AgilePoint BPMS v5.0 R2 for .NET 3.5, .NET 4.0, and Genesis Edition
AgilePoint BPMS v5.0 SP3 for Universal Edition and Azure
© 2014 AgilePoint, Inc. All rights reserved.
| Release Notes v5.0 SP4 / Fixes | |
The following issues have been fixed in this release for AgilePoint Server.
Updates in AgilePoint Server Manager (Windows Service environments) sometimes do not take effect due to file locking issues.
When a user clicks a text box in the AgilePoint Server Manager (Windows Service environments), the selected node in the navigation tree is deselected.
In Advanced Data Population configuration in Envision, when a user selects a table in an Oracle repository, an unhandled exception occurs, and the configuration cannot be completed.
Duplicate overdue emails are being sent when AgilePoint Server is recycled.
RollbackProcInst API throws "Index was outside the bounds of the array" error in a rare scenario.
If an Or AgilePart is in exclusive mode and activities in one of the preceding branch were completed, the Or AgilePart cancels only the currently active activity in the other branch. It should skip all future activities in the non-executing branch.
In a rare scenario where 2 activities are writing back to the same custom attribute row in a table, 1 of the activities overrides the values set by another activity due to a race condition. In other words, the update custom attribute action is not thread safe.
A process delays indefinitely at the And AgileShape.
When a long-running AgilePart completes its processing, the process does not move forward.
In an NLB environment, duplicate emails occur when a slave server has many emails to send, but the master server does not have the emails in memory, so it must retrieve them from the database.
A deadlock occurs sometimes in the table WF_MAIL_DELIVERABLES.
The CSM locker does not work on Oracle Database.
In the File Dropping event of the Event Service AgileConnector, the following issues occur:
After recovering from a network failure, AgilePoint Cluster Server Manager should go into Slave Mode, but instead it stays in Master Mode.
Occasionally AgilePoint Cluster Server Manager takes a long time to stop.
Occasionally when a SQL deadlock occurs, AgilePoint Server restarts unexpectedly.
Occasionally the Delay AgilePart delays longer than configured.
In an NLB environment, Data Services instances may process the same record when the master-slave server switches due to timing overlap among the multiple Data Services instances.
In some environments, Oracle 11g hangs when AgilePoint Server executes a query of type LONG.
When a large number of process instances are running (for example, > 50,000 with > 500,000), AgilePoint Server is slow to restart.
When AgilePoint Server restarts during archiving, child process instances are orphaned in the AgilePoint database. To fix this, all the child process instances are archived prior to archiving the parent process instance.
System.Xml.XmlException: '.', hexadecimal value 0x00, is an invalid character
The database schemas from AgilePoint service packs were not in sync with database schemas in the AgilePoint installers.
Cluster Manager Server does not parse the database connection string (csmServiceexe.config > database node) if it contains a colon (:).
Occasionally the CreateProcessInstance method is executed more than once in an NLB environment.
If the CSM URL table contains no records, Oracle fails to initialize it.
When one or more Cluster Manager Server instance stops, AgilePoint Server takes about 5 minutes to respond.
With this fix, if one CSM instance stops, AgilePoint Server recovers within 5 seconds. If all CSM instances stop, AgilePoint Server stops within about 2.5 minutes.
Only the users from the same domain, which AgilePoint System user belongs to, are synchronized while synchronizing users of multiple Active Directories with the users of AgilePoint.
In an environment with a fast AgilePoint Server machine and database machine, if AgilePoint Server has many transactions in a row, the events may go into more than one thread. This causes issues such as executing AgileParts more than once, or canceling child process instances right after they are initiated.
A performance issue because of memory leak in Schema Mapper module.
Data Service takes longer time (495 ms/per record) to process a record from WF_DATA_SERVICE_TEMPDB.
Duplicate mails are sometimes sent in an NLB environment.
If a subprocess with a custom name has a hyphen (-) AgilePoint does not recognize the name.
If the task due date is set based on the time span in the current version of the process, and a new version of the process is deployed with a different time span, due date of the current activity is not updated when the process is migrated to the new version.
When a participant of a manual activity is an Active Directory group, and the Active Directory group is not reachable, the process hangs at that activity.
Parent and child process archiving does not take the relationships into account, so sometimes child processes became orphans if AgilePoint Server is interrupted or restarts. With this fix, archiving occurs with attention to parent/child process relationships, archiving from child first, and then to parent.
This fix corrects the problem going forward, but it does not correct currently orphaned child processes. To acquire a tool to fix this issue, contact AgilePoint Support.
On the AgilePoint Server Manager window, display errors occur when using Windows Server 2008 R2 in a Simplified Chinese environment.
When running in Debug mode, the Mail Approval AgileConnector is unable to pass debug messages to the log. Instead it throws exceptions related to this logging error.
If you select both a parent and child table that are associated with a foreign key at the same time in Data Population, an error occurs.
Process instance stalls when it can't find the participant.
The UndoWorkItem and AssignWorkItem events cannot be accessed through the AgilePoint API.