Known Issues for Oracle SOA Products and Oracle BPM Products for 12c Release 2 (12.2.1.0.0)

Known Issues for Oracle SOA Products and Oracle BPM Products for 12c Release 2 (12.2.1.0.0)

This page lists issues and workarounds for Oracle SOA and Oracle BPM Products 12c Release 2 (12.2.1.0.0). For a list of issues and workarounds pertaining to Oracle SOA Installation, Upgrade, High Availability, Enterprise Deployment, Performance Tuning, and Web Services, as well as SOA on IBM Websphere, see the Oracle Fusion Middleware Release Notes.

For new features in Oracle SOA Suite 12.2.1, see New Features in SOA Suite.


Oracle Real-Time Integration Business Insight


Search for Real-Time Integration Business Insight 12.2.1.0.0 on Oracle Software Delivery Cloud to locate the patches needed to install Oracle Real-Time Integration Business Insight. You must download and install the patches for the updates related to Oracle Service Oriented Architecture, Oracle Business Activity Monitoring, and Oracle Service Bus. Carefully read and follow the instructions provided in the Readme.txt file (packaged along with the product) to install Oracle Real-Time Integration Business Insight.

OS and Feature Support

This section lists all the known issues/limitations related to OS/Feature support.

Support for Windows OS will be limited for a brief period

  • Bug ID: 22926111 and 22926079
  • Added on: 22 March 2016
  • Platform: Windows

There is a problem with mapping to implementation and editing of imported models on a Windows machine if you are using Windows for server deployment. However, you can run a browser on Windows to access a non-Windows server. All Windows versions are impacted.

This is a limitation for this release.

Workaround

None. However, an additional patch will be provided shortly after the Oracle Real-Time Integration Business Insight product is announced.

Installation/Configuration

This section lists all the known issues/limitations related to Installation/Configuration.

When you roll back the OPatch or Uninstall Insight, some data remains

  • Bug ID: 22858099
  • Added on: 9 March 2016
  • Platform: All

When you uninstall or roll back the Insight OPatch, the Insight related library and applications are left over. These may result in deployment errors after restarting the servers.

Workaround:

You must delete the following Library and applications manually from the Oracle WebLogic Administration Console and restart the Managed Server:

oracle.soa.procmon (12.2.1,12.2.1) - Library

oracle.soa.procmon.agent (12.2.1,12.2.1) - Library
oracle.soa.procmon.ui (12.2.1,12.2.1) - Library
insight - APP
insight-soa-agent - APP
insight-ui - APP
procmon-listener - APP
ProcMonJMSModule - JMS
com.oracle.procmon.engine.management.mbeans.impl:name=InsightConfigBean,type=ConfigMBean ConfigMBean

Oracle Real-Time Integration Business Insight is not compatible with Oracle HTTP Server 12.1.3 Version

The Installation Planning Guide provides a starting point for determining the proper installation path for your environment. The Quick Installation Guide will help you quickly get started for development or evaluation environments. Refer to the Installation Guide for complete information (e.g., installing into a production environment, clustering, de-installation, etc).

  • Bug ID: 22756704
  • Added on: 8 March 2016
  • Platform: All

Oracle Real-Time Integration Business Insight is not compatible with 12.1.3 version of Oracle HTTP Server. Model activation succeeds if you use the 12.2.1 version of Oracle HTTP Server instead.

This is a limitation for this release.

Insight Queue must be Non-persistent

  • Bug ID: 22861157
  • Added on: 8 March 2016
  • Platform: All

Insight agents heavily use JMS for data collection. The throughput of the underlying SOA Composites/Service Bus projects on which the Insight models are based can become degraded.

This is a known issue for this release.

Workaround

It is recommended that the deliver mode of the Insight JMS Queue (named ProcMonEventQueue) is updated to "non-persistent". This can be done from the Oracle WebLogic Administration Console as listed below.

  • Login to Oracle WebLogic Administration Console.
  • Click the Lock and Edit button from the console change center (only required for production domains).
  • Navigate from left panel to Services->Messaging-> JMS Modules.
  • Click ProcMonJMSModule.
  • Click ProcMonEventQueue*. You will find one or two of them based on the clusters you have configured on your domain.
  • Click the "overides" tab.
  • Change the "Delivery Mode Override" to "Non-Persistent" and save the changes.
  • Click "Activate Changes" from the change center (only required on production domains).

Mapping to Database Adapter is not supported in Compact Domain

  • Bug ID:
  • Added on; 22 March 2016
  • Platform: All

Mapping a milestone to a database adapter (request or response) leads to errors at runtime if indicators have been assigned to the milestone. This issue is limited to a compact domain setup and not happen in an expanded domain configuration.

This is a limitation for this release.

Workaround

Either configure an expanded domain or follow the instructions below:
For request: Map the milestone to a BPEL assign activity before the invocation of the database adapter and extract indicator values from a BPEL variable available at this point. If no such assign activity exists, you need to create it.
For response: Map the milestone to a BPEL assign activity after the invocation of the database adapter and extract indicator values from a BPEL variable at this point. If no such assign activity exists that maps the necessary values (required indicator values) to a BPEL variable, you need to create it.

Oracle BPEL PM supports the use of activities with same names across multiple scopes. Mapping a milestone to such an activity may cause unexpected behavior and is not supported

  • Bug ID: 22937770
  • Added on: 16 March 2016
  • Platform: All

Activity names must be unique. If they are not unique, it may cause an unexpected behavior.

Oracle BPEL PM supports the use of variables with same names across multiple scopes. Defining extraction criteria for an indicator from such a variable may cause unexpected behavior and is not supported

  • Bug ID: 22244803
  • Added on: 16 March 2016
  • Platform: All

Variable names must be unique. If they are not unique, it may cause an unexpected behavior

An Indicator with normalize-spaces XPath function does not capture values

  • Bug ID: 22645839
  • Added on: 10 March 2016
  • Platform: All

If a model has any indicator that uses a normalize-spaces XPath function in its extraction criteria, such indicator does not capture any values at runtime.

This is a known issue for this release.

XPath functions are not supported while defining extraction criteria using BPEL variables

  • Bug ID: 22697862
  • Added on: 8 March 2016
  • Platform: All

While defining extraction criteria for an indicator or unique instance identifier based on a BPEL variable, usage of XPath functions is not supported. XPath functions can be used in defining extraction criteria for mapping targets other than BPEL variables.

This is a known issue for this release.

Only “On Entry” mapping supported for Throw/Rethrow Activities

  • Bug ID: 22644657
  • Added on: 8 March 2016
  • Platform: All

When mapping to throw/rethrow activities, only “On Entry” is supported. The throw/re-throw activities for the exit/completion criteria are not captured.

This is a known issue for this release.

Links to Documents in the Help Drawer Do Not Work

  • Bug ID: 22980031
  • Added on: 22 March 2016
  • Platform: All

The links to documents in the Help Drawer on the Home page do not work. An additional patch will be provided shortly.

This is a known issue for this release.

Workaround

None

Links to the Related Topics in Help Content Do Not Work

  • Bug ID: 22980040
  • Added on: 21 March 2016
  • Platform: All

The links in the Related Topics section of each of the Help Topics do not work. An additional patch will be provided shortly.

This is a known issue for this release.

Workaround

None

Refresh button does not work in Custom Dashboard

  • Bug ID: 22811363
  • Added on: 21 March 2016
  • Platform: All

When using custom dashboards, clicking Refresh Data does not refresh the custom dashboards.

This is a known issue for this release.

Workaround

On the Custom Dashboard pane, expand the filter area, edit any filter and click Apply. This will refresh the data.

Input Box for Navigating through Pages in Instance List Page Does Not Work

  • Bug ID: 22838669
  • Added on: 14 March 2016
  • Platform: All

The input box to enter a page number for navigating through pages in the Instance List page does not work.

This is a known issue for this release.

Workaround

Use the other navigational controls like forward or backward to navigate through the pages.

Duplicate Text and Components are seen when you click Reports

Reports

This section lists all the known issues/limitations related to Reports.

  • Bug ID: 22465937
  • Added on: 9 March 2016
  • Platform: All

When you navigate to the Reports tab in Insight, duplicate text and components are seen on the screen.

This is a known issue for this release.

Workaround

Click any of the reports available in the left panel to reload the page. When the page reloads, the screen looks normal.

Summary of criteria used in generating report (at the bottom of the Reports page) uses internal names for model/milestone/status

Reports

This section lists all the known issues/limitations related to Reports.

  • Bug ID: 22817296
  • Added on: 8 March 2016
  • Platform: All

The summary of the criteria chosen by the user to generate the report may not match precisely what was actually chosen.

This is a known issue for this release.

Workaround

None

Hardcoded Strings appear in Milestone Dashboards

Consoles/Dashboards

This section lists all the known issues/limitations related to Consoles/Dashboards.

  • Bug ID: 22858503
  • Added on: 8 March 2016
  • Platform: All

Not all of the content in the Milestone Dashboards like 'Passed Milestones', 'Active Instances', 'Instance Errors' and 'Avg Instance Completion Time' has been localized.

This is a known issue for this release.

Workaround

None

Misleading time displayed in Dashboards

  • Bug ID: 22697557
  • Added on: 8 March 2016
  • Platform: All

In the case of a bar or area chart (either milestone dashboards or custom dashboards) if only one data item is rendered (i.e. one bar in case of bar chart and one point in case of area chart), there could be a time displayed against that data item on the X axis. This time value is to be ignored as the data will always be for one entire day as per the date that is also displayed.

This is a known issue for this release.

Workaround

Ignore the value of the time displayed and go only by the date shown since each point on the X axis represents data for one whole day.

Re-importing a model over a model in “failed” state retains state of “failed”

Import/Export

This section lists all the known issues/limitations related to Import/Export.

  • Bug ID: 22712993
  • Added on: 8 March 2016
  • Platform: All

When the activation of an imported Insight model fails, user can remove the model using the "Discard" action. After the "Discard" if the user re-imports the same model to the system, it will show the newly imported model in the failed state. This is because the initial "Discard" action did not completely cleanup the previous metadata.

This is a known issue for this release.

Workaround

The user can activate the re-imported model (which is in a failed state) and it should succeed if the original issue which has caused the failure has been resolved.

Unique IDs not generated for Models/Connections across Systems

Import/Export

This section lists all the known issues/limitations related to Import/Export.

  • Bug ID: 22492868
  • Added on: 8 March 2016
  • Platform: All

When creating a new Insight model or connection, an internal ID is generated based on the name of the model or connection. These IDs are potentially not unique across environments (but will be unique within a single environment). For example, DuplicateConnectionName-Test and DuplicateConnectionName-Production exported from 2 different source systems will lead to overwriting of model in the target system during import operation.

This is a known issue for this release.

Workaround

Do not select the 'Overwrite Any Existing Drafts' check box during Model import operations if you are using multiple Insight environments for model export.


Partial IWS Report on HA-DR Environment


  • Bug ID: 21914240
  • Added on: 20 April 2017
  • Platform: Windows

On SOA 12.2.1 Disaster Recovery Environment, partial/incorrect IWS Report is generated for certain time intervals. This happens only HA-DR environment. This works fine when you specify another time interval.

WADL File Changes Every Time a New Property is Added or Update

This section lists all the known issues/limitations related to Installation/Configuration.

When you roll back the OPatch or Uninstall Insight, some data remains

  • Bug ID: 21650448
  • Added on: 20 April 2017
  • Platform: All

In JDeveloper the WADL file changes every time when the new property added or updated from the "REST Method Definition" Window. The problem occurs when parameters having any primitive data types. If default attribute is not defined in "REST Method Definition" then default attribute does not exists for primitive types. check the following param declaration.

Workaround:

Create another parameter param2 with int type from "Rest Method Definition". Verify the WADL file.



After the second parameter creation the wadl file is updated. The issue is when the WADL file updated, the default="" is added to previous parameters.

Addition of Tuning Knobs PollInerval and PollTimeout

  • Bug ID: 21511079
  • Added on: 20 April 2017
  • Platform: All

Starting with release 12.2.1, two tuning knobs PollInterval and PollTimeout have been introduced in EDN Configuration.

PollInterval (default value is 5 seconds) allows a user to specify the seconds to delay by an inbound poller thread if no event is received by the EDN subscriber before this poller thread resumes next poll cycle.

PollTimeout (default value is 1 second) allows a user to specify the seconds used by an inbound poller to synchronously receive an event from subscriber. The PollTimeout is the time after which receive call from the poller times out if no event is received.

Additional Parameter Required to Compile Project with MDS

  • Bug ID: 21427061
  • Added on: 20 April 2017
  • Platform: All

An additional parameter soa.oracle.home is introduced in 12.2.1 with the value /soa.This variable is created in JDeveloper. This parameter helps in the working of maven command. The new command to run the compiler is: mvn -settings=/home/$USER/.m2/soa_maven_plugin_settings.xml -DshowProcessLog=false -DoracleHome=/scratch/dchin/fmwhome12 -DsoaOracleHome=/scratch/dchin/fmwhome12/soa-DappHome=/scratch/dchin/tmp/bugs/ bug_21352208/tmp/MavenTestAutoApp compile.

Routing does not work

  • Bug ID: 21389534
  • Added on: 20 April 2017
  • Platform: All

DVM in a Mediator component carries several attributes that govern the routing rule run time behavior, including: EndpointURI, Service, Port, Operation, and so on. We can conditionally set overriding values for each of these attributes.

Workaround

To make this routing work, we should make sure that all the targetnamespaces for ws1 and ws2 are the same. The reference files are

BPELProcess1.wsdl, BPELProcess1.bpel and composite.xml.

Non-durable Subscribers Randomly Lose Events

  • Bug ID: 21117990
  • Added on: 20 April 2017
  • Platform: All

If non-durable subscribers are required, to prevent from losing messages, you need to set endpoint property adapter.jms.SinglePollCycle to false. By default it is true. This will toggle back to old (12.1.3) model.

Workaround

Use durable EDN subscribers.

Enhance Composite Loading on Cluster Nodes

  • Bug ID: 21438699
  • Added on: 20 April 2017
  • Platform: All

You may need to warm up all nodes in cluster before running stress runs. In event of stress scenario, all nodes may not be warmed up and composites configured for lazy loading may not be loaded. In this case, deploy er thread may not find resources on the cold node as it competes with the incoming new requests.

For composites that run into these scenarios, lazy loading may be turned off.

Workaround

NA

Escape Characters Not Supported in XPath Functions

  • Bug ID: 20627840
  • Added on: 20 April 2017
  • Platform: All

The escape characters used to escape the reserve words/keywords in JavaScript overlap which in XPath functions are not supported.

This is a known issue for this release.

Workaround

Use the array selector in JavaScript. Any property that cannot be de-referenced via . notation because of JavaScript language constraints (ie. the - sign) can be de-referenced via: obj["prop-name"].

b2b.hideKeystoreLocation in EM to limit access to keystore

  • Bug: 22013344
  • Added: 28-March-2017
  • Platform: All

In EM, the flag b2b.hideKeystore.Location limits access to the JKS Keystore in Fusion Apps. If the flag is set to true, keystore location will be obscured and uploading certificate will be disabled. The flag has no effect outside of Fusion Apps.

This is a known issue for this release.

Update the maven command

Bug: 21938384

Added: 28-March-2017
Platform: All

You need to update the maven command while running on composites that have MDS from:
mvn --settings=/settings.xml package
-DshowProcessLog=false -DsoaOracleHome=/soa -DoracleHome= 
-DappHome=

To:
mvn --settings=/settings.xml package
-DshowProcessLog=false -DsoaOracleHome=/soa -DoracleHome= 
-DcommonComponentsHome=/oracle_common -DappHome=

ClassNotFound Exception thrown when compiling composite that has REST Adapter

  • Bug: 21903889
  • Added: 28-March-2017
  • Platform: All

A ClassNotFound Exception is thrown while compiling a composite that has a REST Adapter. This is because of the Jersey change.

Clear the local maven repo and remove the following lines from oracle_common/modules:


remove the following files from oracle_common/modules:
jersey-client-1.18.jar
jersey-core-1.18.jar
jersey-json-1.18.jar
jersey-server-1.18.jar
jersey-servlet-1.18.jar
asm-3.1.jar

Perform mvn push using the following command:
mvn --settings=/settings.xml
com.oracle.maven:oracle-maven-sync:push -DpushDuplicate=false
-DfailOnError=false -DoracleHome=

Go to the local repository and open file:
/com/oracle/soa/plugin/oracle-soa-plugin/12.2.1-0-0/oracle-soa-plugin-12.2.1-0-0.pom
Add the following within the  

com.oracle.weblogic
org.glassfish.jersey.core.jersey-server
12.2.1-0-0

Login/Logout to SOA Composer Throws Error on UI/Logs

  • Bug: 21873309
  • Added: 28-March-2017
  • Platform: All

When you try to open multiple instances of SOA Composer, login/logout throw an error. This is because opening multiple instances is not secure.

Workaround

NA

Error thrown while accessing SOA Composer and BPM Worklist

  • Bug: 21873299
  • Added: 28-March-2017
  • Platform: All

The following error is seen in SOA Composer and BPM Worklist logs from soa-infra page




Failed to replicate a non-serializable object.

java.rmi.UnmarshalException: error unmarshalling arguments; nested exception is:

java.io.EOFException

at weblogic.utils.StackTraceDisabled.unknownMethod()

Caused By: java.io.EOFException

Workaround

NA

Subsequent login attempts with Incorrect Credentials displays Blank Page

  • Bug: 21863448
  • Added: 28-March-2017
  • Platform: All

When you login with incorrect credentials to SOA Composer, a blank page is displayed upon subsequent unsuccessful logins with no error message. This is a known issue.

Workaround

NA

Right-click on Partition does not show Deploy option

  • Bug: 21815072
  • Added: 28-March-2017
  • Platform: All

When a user with default deployer role logs in to EM and right-clicks on default partition, the option to deploy the composite is not seen.

Workaround

Navigate to the default partition home page to see the deployment option.

Issue while selecting multiple rows in the endpoint table on system dashboard

Reports

This section lists all the known issues/limitations related to Reports.

  • Bug: 21813223
  • Added: 28-March-2017
  • Platform: All

When you select multiple endpoints from the table and try to perform some actions, for example, enabling the seledcted endpoints, the confirmation window sometimes displays only 1 row as selected. This is a known issue.

Workaround

Select only one row at a time.

Work Manager Count not getting updated in EM

Reports

This section lists all the known issues/limitations related to Reports.

  • Bug: 21799928
  • Added: 28-March-2017
  • Platform: All

Metrics in Work Manager page in EM is not getting updated. This is a known issue.

Workaround

REST BPEL Creates Receive/Reply Block

  • Bug: 21752632
  • Added: 28-March-2017
  • Platform: All

When you use JDEV to create a 12.2.1 type REST Composite (without WSDL), a reference to another composite must be created first and then the BPEL to connect to the reference. This creates a Receive/Reply block in the BPEL connecting to the reference.

This is a known issue.

EM Layer changes to grant user-defined ESS permissions for SOA

  • Bug: 21367448
  • Added: 18-March-2017
  • Platform: All

From EM creating a job definition and specifying access control for native application, doesn't list the AppRoles of other stripes configured for EssNativeHostingApp. This is a problem for SOA and OSB because the user would like to use roles in SOA/OSB stripe to grant access to ESS jobs.

Workaround

Modify the following API in ProcessorMXBean to return the deployedName and all appStripes configured for a hosting App.
public Map getEnabledApplicationsWithStripesMap()
This API now returns a Map of all applications that are enabled and currently active. The keys for this map are the logical names and the corresponding
values are the array of strings with -> 2 elements for an app without MultiStripe configured and
-> 'n+1' elements for an app configured with 'n' multiStripes.
HA SinglePatchCommand: Unrecognized SSL message for patching
Bug: 21157438
Added: 18-March-2017
Platform: All
When you use the same port for EM and SOA Composite, an error is thrown. This is a known issue.
Workaround: Change the port to 7003, which is the port for SOA_Server listed in the WebLogic Console.
Prerequisite Not Executed While Installing 12.2.1 SOA on DOCKER Container
Bug: 21846678
Added: 20-October-2015
Platform: Linux x64
During installation of SOA 12.2.1 on Docker Container, the "Checking Operating system Certification"
pre-requisite is not executed. This is expected behavior . Workaround: You can either click Skip and continue with the installation, or run the SOA installer in silent mode with flags -ignoreSysPrereqs
and -novalidation to ensure that system prerequisite checks are skipped.

Second Right-Click Does Not Work in Target Navigation Tree for Safari

  • Bug: 21965636
  • Added: 20-October-2015
  • Platform: All

The folowing steps reproduce the issue:

  • Use Safari browser to log in to Enterprise Manager Fusion Middleware Control.
  • Expand the target navigation tree on the left.
  • Navigate to the SOA, soa-infra node.
  • Right-click a composite. You get the options.
  • Now, right-click again on any node. The options do not appear. You get a blank context menu.

Workaround

You can use the SOA Infrastructure dropdown menu to access the menu options.


Technology Adapters


Flexfield Support for Open Interface Table Is Not Fully Functional in Oracle E-Business Suite Adapter

  • Bug: 21326097
  • Added: 26-June-2015
  • Platform: Windows

When creating an Open Interface Table partner link configured with flexfields for Oracle E-Business Suite Adapter, the generated schema files do not include the associated flexfield definition or reference.

Please note that this issue is specific to Open Interface Tables. Flexfield support for PLSQL APIs works as expected.

Workaround:

None.

Email Subject and Body Garbled in UMS-Outbound when it is Non_UTF-8 Server Encoding

  • Bug: 20673880
  • Added: 21-August-2015
  • Platform: All

Email subject and body get garbled in UMS-outbound (SendNotification) scenario. This happens when server encoding is Non_UTF-8. The mail corruption happens when message is String Type or Opaque or when it is using XSD.

Workaround:

Restart the server, when the server encoding is changed to be UTF-8.

XSLT Transform Mapper Activity in Oracle BPEL Process Manager

  • Bug: 19558134
  • Added: 25-May-2015
  • Platform: All

When you attempt to map a node from the source node to the target node in the XSLT Mapper for Oracle RightNow schemas, you receive the following error

Could not create element value-of

In addition, the mapped element comes out of Target Parent Node.

Workaround:

Before performing any mapping, check whether the Parent Node on the Target node is gray in color.
Select this parent node on the target (for example, Contact), then right-click and select Create Node in XSLT.
XA class support is deprecated for Oracle JCA Adapter for MQ Series
Bug: 21049656
Added: 07-May-2015
Platform: All
XA class support is deprecated for Oracle JCA Adapter for MQ series.
Do not use XA with Oracle JCA Adapter for MQ Series or Transport.


Business Activity Monitoring


Resource Catalog/Browse dialog in Dash editor doesn't use Business View display names

  • Bug: 21170770
  • Updated: 24-February-2016
  • Platform: Windows

The Resource Catalog (Browse) dialog in the Dashboard editor displays the internal Business View names, not their display names.

Workaround:

None.

Upgrade - Tables - value formatting not retained on some columns even after save

  • Bug: 21634911
  • Updated: 22-February-2016
  • Platform: All

Value formatting is not retained after editing or re-ordering columns in an already-created business view, or upon adding more columns to the business view.

Workaround:

Reset value formatting properties for the edited business view and save again

DO - Varchar size limit changes may purge existing data

  • Bug: 22111778
  • Updated: 19-February-2016
  • Platform: All

In a Logical DO, the simple DOs had a number of fields that were varchar, with a size of 2000. When changing the size of the varchar fields from 2000 to 50, the data that in one of these fields was completely lost. Changing the field size limit purges existing data.

Workaround:

None. This is expected behaviour. edge a

Flow of FOD app in DeveloperAPI suite doesn't complete.

  • Bug: 21376518
  • Updated: 06-October-2015
  • Platform: All

When using BAM 11g Adapter in 12.2.1 SOA env, and if BAM 11g environment does not exist, then we will observe stuck threads and see exceptions.

Workaround:

None

Upgrading logical DO from 11g to 12c.

  • Bug: 21376518
  • Updated: 06-October-2015
  • Platform: All

The issue at hand is that 11g default join type is 'left outer join' and not 'inner join'. In 12c, we added 'left outer join' but did not update the import program and therefore migrating is creating 12c 'default' LDO which is inner join. As a side effect, extra joins are being created during import and data is not shown in the logical DO.

Workaround

After migrating to 12c, select the 'Left Outer Join' select box in the logical DO.

Resource Catalog/Browse dialog in Dash editor doesn't use Business View display names

  • Bug: 21170770
  • Updated: 12-October-2015
  • Platform: All

If the name and display names of business views are different and if you add those business views into the dashboard using the resource catalog, then that dialog shows names of the business views whereas the mandate is to show the specified display name of business views in the UI.

Workaround

None

Messages are missing on fail over using data-based alert event

  • Bug: 21107366
  • Updated: 12-October-2015
  • Platform: All

Messages are missing on fail over using data-based alert event. This often happens during CQ Service migration from one node to another.

Workaround

Data loss depends on the CQL query. If your query condition has some limitations, during query migration, the data that populates in the time frame that the query is started on another server will be lost since it generates data with new modified time before query starts. If we don't want to lose any data we can change the condition to another meaningful one or remove the condition. We can either allow some data loss or allow some duplicated data without the data condition.

Add/Edit query filter: Columns drop-down won't expand in Microsoft Internet Explorer

  • Bug: 21069368
  • Updated: 12-October-2015
  • Platform: All

When creating a filter for a business query, I expect the Columns drop-down list to expand when I click it, but in IE10, the list does not expand.

This is a known issue for this release.

Workaround

None. This is an ADF-based issue.

Action List with too many editable fields causes error

  • Bug: 20730311
  • Updated: 13-October-2015
  • Platform: All

If I have an Action list with editable fields, if there are too many fields that are editable, I will get the following error upon clicking in one of the editable fields.

A connection to the server has failed.(status = 500)

Workaround

None. This is an ADF-based issue.

DO Param saved with default DisplayName inserted into expression with brackets

  • Bug: 20532152
  • Updated: 13-October-2015
  • Platform: All

A parameter saved with the default Display Name is inserted into the expression with brackets {}.

None. This is expected behavior. Without the brackets, any special characters in display names such as a dash ( - ) would be treated as a value (in this case a - a minus sign).

Fix warning for DO row edits when changing pages -- does not always appear

  • Bug: 20173781
  • Updated: 13-October-2015
  • Platform: All

A parameter saved with the default Display Name is inserted into the expression with brackets {}.

Workaround

None. This is expected behavior. Without the brackets, any special characters in display names such as a dash ( - ) would be treated as a value (in this case a - a minus sign).

Properties and Runtime interaction tabs not seen for Scatter and Bubble views

  • Bug: 20076550
  • Updated: 14-October-2015
  • Platform: All

The Properties and Runtime Interaction tabs are missing for Scatter and Bubble views.

Workaround

Resize your browser window for a moment, and then go back your original size to get the '>>' icon to appear. This issue usually occurs with Internet Explorer or Firefox. Switching to a different browser might also fix the problem.


Business Process Management


Option to define "After Task Completion" behavior does not appear in the Oracle BPM Workspace 12.2.1

  • Bug: 24593329
  • Added: 05-SEP-2016
  • Platform: Windows

You can define "After Tasks Completion" behavior in BPM Workspace as discussed in How to Specify After Tasks Completion Behavior of Task Details Pop Up Window. However, this option to define the After Task Completion behavior does not appear in the version 12.2.1.0.0.

Workaround:

For the option to appear, apply the patch 22087208

Approve and reject links in email notification do not work in Microsoft Outlook

  • Bug: 18858843
  • Updated: 20-April-2016
  • Platform: All

Approve and Reject links in Thai, Arabic and Korean languages do not open in email notification when you use Microsoft Outlook as email client. This is limitation of Microsoft Outlook. This happens when email body length exceeds the max limitation in Microsoft Outlook.

This issue might also occur in other languages too if the task title in email subject is too long.

Workaround:

Use Mozilla Thunderbird email client to open tasks with long subject line.

Safari Browser - Links tab missing and Decision Properties Overlaps with Actions

  • Bug: 21964518
  • Updated: 12-October-2015
  • Platform: All

When using Safari browser, the Advanced tab (Rules) in SOA Composer does not display the Links tab and the Decision Properties tab overlaps with Action menu.

Workaround:

None

[Iphone 6 generic ]- Worklist / SOA Composer - App crashes on launching

  • Bug: 21764484
  • Updated: 06-September-2015
  • Platform: All

The Worklist application crashes on launch on iPhone 6 plus while using the Safari browser.

Workaround:

Upgrade to a newer IOS version (for example, 8.4.1).

Human Task form data field not editable

  • Bug: 18998302
  • Updated: 16-June-2014
  • Platform: All

In BPM Studio, human task input in a case activity becomes editable when it is selected as a user input. It is not editable when the human task input is selected as a case data type.

Workaround

None

Migration of PS6 projects to 12.2.1 throws rules upgrade exception in Studio

  • Bug: 21804781
  • Updated: 01-SEP-2015
  • Platform: All

Migrating a project from BPM PS6 to BPM 12.2.1 might cause an error from the Rules SDK similar to the following:

oracle.xml.parser.schema.XSDException: /myjdev/mywork/myproject/../../../xsd/BusinessCaseRuleFact.xsd (No such file or directory)
at oracle.xml.parser.schema.XSDBuilder.build(XSDBuilder.java:621)

This can occur if the stored schema path is expressed relative to the rules directory, not the project directory.

Workaround

Before migrating the PS6 project, edit the problematic .rules file and replace all occurrences of '/../../../xsd/BusinessCaseRuleF act.xsd' with 'xsd/BusinessCaseRuleFact.xsd'. Save the .rules file. Open the BPM PS6 project in BPM Studio 12.2.1 to migrate it.

BPM TASK LIST CONTENTS ARE NOT SHOWING IN PROCESSPORTAL (must use Oracle HTTP Server for WebCenter-BPM integration)

  • Bug: 21772408
  • Updated: 01-SEP-2015
  • Platform: All

If Oracle HTTP Server is not used, the integration of BPM Task information into Oracle Webcenter does not work.

Workaround

In order for the integration of Oracle BPM into Oracle WebCenter to work, you must use Oracle HTTP Server on both installations.

READINESS CHECK FAILED FOR SOA SCHEMA AT TEST_REQUIRED_PROCEDURES

  • Bug: 21762288
  • Updated: 31-AUG-2015
  • Platform: All

When migrating from BPM PS3 to BPM PS5 (as part of an upgrade path to BPM 12.2.1) you may receive an error that indicates 'Readiness check failed for soa schema'

Workaround

This error message can be safely ignored. The upgrade scripts to target BPM 12.2.1 resolve the issue.

BPM PS6->12.2.1 Upg:InvalidMagicMimeEntryException: Invalid Magic Mime Entry

  • Bug: 21535358
  • Updated: 30-JUL-2015
  • Platform: All

When upgrading from BPM PS6 to BPM 12.2.1, you might see an error message similar to 'eu.medsea.mimeutil.detector.InvalidMagicMimeEntryException: Invalid Magic Mime Entry' on server startup.

Workaround

This error message can be safely ignored.

STRESS BPM SQ:tuning-increase AuditEventBatchSize50 from 10

  • Bug: 21444641
  • Updated: 14-JUL-2015
  • Platform: All

You might see an error message similar to the following indicating a long block chain at oracle.bpm.bpmn.engine.service.BPMNAuditTrail.logEvent:

"[ACTIVE] ExecuteThread: '223' for queue: 'weblogic.kernel.Default (self-tuning)'" #2239 daemon prio=5 os_prio=0 tid=0x000000001d631000 nid=0x5da6 waiting for monitor entry [0x000000005ff00000] java.lang.Thread.State: BLOCKED (on object monitor) at oracle.bpm.bpmn.engine.service.audit.persistence.AuditKeyGenerator.nextId(AuditKeyGenerator.java:116) - waiting to lock <0x000000072f41ad70> (a java.lang.Object) at oracle.bpm.bpmn.engine.service.BPMNAuditTrail.logEvent(BPMNAuditTrail.java:125)

Increase the AuditEventBatchSize from 10 to 50. In Enterprise Manager, AuditEventBatchSize can be found under 'SOA Administration / BPMN Properties'.

PENDING JDEV CHANGES FOR SUMMARY FIELD MAPPING FEATURE

  • Bug: 17039859
  • Updated: 01-JUL-2013
  • Platform: All

When creating a summary field mapping, when the Insert Into Expression button is clicked, BPM Studio does not warn if a payload expression is not present in the xpath expression.

Workaround

None

WEB FORMS DO NOT SUPPORT DIGITAL SIGNATURES

  • Bug: 16992692
  • Updated: 20-JUN-2013
  • Platform: All

Web forms do not support digital signatures.

Workaround

None

FUTURE PARTICIPANT EDITING WON'T WORK ON THE DUPLICATE PARTICIPANT

  • Bug: 13990543
  • Updated: 23-APR-2012
  • Platform: All

When a substitution rule is used in a task, the resulting participant list might include duplicate participants that can't be edited.

Workaround

None

Facts Tab - Error 500 thrown on adding a property (Fact tab in BP Composer is now view-only)

  • Bug: 21617657
  • Updated: 11-AUG-2015
  • Platform: All

The Facts tab in BP Composer is now view-only. Some controls to modify existing facts still appear on the GUI. Using them causes errors.

Workaround

Use BPM Studio to modify facts.

GROUP STILL EXISTS AFTER DELETION IN WORKSPACE

  • Bug: 21470654
  • Updated: 20-JUL-2015
  • Platform: All

After deleting an approval group in Process Workspace, the approval group still exists.

Workaround

Repeat the deletion. The approval group is deleted on the second attempt.

JDEV-RULEDITOR-VERBAL RULES Fact usage not honoured

  • Bug: 21310445
  • Updated: 24-JUN-2015
  • Platform: All

Verbal Rules will work only with auto-generated business phrases. When creating verbal rules that include other business phrases used for input/output, error messages in Studio are generated indicating that the business phrases are unreferenced, even though they have been declared. For example, you might see an error message simlar to:

RUL-05163: The fact type "myFact" is not used in any ruleset called by this decision function.

RUL-05164: The fact type "myFact2" is referenced, but is not asserted nor input.

Workaround

Repeat the deletion. The approval group is deleted on the second attempt.

Can't do failover in bpm composer

  • Bug: 20811940
  • Updated: 01-APR-2015
  • Platform: All

BP Composer does not support high availability reliably. When working in BP Composer, if the BPM server becomes unavailable, BP Composer does not deal with the situation by routing requests to other server nodes. Various errors can occur in this scenario.

Workaround

None


Business-to-Business Integration (B2B)


B2B does not use configured element delimiter in translated hl7 native message

  • Bug: 20408793
  • Updated: 27 September 2015
  • Platform: Windows

Please apply Patch 20408793 after B2B/healthcare installation to obtain XEngine related bug fixes. Patch 20408793 is platform specific. Instructions on how to apply this patch are in the README of patch 20408793.

Workaround:

None


Enterprise Scheduling Service


In an environment where SSL only is enabled, ESS cluster healthcheck page may not work as expected

  • Bug: 21886292
  • Updated: 15-September-2015
  • Platform: Windows

In an environment where 'SSL only' is enabled, ESS cluster healthcheck page may not work as expected. This is a known issue which will be addressed in the 12.2.2 release. However, ESS healthcheck for individual servers will work fine in these scenarios.

Workaround:

None


Service Bus


What's New in 12c

For information on new features in Oracle Service Bus 12c, see the following:

What's New in This Guide in Developing Services with Oracle Service Bus

What's New in This Guide in Administering Oracle Service Bus

Accessing the Service Bus Console from http://example.com:port/sbconsole is Deprecated

  • Bug: 21617828
  • Updated: 10-October-2015
  • Platform: All

Accessing the Service Bus Console from http://example.com:port/sbconsole is deprecated. Instead, access the Service Bus Console using the following URL:

http://example.com:port/servicebus

IBM WebSphere MQ Version 7.5 Now Certified with This Release of Oracle Service Bus

  • Bug: no bug number
  • Updated: 16-July-2015
  • Platform: Windows

IBM WebSphere MQ Version 7.5 has been certified for use with Oracle Service Bus 12c Release 12.2.1.

Oracle Service Bus Dashboard Page is Missing in Fusion Middleware Control

  • Bug: 20136341
  • Updated: 23-February-2015
  • Platform: All

In some instances, The Oracle Service Bus Dashboard page may be missing from Fusion Middleware control. If the service-bus node is missing from the SOA folder in the Target Navigation pane, perform the following steps to display it in Fusion Middleware Control:

In the server staging location, the em.war application's weblogic.xml file should have the following entry:
osb.em
If the entry is not present, add it to the weblogic.xml file.
Restart the servers. After the servers are restarted the Oracle Service Bus
dashboard page should now appear in Fusion Middleware Control.

Resequencer Only Applies to Messages in the Pipeline

  • Bug: 19004690
  • Added: 17-June-2014
  • Platform: All

The behavior of the resequencer is not fully documented with regard to business services. When you configure a resequencer for a pipeline, only the messages coming into the pipeline are sequenced per the resequencing rules. The pipeline then forwards the messages, in order, to its target business service. However, it is possible the business service will process the messages in a different order. The resequencer does not guarantee the order messages are processed by the business service.

WSDL Errors Can Occur When Oracle Service Bus and Oracle Business Process Manager Are on the Same Domain

  • Bug: 18856204
  • Added: 10-June-2014
  • Platform: All

If Oracle Service Bus is installed on the same domain as Oracle Business Process Manager (BPM), the following error can occur when creating a WSDL file:

The WSDL is not semantically valid: Failed to read wsdl file from url due to -- java.net.MalformedURLException: Unknown protocol: servicebus.

Workaround:

Contact Oracle Support for a solution to this issue.

Changes Do not Show up on Fusion Middleware Control with Discovery Cache Enabled

  • Bug: 18774509
  • Added: 10-June-2014
  • Platform: All

When the Discovery Cache is enabled in Fusion Middleware Control, all target information is cached upon first login. Consequent logins retrieve the targets from the cache, which means changes, additions, and deletions to targets are not accurately reflected in the navigation tree.

Workaround

Either disable the Discovery Cache, or refresh the WebLogic server domain to update the target information.

Using Only Simple Element Types Causes Parsing Errors in REST Binding

  • Bug: 18695606
  • Added: 10-June-2014
  • Platform: All

Using only a simple element type (as you might do in testing) causes the parsing of the underlying object in a REST binding proxy service creation to fail.

This can occur if the stored schema path is expressed relative to the rules directory, not the project directory.

Workaround

Data types must be wrapped in a complex type. For example:



   
      
         
      
   

OWSM Policies Are Not Applied When Activating a JCA Business Service

  • Bug: 18562166
  • Added: 10-June-2014
  • Platform: All

When you activate a project or deploy a configuration JAR file that includes a JCA business service with a PII policy attached, the policy is not applied until the server is restarted.

Workaround

Restart the server after importing or updating the JCA business service.

Updating a JCA Proxy or Business Service's Policy Mode Causes a Runtime Error

  • Bug: 18562207
  • Added: 10-June-2014
  • Platform: All

If you change the configuration of a JCA proxy or business service from OWSM Policy Store to No Policies, or vice versa, the JCA service generates an exception at runtime. For JCA business services with an OWSM policy attached, any change to the security or endpoint properties results in a runtime exception.

Workaround

For proxy services exhibiting this issue, modify one of the JCA endpoint properties and redeploy. The properties that can be changed depend on the specific JCA adapter used. For business services exhibiting this issue, restart the server after updating the properties.

WSDL Generated Missing Elements for Custom Exception

  • Bug: 18502287
  • Added: 17-June-2014
  • Platform: All

If an EJB contains custom exceptions and you export the WSDL file from the EJB business service, the generated WSDL file does not include the custom exception properties. The issue is limited only to the WSDL generation part of the file. During runtime, the custom exception thrown from the EJB is mapped to the respective elements in the SOAP fault. The elements in the response payload are populated corresponding to the properties of the custom exception.

Workaround

Manually edit the WSDL file to include these custom exception properties.

The Policy Violation Count Does Not Include HTTP Token Policy Violations

  • Bug: 18477417
  • Added: 10-June-2014
  • Platform: All

The policy violation metrics displayed on a service's Policies page in Fusion Middleware Control do not count http_*_token policy violations. This happens for HTTP token-based authentication policies only.

Connection to Application Server is Refused During UMS Adapter Configuration

  • Bug: 18424427
  • Added: 10-June-2014
  • Platform: All

When configuring the JCA Adapter for UMS, a server exception occurs and the connection is refused on the Identity Lookup for the application server.

Workaround

This issue is only observed from a server created outside of JDeveloper. Use an integrated server created within JDeveloper instead. You can also enter the user account manually.

Default Logging Format Does Not Log Authorized User

  • Bug: 18417685
  • Added: 10-June-2014
  • Platform: All

The default extended logging format does not include the name of the authorized user in the access log, which is required when using HTTP basic authorization.

Workaround

Change the logging format to Common. To access the log properties, launch the WebLogic Server Administration Console and navigate to Environment > Servers > server_name > Logging > HTTP (Advanced section) > Format. A server restart is required after this change.

Automatic Policy Configuration for STS May Result in an Incompatible Policy

  • Bug: 18384215, 18416081
  • Added: 10-June-2014
  • Platform: All

Using the automatic policy configuration for STS might not result in a compatible policy for a web service client.

Workaround

For information about this and other web services policy issues, see Web Services in the Release Notes for Oracle Fusion Middleware.

Global Policy Inheritance Rule Does Not Work For Authorization Policies

  • Bug: 18295336
  • Added: 10-June-2014
  • Platform: All

Authorization policies in the global policy set cannot be disabled by locally attaching the oracle/no_authorization_service_policy to the proxy service, where global policy is applied.

Workaround

To disable globally attached authorization policies, select No Policy in the proxy service security policy configuration in either the Oracle Service Bus Console or JDeveloper.

History Changes View Does Not Reflect All Proxy Services

  • Bug: 18058244
  • Added: 10-June-2014
  • Platform: All

In the Oracle Service Bus Console, the Changes sub-tab on the Histories tab does not reflect a newly created proxy service that was generated when creating a pipeline by selecting Expose as a Proxy Service.

Workaround

Refresh the window to show the new proxy service.

Pipeline Info Buttons on the Service Type Page Do Not Work When Launched from the Keyboard

  • Bug: 17959710
  • Added: 10-June-2014
  • Platform: All

With screen reader accessibility enabled, the Pipeline Service Type page has Info buttons next to the service type options. When you use the keyboard to navigate to and launch the Info windows, the window appears and then closes immediately. When you click an Info button with the mouse, the Info window remains open until you press Esc .

Workaround

Click the Info button with the mouse.

Import Config JAR Dialog Has Incorrect Default Focus

  • Bug: 17950304
  • Added: 10-June-2014
  • Platform: All

The Import Config JAR dialog defaults the focus to the "Input Source" flow diagram at the top of the dialog, which does not include active objects

Workaround

Press the Tab key to bring the focus to the Browse button for the configuration JAR input.

KSS Keystore Might Require Updates After Import

  • Bug: 17799824
  • Added: 10-June-2014
  • Platform: All

The default keystore for Oracle Web Services Manager (OWSM) is now KSS. Any imported resources that previously used the JKS keystore might require either updates to the KSS settings or overriding the setting to use KSS. KSS keystore does not use passwords and does not require credential store, so these override properties might not be correct in the imported project.

Workaround

If you have overrides configured for keystore.sig.csf.key and keystore.enc.csf.key to csf-key, you need to change their values to give the direct keystore alias instead of csf-key. If you use default values (that is, no overrides), you can skip this step.

Updating REST Binding Fault Schema Results in Error

  • Bug: 17625172
  • Added: 10-June-2014
  • Platform: All

When you create a REST business service and add the fault binding for the response, you should not be able to change the schema type of the fault. However, the REST Binding wizard allows this. Editing the schema here returns a WADL validation error.

Workaround

Do not change the default schema.

Admin Server Log Contains an aia-log-debug-handler Error Post-Upgrade

  • Bug: 17381835
  • Added: 10-June-2014
  • Platform: All

The following message appears only once during the first re-start of the Admin server after you upgrade to 12c. The error does not indicate any issues with the AIA logging functionality and can be ignored.

The handler 'aia-log-debug-handler' is already defined in the target logging @ configuration document.


Null Pointer Exceptions Occur When Projects and Resources in JDeveloper Are in the Wrong Location

  • Bug: 17299328
  • Added: 10-June-2014
  • Platform: All

When you create projects and resources in a folder location outside of the folder of the containing application, null pointer exceptions occur when you perform certain tasks, such as importing resources. Creating resources outside of the application folder is not currently supported.

Workaround

Make sure to create all projects, folders, and resources under the directory of the containing application.

Runtime Validation for content-type is More Strict

  • Bug: 17187825
  • Added: 10-June-2014
  • Platform: All

In 12c, content-type validation is more strict than in previous versions. Therefore, requests with bad content-type that were previously allowed are no longer allowed in the current version. One example is sending a SOAP message with the content-type set to text/plain. This request would now be rejected.

Workaround

Configure the correct content-type for 12c services.

Creating a Global Policy Generates a Design-Time Conflict

  • Bug: 16894883
  • Added: 10-June-2014
  • Platform: All

If a proxy or business service is configured for OWSM policies and matches the configuration of an existing global policy set but does not support all of the policies in the global policy set, the following error occurs when you update the service in a new session, and a conflict is generated.

[OSB-398128]One of the Web service policy attached is not allowed on service path/resource_name because the service binding does not support it.

Workaround

When you create a global policy set with policies that are not supported by non-SOAP services, make sure that the policy set is configured specifically enough to exclude any non-SOAP services.

Log4j is no Longer Added to the System Classpath

  • Bug: 16729226
  • Added: 10-June-2014
  • Platform: All

Oracle Service Bus no longer adds log4j.jar to the system classpath automatically. If any of your applications or Java callouts depend on log4j being in the system classpath, add it to the classpath manually.

Moving a Service Causes Duplication in Policy Usage Analysis

  • Bug: 16711022
  • Added: 10-June-2014
  • Platform: All

If you move a proxy or business service with locally attached OWSM policies from one folder to another or from one project to another, the service is listed twice on the WebLogic Domain > WSM Policy > Usage Analysis page in Fusion Middleware Control. The expected behavior is that the service no longer appears in its original location on the usage analysis page, but it appears under both the original and the new locations.

Renaming a Service Causes Duplication in Policy Usage Analysis

  • Bug: 16704604
  • Added: 10-June-2014
  • Platform: All

If you rename a SOAP-based proxy or business service with locally attached OWSM policies, the service is listed twice on the WebLogic Domain > WSM Policy > Usage Analysis page in Fusion Middleware Control, once under its original name and once under its new name. The expected behavior is that the service no longer appears with its original name on the usage analysis page.

Deleted Services Still Appear in Policy Usage Analysis

  • Bug: 16704586
  • Added: 10-June-2014
  • Platform: All

If you delete a SOAP-based proxy or business service with locally attached OWSM policies, the service continues to appear on the WebLogic Domain > WSM Policy > Usage Analysis page in Fusion Middleware Control. The expected behavior is that the service no longer appears on the usage analysis page.

Unable to Create Durable Subscriptions on AQ Topics

  • Bug: 14321803
  • Added: 10-June-2014
  • Platform: All

On Oracle AQ, the subscriber name must be between 1 and 30 bytes long. Oracle Service Bus uses an MDB for JMS subscriptions, and generates the MDB name using the format "RequestEJB"+GUID. The GUID alone can be longer than 30 characters. When the MDB name is longer than 30 characters, an error similar to the following is logged:

RequestEJB5599235263010875116XX3906baed.137fe8c76cf.X7fc6 is unable to connect to the JMS destination:
jms.fromSys1toSys2. The Error was: oracle.jms.AQjmsException: JMS-230: Illegal operation on durable subscription with active TopicSubscriber

Assign Activity Produces an Empty Message Body if a New Line Character Exists

  • Bug: 16205232
  • Added: 07-March-2013
  • Platform: All

The resulting message body is empty if a proxy service processes an XML string with a newline character after the XML declaration (), the message flow includes an Assign activity that uses fn-bea:inlinedXML() to assign that XML string to a variable, and the message flow uses the variable to replace the contents of $body. This is related to issue 13902323, caused by the fn-bea:inlinedxml() function trimming whitespace when parsing text to XML.

Pass-Through Pipeline Might Marshall and Unmarshall Message Body

  • Bug: 15908859
  • Added: 29-January-2013
  • Platform: All

If the message flow in a pipeline is configured to simply pass messages through with no modification, the messages are still parsed by the XML parser if the message content is accessed and read during processing. This means the original text is materialized into XMLBeans and then back into XML. For example, the message might be processed in a pass-through scenario, but be accessed and read by a logger during processing. In this case, the message is materialized into XMLBeans. When the message body includes decimal or hexadecimal representation of characters, the XML parser resolves these representations to the actual characters. Note that only the character representation of the information is changed and not the information itself.

Workaround:

If there are character representations in the message content and you want to preserve the original character representations instead of converting them to the actual characters, either make sure the message body is unread during processing to ensure it is truly pass-through, or specify the representations inside a CDATA section before the message is received by Service Bus. For example:



   

JCA Polling Proxy Services Generate Stuck Thread Errors

  • Bug: 14831608
  • Added: 29-January-2013
  • Platform: All

Proxy services configured for JCA polling adapters (such as the JCA File Adapter, DB Adapter, AQ Adapter, and so on) might generate errors in the server log indicating that the thread is stuck, which causes the server's health status to change to WARNING. The errors are caused by the single listening thread created in the default Work Manager. This thread is non-terminating so the services can continue to poll for new events, but this means the thread necessarily remains busy longer than the configured maximum time. The error message indicates that the thread has been busy for "xx" seconds working on the request, which is more than the configured time (StuckThreadMaxTime) of "xx" seconds. The errors can be ignored, but might also fill up the server log.

Workaround:

Because these error messages do not indicate a true problem, they can be suppressed. To do this, create a new Work Manager with the following setting:

true

Then modify the service's Dispatch Policy in the transport settings to use the new Work Manager. Any new stuck thread messages are written to the new Work Manager instead of cluttering up the default Work Manager.

Harmless Warning for JEJB 3.0 Proxy Service

  • Bug: 14321779
  • Added: 29-January-2013
  • Platform: All

After a JEJB 3.0 proxy service is created and activated, the following warning message appears in the managed server logs:

  <"config-root"///plan was not found>

This warning is harmless and can be ignored.

JCA Polling Proxy Services Generate Stuck Thread Errors

  • Bug: 14831608
  • Added: 29-January-2013
  • Platform: All

Proxy services configured for JCA polling adapters (such as the JCA File Adapter, DB Adapter, AQ Adapter, and so on) might generate errors in the server log indicating that the thread is stuck, which causes the server's health status to change to WARNING. The errors are caused by the single listening thread created in the default Work Manager. This thread is non-terminating so the services can continue to poll for new events, but this means the thread necessarily remains busy longer than the configured maximum time. The error message indicates that the thread has been busy for "xx" seconds working on the request, which is more than the configured time (StuckThreadMaxTime) of "xx" seconds. The errors can be ignored, but might also fill up the server log.

Workaround:

Because these error messages do not indicate a true problem, they can be suppressed. To do this, create a new Work Manager with the following setting:

true

Then modify the service's Dispatch Policy in the transport settings to use the new Work Manager. Any new stuck thread messages are written to the new Work Manager instead of cluttering up the default Work Manager.

HTTP Business Service Connection Timeout Not Honored on Windows 2008 64-bit

  • Bug: 13097895
  • Updated: 13-January-2012
  • Platform: All

Connection timeouts set on HTTP business services ("Connection Timeout" property in the HTTP transport configuration) are not honored in Windows 2008 64-bit environments.

HTTP Business Service Connection Timeout Not Honored on Windows 2008 64-bit

  • Bug: 13097895
  • Updated: 13-January-2012
  • Platform: All

Connection timeouts set on HTTP business services ("Connection Timeout" property in the HTTP transport configuration) are not honored in Windows 2008 64-bit environments.

Starting Managed Server with IPv6 Host and Port Throws BEA-000000 Error

  • Bug: 10029204
  • Updated: 14-January-2011
  • Platform: All

In the development environment, when starting a managed server by passing the managed server's IPv6 address and port, Oracle Service Bus throws a BEA-000000 "Cannot contact Admin server" error.

Workaround:

Create a new key-pair, then reconfigure SSL on the Admin server so that the key-pair in the server-side SSL certificate uses the host name of your IPv6 managed server. Then restart your managed server. For more information on SSL, see Configuring SSL in Administering Security for Oracle WebLogic Server.

Password Field Not Editable in Domain Template on Linux

  • Bug: 10394900
  • Updated: 14-January-2011
  • Platform: Linux

On Linux, when creating a new domain using a developer template in the Oracle Fusion Middleware Configuration Wizard, the password and confirm password fields are not editable, and you cannot enter a password to create a domain.

Workaround:

To work around the issue each time it occurs:

Click the close window "x" button in the upper right corner of the Configuration Wizard. In the confirmation dialog that appears, click "No" to return to the Configuration Wizard. Now you can enter the password and confirmation password.

To fix the issue permanently:

Kill a scim processes. For example, kill "pgrep scim"
Modify (or create) the file ~/.scim/config to include the following line:
/FrontEnd/X11/Dynamic = true (follow case sensitivity). If you are running VNC, restart the VNC server.
Run the Configuration Wizard again.

MQ Transport in Binding Mode Throws Error on Linux

  • Bug: 10020344
  • Updated: 14-January-2011
  • Platform: Linux

When Oracle Service Bus and IBM WebSphere MQ are installed on the same Linux system, configuring the Oracle Service Bus MQ transport to run in binding mode causes an error.

Workaround:

Make sure Oracle Service Bus and IBM WebSphere MQ are installed using the same user credentials.

Click the close window "x" button in the upper right corner of the Configuration Wizard. In the confirmation dialog that appears, click "No" to return to the Configuration Wizard. Now you can enter the password and confirmation password.

Add the MQ Java library path to the Oracle Service Bus DOMAIN_HOME/bin/setDomainEnv.sh file. For example:
-Djava.library.path=/opt/mqm/java/lib
Oracle JRockit: When using Oracle JRockit on Linux, add the LD_PRELOAD property to setDomainEnv.sh as well. For example: LD_PRELOAD=${JAVA_HOME}/jre/lib/i386/libjsig.so

Resources Disappear When Server Encoding Changes

  • Bug: 9488927
  • Updated: 14-January-2011
  • Platform: Linux

After creating resources in the Oracle Service Bus Console, then changing the server encoding and restarting the server with the different encoding, only English language resources (if any) appear in the Oracle Service Bus Console. Non-English resources do not appear. For example, if you start the server using ja_JP.utf8 encoding and create resources in the Oracle Service Bus Console with Japanese, Korean, and English names, then restart the server using a different encoding such as ko_KR.euckr, only the English-name resources appear in the Oracle Service Bus Console. In this situation, the missing resources still exist but are not displayed because of the encoding change.

Workaround:

Switch back to the original server encoding and restart the server. Oracle Service Bus does not support server encoding changes.


Healthcare Integration


B2B does not use configured element delimiter in translated hl7 native message

  • Bug: 20408793
  • Updated: 27 September 2015
  • Platform: All

Please apply Patch 20408793 after B2B/healthcare installation to obtain XEngine related bug fixes. Patch 20408793 is platform specific. Instructions on how to apply this patch are in the README of patch 20408793.

Workaround:

None


Managed File Transfer


Multibyte file name is garbled in transfer instance page

  • Bug: 24611740
  • Updated: 12-September-2016
  • Platform: All

Multibyte file name is garbled in the transfer instance page.

To avoid this, change the server encoding configuration to UTF-8 from the MFT user interface while designing the connection to SFTP Source/Target. Use Linux "locale charmap" to know local characterset and set the same in MFT.

MFT Upgrade: Cannot update properties in EM console MFT Config page

  • Bug: 23152243/23517580
  • Updated: 03-June-2016
  • Platform: All

After you upgrade from 11g to 12.2.1.1 or 12.1.3 to 12.2.1.1 and try to modify MDS related properties like (Time delay between commits) in MFT config page in Enterprise Manager, you may run into an exception in the MFT server log as below:

javax.management.remote.rmi.RMIConnectionImpl.invoke

(Ljavax.management.ObjectName;Ljava.lang.String;Ljava.rmi.MarshalledObject;
[Ljava.lang.String;Ljavax.security.auth.Subject;)
 javax.management.RuntimeMBeanException:
javax.management.RuntimeMBeanException: Property not found for the key:
mdsCacheSize

Workaround:

Use the system MBean browser in EM console to modify/update mdsCacheTTL and mdsNonCoreCacheSizeRatio properties.

Importing SSL certificates for UCM/WebCenter endpoints

  • Bug: 22588159
  • Updated: 08-February-2016
  • Platform: All

MFT Keystore Service is not used for UCM endpoint certificates

Workaround:

To connect WebCenter source/targets via HTTPS URL, import the WebCenter endpoints certificate into the Demo Trust and JDK keystorestores. Restart the server after importing the certificates.

The default location of the keystores are:
JDK - $JAVA_HOME/jre/lib/security/cacerts
Demo Trust - $WL_HOME/server/lib/DemoTrust.jks
For more information, see Using the Demonstration Keystores in the Administering Security for Oracle WebLogic Server guide.

MFT Remote sFTP endpoint port missing in config plan file

  • Bug: 22594150
  • Updated: 08-February-2016
  • Platform: All

When sFTP source and target are deployed, the config plan file mftconfig.xml is missing the port on download.

Workaround:

After downloading the config plan file, use global search and replace to update the new value.

Backward compatibility of SOAP services

  • Bug: 20476899
  • Updated: 12-January-2016
  • Platform: All

When upgrading from 12.1.3 to 12.2.1 release, if users have MFT specific projects created in older versions using JDev, specifically when MFT is the target for SOA composite (not when it is a source for SOA), users need to correct the WSDL definition of existing SOA/SOAP projects by opening them via JDev and redeploying the composite.

To run old composites without upgrading to 12.2.1, set the system property "oracle.xdkjava.compatibility.version=11.1.1".

Workaround

Swap the parts in the WSDL definition, first part is the payload and all remaining parts are headers.

For example,




FIPS Compliance Restrictions in MFT 12.2.1

  • Bug: 21973718
  • Updated: 09-October-2015
  • Platform: All

The following are the new FIPS Mode compliance restrictions in MFT 12.2.1:

  • In FIPS mode, MFT now requires 2048-bit or above SSL key sizes for Embedded FTPS SSL Server configuration and startup in FIPS Mode.
  • In FIPS mode, MFT now requires 2048-bit or above PGP keys sizes for PGP encryption and decryption.

FIPS Mode Known Issues in MFT 12.2.1

  • Bug: 21975153
  • Updated: 09-October-2015
  • Platform: All

The following are FIPS Mode known issues in MFT 12.2.1:

  • The MFT Embedded (SSH) SFTP Server does not support FIPS Mode. However, the MFT Embedded FTP/FTPS Servers support FIPS Mode. Also refer to related release note "FIPS Compliance Restrictions in MFT 12.2.1" (Bug21973718).
  • The MFT WLST command generateKeys is not supported in FIPS Mode.

Oracle API Manager


Native REST Proxy Services Are Not Publishable as APIs

  • Bug: 20877715
  • Added: 19-October-2015
  • Platform: All

Native REST services (both typed and untyped) created in Service Bus 12.2.1 are not supported. Only 12.1.3-style WSDL-based REST services are publishable as APIs. Refer to How to Create WSDL-Based REST Services for Service Bus Using JDeveloper in Developing Services with Oracle Service Bus to create this type of REST service in Service Bus 12.2.1.

This is a known issue for this release.


Oracle Stream Analytics


BigDecimal and BigInteger are not supported in JMS Outbound Adapter

  • Bug ID: 23184640
  • Added: 28-April-2016
  • Platform: All

Currently only objectified primitive object types (Integer, Double, Long, Float, Short), String objects, and byte arrays are supported in outbound JMS adapter. BigDecimal and BigInteger are not supported.

This is a known issue for this release.

Workaround:

None

Draft Explorations Break when you change the Source

  • Bug ID: 23072674
  • Added: 26-April-2016
  • Platform: All

The draft exploration breaks when you modify/delete the source or unpublish/republish the exploration.

Workaround:

None

Related Topics Link in the Topics "Configure a Target" and "Topology Viewer" do not work

  • Bug ID: 23120202
  • Added: 19-April-2016
  • Platform: All

The Related Topics links in the help topics "Configure a Target" and "Topology Viewer" that take you to the corresponding topics in the User Guide do not work. A fix will be provided soon for this.

Workaround:

None

Second Whitespace Disappears in the Business Rules Action Editor

  • Bug ID: 22675944
  • Added: 18-April-2016
  • Platform: All

When you specify strings in the business rules action editor with more than one whitespace, the second whitespace disappears. For example, if you specify "happy spring festival", it becomes "happy springfestival". This happens only while creating the business rule. When you edit the same business rule, it appears fine. This happens only on Firefox.

Workaround:

None

Search Result Location not correct in the Map Initially

  • Bug ID: 22863116
  • Added: 18-April-2016
  • Platform: All

Initially, when you search for a location on the map, the search result is not correct.

Workaround:

Zoom in or zoom out and you can see the searched location on the map.

Sort By Name Does not Work as expected on Chrome Browser for Patterns

  • Bug ID: 22863116
  • Added: 18-April-2016
  • Platform: All

On a Chrome Browser, when you try to sort the patterns in the Patterns page by name, the result of the sort operation is not as expected.

Workaround:

None

JMS Exception Encountered for Big Decimal Data Type

  • Bug ID: 23065732
  • Added: 18-April-2016
  • Platform: All

JMS Exception is encountered when you try to send an event with Big Decimal Data Type in the stream that uses a CSV source.

Workaround:

None

No Real Output in Target CSV File on Standard Deviation Pattern for int Input Property

  • Bug ID: 23065745
  • Added: 18-April-2016
  • Platform: All

When you configure a CSV target for a Standard Deviation Pattern which creates mapping between the standard deviation result on int property and a double property in target, and publish the pattern, there is no real content outputted to the target CSV file and an exception like <> is thrown repeatedly on the Oracle Stream Analytics console.

Workaround:

None

Deploying an Exported File Fails for Business Rules based on Aggregation

  • Bug ID: 23065797
  • Added: 18-April-2016
  • Platform: All

When you export an exploration that contains business rules based on aggregation and try to deploy it in an instance, the deployment fails.

Workaround:

None

Exception Thrown when you Export a Kafka Application that contains Oracle Business Rules

  • Bug ID: 23065848
  • Added: 18-April-2016
  • Platform: All

When you export an exploration based on a Kafka source and configured with Oracle Business Rules, an exception like oracle.rules.rl.RuleSession logIt SEVERE: UndefinedException: The symbol "oracle.wlevs.strex.generated.sx_4_14_Exploration1Sean_1" is undefined is thrown.

Workaround:

None

Spatial Pattern does not retain all statuses in the Live Output

  • Bug ID: 23065863
  • Added: 18-April-2016
  • Platform: All

In an exploration based on a spatial pattern using a CSV file as a stream source, if the CSV stream has the Loop the CSV File option selected, the Live Output Stream shows ENTER, EXIT, NEAR and STAY initially, on the first loop of the CSV file. Later on, on consecutive loops of the CSV file, only STAY status is retained in the output.

Workaround:

None

Exception Thrown when Deploying Exported Twitter Exploration that contains Oracle Business Rules

  • Bug ID: 23126540
  • Added: 18-April-2016
  • Platform: All

When you export an exploration which is based on Twitter source and configured with Oracle Business Rules, try to deploy the JAR to an Oracle Stream Analytics Server instance, the deployment fails with Failed to set property 'name' in output event exception.

Workaround:

None

Cannot find the Summary Field using Double Property in Event Mapping

  • Bug ID: 23022885
  • Added: 18-April-2016
  • Platform: All

When you use aggregate functions like max, min, avg on long, double, timestamp type, you may not be able to find the corresponding summary fields in the target event mapping dialog.

Workaround:

Create the target before the exploration.

Configure the target in the exploration and save the exploration. Re-open the Configure Target dialog in the Exploration Editor to configure the mapping completely.

Shape Details are not displayed in Stream Wizard when you select an Existing Shape

  • Bug ID: 23065916
  • Added: 18-April-2016
  • Platform: All

While creating a HTTP based stream, when you select an existing shape, the shape details are not displayed

Workaround:

None

Unexpected Output from Spatial General Pattern

  • Bug ID: 23007524
  • Added: 18-April-2016
  • Platform: All

The Spatial General Pattern keeps displaying Enter status in the Live Output Stream for both Enter and Exit statuses used in the Spatial General Pattern. The pattern outputs some records whose records are empty.

Workaround:

None

IPv6 Format URL is Detected an Invalid

  • Bug ID: 23066531
  • Added: 18-April-2016
  • Platform: All

The IPv6 format URL is detected as invalid on Oracle Stream Analytics User Interface, like the url http://[2606:b400:2010:504f:221:f6ff:fea6:331a]:9002/resttarget for rest target url.

Workaround:

None

Kafka Topics Group is not set in the Spark Exploration

  • Bug ID: 23081310
  • Added: 18-April-2016
  • Platform: All

Though a Kafka Topic Group is specified in the stream, the group is not visible in the corresponding application.

Workaround:

None

Exploration not displayed in Catalog when the deployment fails in Spark Runtime Environment

  • Bug ID: 23081450
  • Added: 18-April-2016
  • Platform: All

An exploration whose deployment fails on the Spark Runtime environment does not appear in the Catalog.

Workaround:

Refresh the web page or the Catalog to see the exploration.

Hide/Unhide Behavior is Incorrect in Change Detector Pattern Exploration

  • Bug ID: 23082097
  • Added: 18-April-2016
  • Platform: All

In Change Detect Pattern, when you add a non-first (like second or third) change criteria, the original change criteria property column does not appear in the live output table. When you delete change criteria, the last deleted change criteria related property columns are still displayed in the live output table.

Workaround:

Click Restore Columns, or change the Windows Range. You can also navigate to Catalog and then come back to the exploration.

Incorrect Expression Generated for Filter with Byte Property

  • Bug ID: 23094654
  • Added: 18-April-2016
  • Platform: All

If you create a filter with a property of Byte type, an incorrect expression is generated.

Workaround:

None

K-means.Anomaly Pattern Throws CCE

  • Bug ID: 23094768
  • Added: 18-April-2016
  • Platform: All

The K-means.Anomaly pattern supports only Double data type for Value1 and Value2 fields. If you specify any other data type for these fields, the pattern throws CCE.

Workaround:

None

Spark Profile Application may not receive events after Undo/Redo

  • Bug ID: 23095620
  • Added: 18-April-2016
  • Platform: All

When you make any changes to an exploration in the Spark runtime environment and perform Undo/Redo operations on the exploration, the events are not received any more in the Live Output Stream.

Workaround:

None

Events not received while using non pre-defined Kafka Topic

  • Bug ID: 23095677
  • Added: 15-April-2016
  • Platform: All

When you use a non-existing Kafka topic name while creating a stream, events are not received in the Live Output Stream.

Workaround:

Create a Kafka topic before creating an exploration that uses the stream with the Kafka topic

Empty Exploration Exception Thrown After Swapping Sources

  • Bug ID: 23103867
  • Added: 15-April-2016
  • Platform: All
When you swap the sources within an exploration, an empty exploration exception like SXException: Empty exploration is thrown occasionally.

Workaround:

None

SQL Column Types Map to Incorrect CQL Type

  • Bug ID: 23102546
  • Added: 15-April-2016
  • Platform: All

The following SQL Column Types are mapped to incorrect CQL types:

  • REAL
  • SMALLINT
  • NUMERIC
  • FLOAT
  • INTEGER
  • DECIMAL
  • BINARY_FLOAT

Workaround:

None

Alert for Change Does Not Happen Correctly for the Change Detect Pattern

  • Bug ID: 23035266
  • Added: 05-April-2016
  • Platform: All

You need to get alert immediately after any change happens in the specified duration. But, this does not happen. You cannot receive alert immediately when a value has changed and can only receive the alert either at the moment when changes do not happen anymore or at the end of the specified duration.

Workaround:

Minimize the window value to receive events just after the changes happen

GeoFence Does Not Work with 3857 SRID

  • Bug ID: 21259269
  • Added: 03-April-2016
  • Platform: All

The Spatial pattern does not work when you use 3857 as the SRID.

Workaround:

None

Catalog Filter Remains after Logout and Login

  • Bug ID: 22662900
  • Added: 03-April-2016
  • Platform: All

The filters applied to the Catalog are retained even after you logout of the application and login again. The filters must ideally be removed between sessions, but they are retained.

Workaround:

None

Mouse Cursor Shows Poor Alignment While Adding Business Rules

  • Bug ID: 22736268
  • Added: 03-April-2016
  • Platform: All

When you try to add Business Rules to an Exploration, the mouse alignment is not correct when you hove the mouse over the Add Rule field. This issue is specific to Firefox browser.

Workaround:

None

Special Characters are not supported in the Exploration HTTP Pub-Sub Channel Names

  • Bug ID: 23014064
  • Added: 03-April-2016
  • Platform: All

If the HTTP Pub-Sub channel used in an exploration has special characters like ".", the exploration fails. Special characters are not supported in the names of any channel or exploration.

Workaround:

None

Boolean Shape Type Confusion in Streams

  • Bug ID: 23027340
  • Added: 03-April-2016
  • Platform: All

Oracle Stream Analytics supports the Boolean values of true and false. Any other values including 1 and 0 are treated as false.

Workaround:

None

Wrong Shape Details Shown While Using Existing CSV Shape

  • Bug ID: 23027494
  • Bug ID: 23027494
  • Platform: All

When you use an existing shape in a new stream, the shape detail shown are incorrect

Workaround:

None

NullPointerException Thrown while Saving Twitter Stream of an Exploration

  • Bug ID: 23021791
  • Added: 03-April-2016
  • Platform: All

When you edit and save the Twitter stream of an exploration, a NullPointerException is thrown occasionally. When this exception is thrown, the stream does not work

Workaround:

None

EDN Exploration cannot publish results to EDN Target

  • Bug ID: 23026764
  • Added: 03-April-2016
  • Platform: All

If the field of EDN's target is a required field (as specified in the EDN schema) but is not mapped properly, the exploration cannot publish results to the EDN target.

Workaround:

For EDN targets propagate the required flag to the target field mapping UI and do not let users omit specifying mapping in fields. For required fields, if the mapping is not specified, send NULL values.

IllegalStateException after Editing a Twitter Stream

  • Bug ID: 23021690
  • Added: 03-April-2016
  • Platform: All

When you edit and save a Twitter stream, an IllegalStateException is triggered occasionally.

Workaround:

None

Swapping Sources does not work in Patterns

  • Bug ID: 23037709
  • Added: 03-April-2016
  • Platform: All

When a pattern has multiple sources, and if you swap the sources, the change is not reflected in the Live Output Stream.

Workaround:

None

Encoding Issue with "%" character in Inbound REST Adapter

  • Bug ID: 22068888
  • Added: 27-October-2015
  • Platform: All

"%" character without any trailing hexa character must be encoded with "%25" in message payload to Inbound REST adapter. The following exception is thrown if you do not encode: java.lang.IllegalArgumentException: URLDecoder: Illegal hex characters in escape (%) pattern - For input string

Workaround:

None

Output for combination of UNIONALL with MINUS/UNION on subquery is not as expected

  • Bug ID: 20055734
  • Added: 13-October-2015
  • Platform: All

The output in case of sub queries with a combination of the CQL Operators UNIONALL and MINUS or UNION is not as expected. This is a precedence issue in CQL grammar for binary operators.

Workaround:

Use the parenthesis at appropriate places in the sub query based query to get the appropriate behavior. For example, instead of

select c1,c2,c3 from ch0[rows 1000] union all select c1,c2,c3 from ch1[rows 1000]

minus select c1,c2,c3 from ch2[rows 1000]use
(select c1,c2,c3 from ch0[rows 1000] union all select c1,c2,c3 from ch1[rows 1000]) 
minus 
select c1,c2,c3 from ch2[rows 1000]

WLEVSJConsole does not work on Windows because of wrong Class path

  • Bug ID: 21972306
  • Added: 12-October-2015
  • Platform: Windows

The wlevsjconcole.cmd does not work correctly on a Windows machine. This is because the class path points to an incorrect version of files.

This is a known issue.

Workaround:

Directly use the JConsole command included in the JDK.

Special Handle for Timestamp Type Conversion between CQL and JDBC with Derby

  • Bug ID: 18311700
  • Added: 8-June-2015
  • Platform: All

When you join a stream with the Derby database, where the join condition compares two timestamp values where one value is from the stream attribute and the other value is from the Derby data source attribute, the Derby database performs the predicate evaluation. However, the Derby database supports only the yyyy-mm-dd-HH.mm.ss[.nnnnnn] format. For the Derby database to perform the evaluation correctly, the stream timestamp value must use the Derby database format.

Workaround:

Convert the timestamp from the stream to the Derby database format. The to_timestamp expression in the WHERE clause in the following example performs the conversion.

SELECT t.testTimestamp

FROM S1[NOW] as s, T1 as t
WHERE to_timestamp(to_char(s.testTimestamp, 'yyyy-MM-dd HH:mm:ss'),
'yyyy-MM-dd HH:mm:ss') = t.testTimestamp
If the timestamp value contains a fractional second component,
 use the appropriate format to parse the fraction 
component as shown in the following example:
SELECT t.testTimestamp
FROM S1[NOW] as s, T1 as t
WHERE to_timestamp(to_char(s.testTimestamp, 'yyyy-MM-dd HH:mm:ss.SSS'),
'yyyy-MM-dd HH:mm:ss.SSS') = t.testTimestamp

Cannot Get Event Type of Complex Element to Create EDN Stream

  • Bug ID: 20700113
  • Added: 3-March-2015
  • Platform: All

Creating a stream with EDN as a source has a few limitations. The other EDN types are not visible for selection. In this release:

Only the XML element that has a flat structure is supported.

The XML schema does not import other schemas.