You see an SQL error when you run a task created by a user whose user name contains a single quote. For example, you see an SQL error when you open a task created by a user whose name is O'Leary.

When you submit a form that contains an attachment, ActiveVOS does not decode the Unicode characters in the form correctly. For example, if you submit a form that contains Mandarin text and an attachment, ActiveVOS does not correctly decode the Mandarin text in the form. You see small boxes in the place of Mandarin characters.

A web service call that requires authentication does not generate a response. This is because of a configuration issue in the javax.xml.xpath.XPathFactory file that occurs when you use Tomcat with Java version 8.

When you deploy a contribution that contains WSDL files with complex XML schema definitions, you might see an Out of Memory error. You see errors if there are duplicate imports of complex WSDL schema definitions, which occupy a large chunk of Java heap memory.

ActiveVOS uses an outdated Struts version that potentially enables users to perform malicious code execution.

If recovery takes a long time, the ActiveVOS Alarm Manager sometimes sends duplicate alarms. When this happens, the ActiveVOS console indicates correlation violations.

If you use Microsoft SQL Server and perform process recovery, you see syntax errors. The errors you see are QueueStorage.GetQueuedMessageReceivers and QueueStorage.GetAlarmsJoinProcess.

Recovery time is slow because the ActiveVOS server is unable to get certain resources. This is because these resources are at maximum capacity as the number of processes that the server needs to recover is higher than the engine process count.

The getTask Human Task request fails with the error "ORA-01795: maximum number of expressions in a list is 1000". This happens if you use an Oracle database and the user for whom you are sending the request has more than 1000 roles.

There is a deadlock during subprocess coordination if more than two processes are involved.

ActiveVOS Process Migration converts the XML variable infaTask to a string.

Sometimes, ActiveVOS incorrectly auto-claims Human Tasks to a single user. This happens when the potential owner list of a human task contains both a single user and a user group.

When you restart the ActiveVOS server, sometimes, completed processes also undergo recovery.

You get a Null Pointer Exception when you deploy a Java Message Service (JMS). This happens because of a Fixed Listener name lookup issue.

When you claim a task that belongs to a multiple delegates, you see the XQuery error "Attempting to access a sequence as a singleton item" followed by a correlation violation.

If you use ActiveVOS on an Oracle WebLogic 12.1.3 server without a specific JVM startup parameter, you find that startup is slow. You may see messages like "Unable to locate catalog resource". This happens because of a change to the WSDLFactory preference in Oracle WebLogic 12.1.3. If you are an ActiveVOS 9.2.4.1 or 9.2.4.2 customer who uses Oracle WebLogic 12.1.3,set the following JVM startup property:Djavax.wsdl.factory.WSDLFactory=com.ibm.wsdl.factory.WSDLFactoryImpl

When the ActiveVOS server receives JMS message that results in a correlation exception, the process either hangs or times out. This is because the JMS client expects but does not get a response at the replyTo address as the ActiveVOS server does not send any system-generated response to it. The JMS client is expected to process the error condition and the following fault: < aex:Fault xmlns:aex="http://www.active-endpoints.com/2004/06/bpel/extensions/"> <faultcode xmlns:ns="http://docs.oasis-open.org/wsbpel/2.0/process/executable" name="correlationViolation" namespace="http://docs.oasis-open.org/wsbpel/2.0/process/executable"/>

When you use the Human Tasks query getTasks, the server is slow to respond when the data set is large.

If you use JMS queue to send a message to the ActiveVOS server, you see a correlation violation even though there is a process that is waiting for the message.

Fixed a defect calling the WS-HumanTask getAttachment operation which could result in a fault.

Fixed a defect where the Additional query dialog of the ActiveVOS Console's process list page did not pop up when using Internet Explorer 7.

Database queries were enhanced to eliminate contention when performing WS-HumanTask getTasks operation that resulted in poor response times. See also defect 16607.

Fixed a defect installing ActiveVOS Server 9.2.2 on IBM WebSphere 7 on Solaris and RHEL.

Fixed a deadlock in the ActiveVOS server's AeResource cache. Under specific circumstances, a few minutes after starting the server, several threads could get stuck resulting from a deadlock. This would stop further server processing. This deadlock condition was caused in the caching layer of the AeResource cache.

Note: ActiveVOS 9.2.2 patch Server922-5.zip is also available. This patch resolves defects 16269, 17041, and 17115, which were fixed with ActiveVOS 9.2.4.

Fixed a problem with reports that did not launch in the ActiveVOS Console when the console is launched using the Reverse Proxy URL. Note: ActiveVOS 9.2.2 patch Server922-2.zip is also available.

Fixed a problem with BIRT reports saved with the ActiveVOS Designer 9.2.2 that would fail to open up in the ActiveVOS Console.

Fixed a problem with the WebSphere servlet/filter that interfered with ActiveVOS reporting. Note: ActiveVOS 9.2.2 patch Server922-1.zip is also available.

For database servers other than MySQL Process Recovery was slow due to incorrect/missing index. An additional index was added and included in the upgrade DDL

Database queries were enhanced to eliminate contention when performing the WS-HumanTask getTasks. For database servers other than MySQL Process Recovery was slow due to incorrect/missing index. An additional index was added and included in the upgrade DDL operation that resulted in poor response times.

Fixed a problem with queue receive IDs going over 2 billion. Customers running in an ActiveVOS MultiSite configuration are likely to encounter this issue. For details on MultiSite, see

http://infocenter.activevos.com/infocenter/ActiveVOS/v92/topic/doc.server_console/html/SvrUG6-11.html.

Fixed a problem with the WS-HumanTask getInput operation that could fail due to strict XML Schema validation when the schema used a substitution group or when the task input used xsi types. Note: ActiveVOS 9.2.2 patch Server922-4.zip is also available.

Fixed a problem with the <abp:timeout> partner policy on an endpoint reference not being applied correctly for certain cases that allowed processes to remain in the running state. Note: ActiveVOS 9.2.2 patch Server922-3.zip is also available.

Fixed a problem with the expiration of a human tasks if the date of expiration is greater than 12 days.

Fixed a problem saving process state causing the server to stop processing. This issue would arise when variable state saves were attempted when multiple versions of a variable were being saved. This could sporadically occur and be reported with the following stack trace in the server log, or execution markers not showing in the process detail view.

Apr 16, 2014 9:19:30 AM org.activebpel.rt.AeException logError SEVERE: Warning: probably need to use a type-specific null value for parameter 1! org.activebpel.rt.bpel.server.engine.storage.sql.AeQueryRunner$AeWarning ... Note: ActiveVOS 9.2.2 patch Server922-5.zip is also available. This patch resolves defects 16269, 17041, and 17115 fixed with ActiveVOS 9.2.4.

Fixed a problem that could cause threads to be blocked in a busy environment because the default counter block size for the JournalID journal counter was set too low. The problem materialized in the logs as follows: WorkManager.aeWorkMgr : 124" daemon prio=3 tid=0x064df000 nid=0x5d1b waiting for monitor entry [0x1a92e000] java.lang.Thread.State: BLOCKED (on object monitor) at org.activebpel.rt.bpel.server.engine.storage.AeCounter.getNextValue(AeCounter.java:160) - waiting to lock ... Note: ActiveVOS 9.2.2 patch Server922-5.zip is also available. This patch resolves defects 16269, 17041, and 17115 fixed with ActiveVOS 9.2.4.

Fixed a contention problem in the WSDL Locator object. This materialized as thread dumps in a highly loaded environment; for example: "WorkManager.aeWorkMgr : 210" daemon prio=3 tid=0x04c40800 nid=0xae7 waiting for monitor entry [0x0f67e000] java.lang.Thread.State: BLOCKED (on object monitor) at org.activebpel.rt.wsdl.def.AeBPELExtendedWSDLDef. ... Note: ActiveVOS 9.2.2 patch Server922-6.zip is also available.

Corrected a situation where a single process being recovered by the engine containing thousands of records returned from a service invoke would take an excessive amount of time, causing thread starvation as the process recovery thread would hold up a majority of threads in the system. Process recovery now uses a 30 second timeout. This limits the amount of time a single process may take for recovery. If recovery of a single process exceeds this limit, it will be faulted and information is displayed in the ActiveVOS Console. The server log will record the server trace of the recovery thread at the time it was interrupted.

Fixed a problem with correlation not working when instantiating a process using a Pick activity when a process has correlation set to one message (but not all messages) and there is a subsequent intermediate Pick to be correlated in the same Pick branch that started the process Note: ActiveVOS 9.2.2 patch Server922-7.zip is also available.

Fixed a problem reconnecting to a remote JMS broker that was failing over to the default initial context when a JMS Manager was configured to use a remote JNDI location. This occurred after the loss of connection to a remote JMS provider. The server incorrectly reverted to using the local JNDI context instead of the remote one, and never successfully reconnected to the remote queue. The only work around was to restart ActiveVOS after remote queues were back up and available.

An ActiveVOS 9.2.2 patch is available upon request.

Fixed a problem with the businessAdministrators and potentialOwners groups variables getting lost after changing the logging level of deployed process in ActiveVOS console.

Fixed a problem where the search filter in "Deployment Log" does not work properly when errors are present in the deployment logs.

Fixed SQL performance issues when taking contributions offline These issues occurred during recovery and taking contributions offline in large databases (millions of records in the AeProcess table) and resulted in long running SQL operations. The tables primarily affected are the AeProcess and AePlan tables. Changes to existing indexes were made.

Customers should follow the instructions found in the "Upgrading Database Schemas" below.

To improve process purge performance, foreign keys were added as indices to a number of tables.

Corrected an issue with the Dispatch service not decrementing the executing request count.

Fixed a problem where getProcessListIO did not execute without all of its arguments.

Fixed a problem where the documentation did not indicate that when deploying ActiveVOS tomcat server, the activebpel-cert.war file is not automatically deployed.

Fixed a problem where threads became stuck when a server was started and a node failure occurred.

Fixed a getAttachment() error where the request failed because of a timeout.

Fixed a problem in ActiveVOS Central where no Screenflow guide tabs were displayed.

Fixed a problem with the timestamp policy for MyRole where the created timestamp would not accurately reflect the time for which a message would be valid.

Fixed a problem where a queued receive process ID was not correct when process IDs were large.

Fixed a problem where the outer variable (the counter) in a for loop was not properly restored from the state document. This most often occurred when the parallel flag was enabled in the outer for loop.

Fixed a problem where changes made to an XQuery within the ActiveVOS console did not take effect until the application server is restarted.

Fixed a problem where BPRD execution is always successful even if deployment failed.

Fixed a problem where scheduled purges do not occur after a node was restarted.

Fixed a problem with the ActiveVOS implementation of OAuth 2.20 where it did not recognize the correct payload.

Fixed a problem with a tech note on configuring the ActiveVOS console with CAS was incorrect.

Fixed a problem with a correlated subquery accessing the AEPROCESS table taking excessive time to execute. (This is related to the DisableExpiredPlans function.)

Fixed a problem where an Ant script could not find an automatic step's WSDL file.

Fixed a problem where the content type was changed in a REST invoke.

Fixed a problem where the installer wouldn't execute for the JBoss server when the JDK is 1.7 on Linux.

Fixed a problem where a meaningful error message did not display when there was a null pointer exception error during simulation.

Fixed a problem with the timestamp policy for MyRole where the created timestamp would not accurately reflect the time for which a message would be valid.

Fixed a problem with parameter ordering in the JSON binding when starting a task.

Fixed a problem when a null pointer exception occurred when rendering a process within the ActiveVOS console.

Fixed a problem where users where unable to login to ActiveVOS Central 9.2.1 running on JBoss 7.1.1 running on Linux.

Fixed a problem where the Administration Console did not always display the correct links when running under JBoss 7.

Fixed a problem where the additional information bar was not resizable.

Fixed a problem where a link definition error occurred. This occurred when there are a number of scopes that are linked with conditional or unconditional links.

Fixed a problem where the Designer generates an unsorted guide-catalog.xml file.

Fixed a problem where the getAllHttpHeaders() function did not return a result message.

Fixed a problem where process service times do not display a graphic image in Firefox.

Fixed a problem where the server start-up aeEngineResource status is reported inconsistently.

Fixed a problem where data validation did not occur when an attachment was not added during an Add Attachment Automated step in Internet Explorer.

Fixed a problem where some themes did not work for Internet Explorer 7 and 8.

Fixed a problem where guide action buttons are re-enabled before a long running automated step completes.

Fixed a problem where Screenflow would hang after previously using either pressing the Go Back One Step button or an entry in the History list.

Fixed a problem in the documentation where the WebLogic version numbers were incorrect.

Fixed a problem where data validation did not occur after reentering invalid the first time.

Fixed a problem where the Synchronize button was not enabled in the Cloud Extend Setup page in Internet Explorer and Firefox.

Fault Data in the Retry Service Input Fixed: When triggering an alert service with a "suspend on uncaught fault" flag, you may have difficulties determining if the process actually completed This problem will not occur when using the Retry service.

Signal Events are not shown in an existing workspace This problem can occur if you have a customized palette in your workspace. Copy the signal events entries over or removing the customized palette allows these events to be shown.

Getting error transmitting multisite message after restarting one site server in a multisite You may see the following message: Error transmitting MultiSite message from Site streamsa to http://streams:8080/active-bpel/services/activevosSiteComm.

This message appears periodically and disappears when a site is restarted. After the site is fully online, the message stops.

Fixed a problem whereby an invalid WS-HumanTask setOutput API payload request would timeout. Reference: http://forums.activevos.com/showthread.php?t=5945

Fixed a problem where by "Delegate" (assign) and "Release" task commands was not limited to business admins only, but incorrectly to all potential owners.

Fixed an issue where server startup taking long time on a query to AeCatalog. This results in significant improvement in server startup time.

Fixed a problem where reconnection of listeners to a remote Oracle WebLogic queue would sometimes fail due to recursive calls to the JMS exception listener Reference: http://forums.activeendpoints.com/showthread.php?t=6336

Fixed a problem restoring counter variables in parallel forEach scopes after the process has been restored from the state document. Reference: http://forums.activeendpoints.com/showthread.php?t=6234

Versions of WebLogic prior to 12.2 have file upload issues You may encounter issues that include not being able to deploy a BPR, import identity configuration settings, adding attachments, and the like.

The workaround is update a later release of WebLogic or download and install patch 13657792. Enter this value in the patch number search field on the patches & updates tab of my oracle support.