies
- Configure process engines
- Create a shared Camunda library
- Install the Camunda archive
- Install the Camunda 7 web applications
- Associate existing applications with shared library
In each of the following steps, the identifiers $*_VERSION
refer to the current version and the new versions of the artifacts.
Changing Platform Configuration
Depending on your chosen feature set for Camunda 7, some of the (optional) migration steps may require to change the configuration of Camunda 7. The Camunda enterprise archive (EAR) contains a default platform configuration. If you want to change this configuration, you can replace it as described in the deployment descriptor reference.
1. Uninstall the Camunda Libraries and Archives
First, uninstall the Camunda web applications, namely the Camunda REST API (artifact name like camunda-engine-rest
) and the Camunda applications Cockpit, Tasklist and Admin (artifact name like camunda-webapp
).
Uninstall the Camunda EAR. Its name should be camunda-ibm-was-ear-$PLATFORM_VERSION.ear
. Then, uninstall the Camunda job executor adapter, called camunda-ibm-was-rar-$PLATFORM_VERSION.rar
. Delete the J2E connection factory and activation specification created for Camunda 7.1.
After stopping the server, delete all Camunda 7 related libraries from the $WAS_HOME/lib/ext
directory. These should be:
camunda-engine-$PLATFORM_VERSION.jar
camunda-bpmn-model-$PLATFORM_VERSION.jar
camunda-xml-model-$PLATFORM_VERSION.jar
mybatis-$MYBATIS_VERSION.jar
java-uuid-generator-$VERSION.jar
joda-time-$VERSION.jar
slf4j-api-$VERSION.jar
Note that the non-Camunda artifacts may be depended on by other applications.
2. Add the New Camunda Libraries
Camunda 7.2 ships a new installation approach for Camunda 7 on IBM was. The Camunda 7 libraries are now available through a shared library the server administrator has to create.
Copy the Camunda core modules from the folder $WAS_DISTRIBUTION/modules/lib
to a folder that can be referenced from your IBM was installation when you later create the shared library. In detail, these are the following artifacts:
camunda-engine-$PLATFORM_VERSION.jar
camunda-bpmn-model-$PLATFORM_VERSION.jar
camunda-cmmn-model-$PLATFORM_VERSION.jar
camunda-xml-model-$PLATFORM_VERSION.jar
mybatis-$MYBATIS_VERSION.jar
java-uuid-generator-$VERSION.jar
joda-time-$VERSION.jar
slf4j-api-$VERSION.jar
If you have been previously using the Camunda LDAP plugin, also copy the following artifact:
camunda-identity-ldap-$PLATFORM_VERSION.jar
3. Install Optional Camunda Dependencies
There are artifacts for Camunda Connect, Camunda Spin, the Freemarker template language and Groovy scripting that may optionally be added to the shared library folder. Since all these artifacts add new functionality, the following steps are not required for migration.
Note: The default Camunda configuration file contained by the Camunda EAR automatically activates the newly introduced, optional Camunda dependencies, Camunda Spin and Connect. If you do not use a custom Camunda 7 configuration as described here and do not intend to do so, you must install the Camunda Spin and Connect core libraries to the shared libraries folder.
Not Using Connect/Spin
If you do not want to use Camunda Connect or Camunda Spin, you cannot use the default Camunda 7 configuration that is contained in the Camunda EAR. In this case, make sure to change the configuration location as described [here][configuration-location]. As a starting point, you can copy the default configuration from $WAS_DISTRIBUTION/modules/camunda-ibm-was-ear-$PLATFORM_VERSION.ear/camunda-ibm-was-service-$PLATFORM_VERSION.jar/META-INF/bpm-platform.xml
and remove the <plugin>
entries for the classes ConnectProcessEnginePlugin
and SpinProcessEnginePlugin
.
Camunda Connect
If Camunda Connect is intended to be used, copy the following library from $WAS_DISTRIBUTION/modules/lib
to the Camunda
shared library folder:
camunda-connect-core-$CONNECT_VERSION.jar
camunda-commons-logging-$COMMONS_VERSION.jar
camunda-commons-utils-$COMMONS_VERSION.jar
slf4j-api-$SLF4J_VERSION.jar
If you use a custom Camunda 7 configuration file, Camunda Connect functionality has to be activated for a process engine by registering a process engine plugin (note that if you use the default configuration, this step is not necessary):
<?xml version="1.0" encoding="UTF-8"?>
<bpm-platform ... >
<process-engine name="default">
...
<plugins>
... existing plugins ...
<plugin>
<class>org.camunda.connect.plugin.impl.ConnectProcessEnginePlugin</class>
</plugin>
</plugins>
...
</process-engine>
</bpm-platform>
Camunda Spin
If Camunda Spin is intended to be used, copy the following library from $WAS_DISTRIBUTION/modules/lib
to the Camunda
shared library folder:
camunda-spin-core-$CONNECT_VERSION.jar
camunda-commons-logging-$COMMONS_VERSION.jar
camunda-commons-utils-$COMMONS_VERSION.jar
slf4j-api-$SLF4J_VERSION.jar
If you use a custom Camunda 7 configuration file, Camunda Spin functionality has to be activated for a process engine by registering a process engine plugin (note that if you use the default configuration, this step is not necessary):
<?xml version="1.0" encoding="UTF-8"?>
<bpm-platform ... >
<process-engine name="default">
...
<plugins>
... existing plugins ...
<plugin>
<class>org.camunda.spin.plugin.impl.SpinProcessEnginePlugin</class>
</plugin>
</plugins>
...
</process-engine>
</bpm-platform>
Groovy Scripting
If Groovy is to be used as a scripting language, add the following artifacts to the Camunda
shared library folder:
groovy-all-$GROOVY_VERSION.jar
Freemarker Integration
If the Camunda integration for Freemarker is intended to be used, add the following artifacts to the Camunda
shared library folder:
camunda-template-engines-freemarker-$TEMPLATE_VERSION.jar
freemarker-2.3.20.jar
camunda-commons-logging-$COMMONS_VERSION.jar
camunda-commons-utils-$COMMONS_VERSION.jar
slf4j-api-$SLF4J_VERSION.jar
4. Configure Process Engines
Script Variable Storing
As of 7.2, the default behavior of script variables has changed. Script variables are set in e.g. a BPMN Script Task that uses a language such as JavaScript or Groovy. In previous versions, the process engine automatically stored all script variables as process variables. Starting with 7.2, this behavior has changed and the process engine does not automatically store script variables any longer. You can re-enable the legacy behavior by setting the boolean property autoStoreScriptVariables
to true
for any process engine in the bpm-platform.xml
:
<?xml version="1.0" encoding="UTF-8"?>
<bpm-platform ... >
...
<process-engine name="default">
...
<properties>
... existing properties ...
<property name="autoStoreScriptVariables">true</property>
</properties>
...
</process-engine>
...
</bpm-platform>
As an alternative, process application developers can migrate script code by replacing all implicit declarations of process variables in their scripts with an explicit call to execution.setVariable('varName', 'value')
.
5. Create a Shared Camunda Library
Next, start up the server and create a shared library named Camunda
for these artifacts. In detail, go to Environment / Shared libraries, choose the correct scope in which your EAR and applications will run and define a new shared library. Name it Camunda
. This name is mandatory. Enter as classpath the path where you have copied the Camunda shared libraries, e.g., /opt/IBM/was/AppServer/profiles/AppSrv01/camunda-shared-libs
. Enable the Use an isolated class loader for this shared library checkbox.
JDBC Drivers, DataSource and WorkManager settings stay the same as compared to 7.1.
6. Install the Camunda Archive
Install the Camunda EAR, i.e., the file $WAS_DISTRIBUTION/modules/camunda-ibm-was-ear-$PLATFORM_VERSION.ear
. During the installation, the EAR will try to reference the previously created Camunda
shared library.
As of version 7.2, the Camunda job executor resource adapter (RAR) that you uninstalled in step 1 is part of the Camunda EAR and therefore does not need to be installed separately.
7. Install the Camunda Web Applications
Camunda REST API
The following steps are required to update the Camunda REST API on an IBM was instance:
- Deploy the web application
$WAS_DISTRIBUTION/webapps/camunda-engine-rest-$PLATFORM_VERSION-was.war
to your IBM was instance. - Associate the web application with the
Camunda
shared library.
Camunda Cockpit, Tasklist, and Admin
The following steps are required to update the Camunda web applications Cockpit, Tasklist, and Admin on an IBM was instance:
- Deploy the web application
$WAS_DISTRIBUTION/webapps/camunda-webapp-ee-was-$PLATFORM_VERSION.war
to your IBM was instance. - Associate the web application with the
Camunda
shared library.
8. Associate Existing Applications with Shared Library
The new shared Camunda
library affects process applications as well because the Camunda 7 libraries which are available with 7.1 on the global classpath are gone. In order to make the Camunda libraries available to applications, the Camunda
shared library has to be associated with every process application. If this is not done, the applications fail due to missing resources.
LDAP Entity Caching
With 7.2, it is possible to enable entity caching for Hypertext Application Language (HAL) requests that the camunda web applications make. This can be especially useful when you use camunda in combination with LDAP. To activate caching, the camunda webapp artifact has to be modified and the pre-built application cannot be used as is. See the REST Api Documentation for details.