Version 6.1 MP2

Version 6.1 MP1 Patch 1

Version 6.1 MP1

AquaLogic Interaction 6.1 MP2 released June 2008

AquaLogic Interaction 6.1 MP1 Patch 1 released August 2007, for AIX, HP-UX, Red Hat Linux, Solaris, SUSE Linux, and Windows

AquaLogic Interaction 6.1 MP1 released April 2007, for AIX, HP-UX, Red Hat Linux, Solaris, SUSE Linux, and Windows

AquaLogic Interaction 6.1.0.1 released December 2006, for AIX, Solaris, and Windows

AquaLogic Interaction 6.1 released October 2006, for AIX, Solaris, and Windows

These release notes cover all releases of AquaLogic Interaction 6.1, including patches and maintenance packs. For information on a particular release, please go to the appropriate section of these release notes.

Release notes are occasionally updated after the release date. For the most up-to-date release notes, go to edocs.bea.com.

The BEA licensing framework has been removed from this product entirely. There is no need to obtain or use a license key to enable any functionality in this version.

Refer to the Interoperability page in the Product Center at one.bea.com/support for the latest information on supported BEA products.

At the time of release, AquaLogic Interaction 6.1 MP2 supports the following:

• AquaLogic Interaction Collaboration 4.2 MP1, Publisher 6.4 and 6.4 MP1, Studio 2.2 MP1, and Analytics 2.1
• AquaLogic Interaction Content Service - Windows Files 6.1, Identity Service - Active Directory 6.3, and Identity Service - LDAP 2.2
• For information on additional products, refer to the Interoperability page.

• Portlets may fail to load if the remote server host name contains an underscore character ('_'). (Issue #56709)
• Under certain configurations with previous releases of WebLogic Server, the portal CPU utilization may increase to 100% over an extended period of time. (Issue #61632)

• An error appears when adding a page to a community when the page is based on a page template that was returned by a search. (Issue #57030)
• Portlets inherited from page templates can be edited but changes are not saved. (Issue #60054)
• Community pages built from a non-inherited page template revert to the settings of the page template after the community is copied. (Issue #60507)
• Community pages do not properly inherit content canvas portlets. (Issue #60903)
• Page Template portlets are not inherited when creating a new community page from the Community Editor. (Issue #61042)
• When changing content canvas portlets on a page, the page cannot be saved. (Issue #61272)
• The page template editor allows more than one content canvas portlet to be added. (Issue #61276)

• Months are displayed in English for non-English locale users. (Issue #57807)
• Community names that contain Korean characters are displayed incorrectly in search results. (Issue #60349)
• Day and month names in ALI Collaboration calendars are not localized. (Issue #62745)

• The openhttp:LoadBalancedIPs setting is not properly processed. (Issue #55532)
• MPPE load balancing fails to switch over to another TCP port when there are multiple application server instances running on different ports on the same host. (Issue #61325)

• Field names in fielded search are improperly case sensitive. (Issue #56073)
• The Activity Rank gets removed with document updates that do not update the Activity Rank. (Issue #58868)
• Search server cluster nodes shut down or fail under high load from AquaLogic Pathways. (Issue #61662)
• Search server archive files have '..' in the file name. (Issue #62118)
• A fielded search query of the form "property:" without specifying a value has inconsistent results. (Issue #62304)
• Documents with an Activity Rank value occasionally cause “FILE_REMOVE_FAILED” errors in the search server log file. (Issue #62306)
• In AquaLogic Pathways, the Activity Rank scoring is not properly assigned for documents on search nodes other than the one receiving a given search query. (Issue #62617)
• If a particular document property is unable to be indexed, a small memory leak will occur for every subsequent index of that document. (Issue #62703)

• On HP-UX, adding more than one crawler operation to a job will cause both crawlers to fail. (Issue #57589)
• When a Knowledge Directory card that was created by a crawler is deleted, duplicate cards may be created through subsequent runs of the crawler. (Issue #60506)
• When running a mirrored Documentum crawler, documents are not consistently imported. (Issue #61666)
• Mirrored references do not get auto-approved even when "Automatically approve imported documents" is checked in the crawler settings. (Issue #61677)
• After running a mirrored crawl, if one of the documents crawled in is deleted from the Knowledge Directory, all subsequent runs of the crawler will fail. (Issue #62021)

• The suspension time setting for a content crawler job is not saved. (Issue #57319)
• Under some circumstances when a job crashes or is unresponsive, it may not be cleared from the automation server after the automation server is restarted. (Issue #59906)

• A content canvas portlet will not initially display after it is added to a community page template. Workaround: Open the editor for the community page. Once the community is edited and saved, the content canvas portlet will be displayed. (Issue #61274)

• If there are AquaLogic Interaction services running when the installer is executed, the ALI 6.1 MP1 Patch 1 installer will report error messages in the installer log file and not overwrite locked files. Workaround: Shut down all running AquaLogic Interaction services and re-run the installer. (Issue #62903)

Refer to the Interoperability page in the Product Center at one.bea.com/support for the latest information on supported operating systems, application servers, databases, and browsers.

AquaLogic Interaction 6.1 MP1 supports the following:

• Operating Systems: AIX 5.3 (on POWER3, POWER4, POWER5); HP-UX 11i V2 (on Itanium); RedHat Enterprise Linux 4 update 3 (on x86); Solaris 8, 9, and 10 (on SPARC); SUSE Enterprise Linux 9 (on x86); Windows Server 2003 SP1 or R2, SP2 (on x86)
  
  Solaris 8 requires these patches:
  • 108434-21 (32-Bit Shared library patch for C++)
  • 108435-21 (64-Bit Shared library patch for C++)
  
  Solaris 9 requires these patches:
  • 111711-15 (32-bit Shared library patch for C++)
  • 111712-15 (64-Bit Shared library patch for C++)
  
  
• Databases: Microsoft SQL Server 2000 SP3a or SP4, Microsoft SQL Server 2005 (with 2000 Compatibility Level), Microsoft SQL Server 2005 SP2 (with 2000 Compatibility Level), Oracle 9i (9.2.0.4 and above), Oracle 10g (10.1.0.3 and above), Oracle 10g R2 (10.2.0.1 and above)
  
• Application Servers:
  • AIX:
    • IBM WebSphere 6.0.1 with IBM (32-bit) SDK 1.4.2 SR6, IBM WebSphere 6.1 with IBM (32bit) SDK 5.0 SR2
  • HP-UX:
    • BEA WebLogic 9.2 with HP-UX JDK 5.0.03 with Java HotSpot Server VM (32-bit), BEA WebLogic 9.2 MP1 with HP-UX JDK 5.0.04 (32-bit) with Java HotSpot Server VM
  • Red Hat Enterprise Linux:
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with BEA JRockit 5.0 (R26.0.0) JDK (32-bit); BEA WebLogic 9.2 MP1 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit), BEA WebLogic 10 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit) and all later 5.0 releases of JRockit
  • Solaris:
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06; BEA WebLogic 9.2 MP1 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06; BEA WebLogic 10 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06
    • IBM WebSphere 6.1 with IBM (32bit) SDK 5.0 SR2
  • SUSE Enterprise Linux:
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with BEA JRockit 5.0 (R26.0.0-189_CR269406) JDK (32-bit); BEA WebLogic 9.2 MP1 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit), BEA WebLogic 10 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit) and all later 5.0 releases of JRockit
  • Windows:
    • Microsoft IIS 6.0 with .NET Framework 1.1 SP1
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with Sun Java 2 JDK 5.0 (32-bit) with Java HotSpot Client VM, version 1.5.0_06; BEA WebLogic 9.2 MP1 with Sun Java 2 JDK 5.0 (32-bit) with Java HotSpot Client VM, version 1.5.0_06; BEA WebLogic 10 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06
• Browsers: Administrative Use - Firefox 2.0, Internet Explorer 6.0, 7.0; Browsing Use - Firefox 1.5, 2.0; Internet Explorer 6.0, 7.0; Netscape 7.2, 8.0; Safari 2.0 (Mac only)

Note: AquaLogic Interaction 6.1 MP1 does not support IPv6. If a system is incorrectly configured to enable IPv6, errors may occur with your system. You must ensure that IPv6 is not enabled on your system before installing AquaLogic Interaction 6.1 MP1. For help verifying whether IPv6 is enabled, please contact Customer Support.

Refer to the Interoperability page in the Product Center at support.plumtree.com for the latest information on supported BEA products.

At the time of release, AquaLogic Interaction 6.1 MP1 supports the following:

• AquaLogic Interaction Collaboration 4.2 MP1, Publisher 6.4, Studio 2.2 MP1, and Analytics 2.1
• AquaLogic Interaction Content Service - Windows Files 6.1, Identity Service - Active Directory 6.3, and Identity Service - LDAP 2.2
• For information on additional products, refer to the Interoperability page.

As more platforms are supported, additional upgrade paths to the new supported platforms will be included.

AquaLogic Interaction 6.1 MP1 Patch 1 only supports upgrades from AquaLogic Interaction 6.1 MP1.

AquaLogic Interaction 6.1 MP1 currently supports upgrade directly from the following previous releases:

• AquaLogic Interaction 6.1.0.1 and 6.1.
• Plumtree Foundation 6.0 and 6.0 SP1.
• Plumtree Corporate Portal 5.0.2, 5.0.3, 5.0.4, and 5.0.5 on Windows.
• Plumtree Corporate Portal 5.0.4J and 5.0.5J Solaris.
• Plumtree Corporate Portal 4.5WS SP3 on Windows.

The following product documentation is available for download at edocs.bea.com/alui/ali/docs61:

Document Title and File Name	Document Description
Installation and Upgrade Guide for BEA AquaLogic Interaction 6.1 install.pdf	Describes how to install and upgrade the components of AquaLogic Interaction 6.1.
Administrator Guide for BEA AquaLogic Interaction 6.1 admin.pdf	Describes how to use, manage, and administer AquaLogic Interaction 6.1.
Online Help	Describes how to use the user interface features delivered with AquaLogic Interaction 6.1.
Installation Worksheet for AquaLogic Interaction 6.1 - Unix Worksheet_ALI_Unix.pdf	Print this document and use it to gather the host computer, portal, database, and additional information that you need to install and configure all of the various AquaLogic Interaction components.
Installation Worksheet for AquaLogic Interaction 6.1 - Windows Worksheet_ALI_Windows.pdf	Print this document and use it to gather the host computer, portal, database, and additional information that you need to install and configure all of the various AquaLogic Interaction components.

The release package for AquaLogic Interaction no longer contains a "ThirdParty" folder with a copy of Apache Tomcat and Apache HTTP Server. The correct locations to obtain a copy of Apache Tomcat and Apache HTTP Server for your OS are:

• Apache Tomcat: http://tomcat.apache.org
• Apache HTTP Server: http://httpd.apache.org

Additionally, the Sun Java Runtime Environment can be obtained from http://java.sun.com.

• It is now possible to clear a reference property value from an object. (Issue #49362)
• When using the File Upload to submit PDF files on a .NET portal, certain date attributes from the PDF can now be mapped to the document properties. (Issue #52252)
• Document title and description changes made from the Approve Directory Content utility are now saved. (Issue #53015)
• The Audit Manager will delete messages after selecting the 'Delete Messages and Archives when settings are saved?' checkbox and saving the Audit Manager settings. (Issue #53238)
• Incorrect physical memory size is no longer reported in the Search Server Cluster Status page. (Issue #54100)
• From the Administration directory, editing a group with a large number of members (16,000 or more) will no longer cause a high amount of CPU utilization. (Issue #57861)
• On Windows, clicking the 'Finish' button in Cluster Manager no longer returns you back to the Home Page. (Issue #58040)

• The following error is no longer displayed in the job log and Logging Spy when refreshing documents: Error in function PTCard.UnlockObject (): -2147205112 - Object not in locked state. (-2147205112) (Issue #52859)
• Crawl jobs will no longer hang when a crawler has a large number of Knowledge Directory folders with filters assigned. (Issue #53106)
• Knowledge Directory documents are now automatically approved when the target document is a linked card in the Documentum Content Server. (Issue #53324)
• Crawler and Document Refresh Agent operations will now remove Knowledge Directory documents that have been deleted from their original repositories. (Issue #53520)
• Two concurrent instances of the same mirror crawler will no longer delete each others cards. (Issue #54819)
• Documentum and Windows File Crawlers will no longer re-import all Knowledge Directory cards if 'Refresh ACLS' is selected. (Issue #56225)

• JavaScript with line breaks after a plus symbol (+) are now transformed properly when gatewayed. (Issue #53221)
• In the OnAfterLogin PEI, when trying to construct the logout URL redirect, the current user will be logged off properly. (Issue #53270)
• Using the adaptive tag <pt:ptui.loginauthsource pt:defaultselection='' pt:hidden='true'/> now selects the ALUI User Database as the authentication source. (Issue #54220)
• When using Firefo, when gatewaying an HTML form with inline-refresh enabled, if the form is posted empty, it will no longer generate an error. (Issue #54300)
• There is now a UICI for the AIX version of AquaLogic Interaction. (Issue #55945)
• With SSO enabled, if you write any portlet or login tag with the <pt:standard.loginlink> tag, it will now redirect to the SSO login page instead of the standard login page. (Issue #58441)

• The LocalMachineOnly setting in serverconfig.xml functions correctly now. If it is set to 'true', Logging messages will not be sent over the network. (Issue #36806)
• Using a Remote Server with a hostname that contains a dash (-) and a number (e.g. host-01) will no longer cause an error. (Issue #46871)
• The Guest user can no longer use the site map to list all users. (Issue #48648)
• Logging information from Search is now fully enabled on AIX or Solaris. (Issue #49527)
• Users will no longer be prompted for logins multiple times when using Security Mode 1 and accessing Collaboration or Publisher Web Folders. (Issue #50181)
• Mandatory Community tabs now highlight the entire button on mouse over. (Issue #52965)
• The navigation dropdown menus have been fixed so that certain HTML SELECT elements will no longer display through the background. (Issue #54689)
• Activity Server portlets are now able to generate Image Server and JS base URLs when the PT-Imageserver-URI string is a relative URL. (Issue #55592)
• Add Portlets and Create New Page icons are now displayed correctly in the My Communities dropdown. Previously the icons were mismatched. (Issue #57217)

• The Oracle Global Deployment Scripts should not be run on the AquaLogic Interaction database as it is not supported (Issue #27272)
• The portalconfig.xml file now correctly describes how to order Authentication Sources. (Issue #51484)
• Uninstalling AquaLogic Interaction before uninstalling the Logging Utilities will no longer cause the uninstaller for the Logging Utilities to fail. (Issue #51543)
• When using Microsoft SQL Server, it is now possible to use a database password that contains the equals '=' or semicolon ';' characters. (Issue #54497)
• The UICI java installer now works on Apache Tomcat 5.5. (Issue #54739)
• The ALI 6.1 MP1 Installer now lists WLS 8.1, 9.1, and 9.2 for auto-deployment. (Issue #55862)
• In the Install guide, under the 'Windows User and Group Requirements' section, the 'Dedicated Windows User' reference has been removed. (Issue #56144)
• The Install Guide now mentions the required import of SearchClusterAdminUI.pte in both the Install and Upgrade sections. (Issue #56405)
• Documentation no longer incorrectly states SQL Server 2005 requires SQL Server 2000 Compatibility 'Mode' instead of 'Level'. (Issue #56523)

• Uploaded or crawled files in the Knowledge Directory that contain '#' in the file name will now have the '#' symbol escaped properly to prevent browsers from assuming that it is an HTML reference. (Issue #50424)
• The Knowledge Directory sort order is now correct when a localized name is assigned to the folder properties. (Issue #52931)
• Users who are explicitly granted Edit (or above) permissions on a Knowledge Directory folder and a document therein will now be able to delete the document without having to belong to a group with Edit (or above) rights. (Issue #53764)
• Folder Sorting in Edit Mode is now consistent with Browse Mode. (Issue #54344)

• Clicking on 'more . . .' link in a Snapshot Query Portlet when editing portlet preferences or previewing the portlet will no longer return an error to the user. (Issue #30485)
• Snapshot Query Portlets will no longer generate an error when users do not have Select permissions on any of the custom properties used for sorting. (Issue #53075)

• When running examinearchive on a large Search Server index, the process will no longer hang or core dump. (Issue #50454)
• Cluster nodes no longer need to be reset after aborting a checkpoint restore (either from the UI or command line). (Issue #55021)
• Search Server cluster checkpoint creation during a time of heavy search indexing will no longer lead to a deadlock on one or more nodes. (Issue #56537)
• Community pages are now searchable after a Search Server rebuild. (Issue #56664)
• On Solaris 10, with the kernel patch 118833-24, the Search Server will no longer report SOCKET_TIMEOUT errors. (Issue #56708)
• If a request transaction delta file is missing in folder where Search Server expect it, indexing will no longer fail and Search Server will no longer to switch into 'read-only' mode as a result. (Issue #58677)

• Users in the Korean locale will not be able to adjust AM/PM when editing jobs. The job will always be saved as AM. (Issue #56275)
• Audit log messages are not deleted the 'Message types to delete' is set to 'All' and the 'Delete Messages and Archives when settings are saved' option is selected. (Issue #57276)
• When indexing a large document set, the Search Cluster UI may display somewhat misleading information in the 'Document Load' bar. Under some circumstances, the 'Document Load' value may be displayed over 100%. This should not be considered an indication of any problem with the Search Cluster. (Issue #58042)
• The Top Best Bet icon will not be displayed even when adding the HTML code for it. (Issue #58216)

• The Portal Diagnostic Tool will not work in Java when you create and register a custom tag library. (Issue #53063)

• The Content Upload Service requires the Document Repository Service to be running before starting. (Issue #56575)
• On the Linux operating system: when using non-English locales, you must run 'export LANG=en_US' before running ptspy.sh to launch the Logging Spy graphical utility. (Issue #58101)

• If the Search component is missing from the AquaLogic Interaction license key, both the diagnostic tool and the application server log will display a warning stating "Missing, invalid, or expired license for AquaLogic Interaction Grid Search-6.1-Search. The search functionality has been disabled." This message can be ignored. If the Search component is missing from the AquaLogic Interaction key, search will be limited to 2 nodes. (Issue #60468)
• The installer may garble the names of the Server Side Includes and Intern Data Connector web extensions. You can use MetaEdit to change the name back to the original names. (Issue #54147)
• The installer may take longer to complete the installation than indicated by the status bar. (Issue #57235)
• On HP-UX: if you set the location where the installer unpacks files from /tmp to another location (this can be done by setting IATEMPDIR environment variable to a directory) the installer will still use approximately 90MB of space from /tmp. Clear out at least 90MB of space from /tmp and shut down any processes on the server that may also be using space on /tmp before running installer. (Issue #58874)
• On HP-UX: Re-installations of all components of the ALI installer occasionally cause the installer to hang. The hang always occurs when the installer is performing the post-installation step of deploying the Image Service. To locate the hanging process, run 'ps -eaf | grep jre142'. This should show at least one process running jre142/bin/IA64N/java in your installation directory. Stop this process and let the installer run to completion. The installer will install everything properly except the Image Service. Run the installer again and choose only the 'Image Service' for installation. The installer should run to completion without hanging and properly install the Image Service. (Issue #58954)
• Updates to the BEA ALI license file will only take effect only after your application server is restarted (Issue #59412)

• Copying a saved query portlet will unlink it from the original snapshot query. (Issue #58282)

• On HP-UX: The JVM on HPUX-IA64 tends to agressively consume and fragment memory. At the same time, the default kernel parameters limit the data segment size to 1GB by default. The combination of these factors can lead to out-of-memory errors. This can affect any Java process, particularly those which use the search client, which include Portal, Automation Server, Collaboration Server and Publisher. We recommend increasing the maximum data segment size to at least 2GB. This can be done using the kctune or sam tools. For instance: /usr/sbin/kctune maxdsiz=2147483648 If memory usage continues to be a problem, it may be necessary to tune the Java heap parameters. In order to minimize heap fragmentation, it is best to set the initial and maximum heap parameters to the same value. For instance -Xms 320m -Xmx 320m (Issue #57256)
• If a search node has been offline for some period of time and several checkpoints have been made while it was offline, the node may be unable to transition into RUN state when started. Because the checkpoints have deleted the old requests, the search server logs will mention that it is looking for an index queue segment file which no longer exists. The server should still come up and stay in standby mode. The node needs to be issued a reset command through the cadmin tool (cadmin reset <node-name>). This reset will clear out the contents of the node’s index, and the node will restore to the most recent checkpoint and catch up from there. (Issue #57464)
• On HP-UX: adding more than one crawler operation to a job will cause both crawlers to fail. The workaround is to associate only one crawler operation with each job. (Issue #57589)
• Search Server index directory and cluster request directory should be excluded from virus scanning, as indicated in product documentation. Under certain circumstances a node may crash if a virus scanner interferes with file I/O. (Issue #57858)
• The Search Cluster Manager allows the administrator to schedule cluster checkpoint operations. Once a checkpoint schedule has been established, you may change the schedule using the same user interface, however there is no facility in the user interface to completely turn off scheduled checkpoints. To completely remove a scheduled checkpoint, you may edit the cluster.ini file, found in the cluster home directory. Find the line with the parameter RF_CHECKPOINT_SCHEDULE and comment it out with a '#' character or delete it completely. (Issue #58443)
• On HP-UX: The Search Update Agent may occasionally fail to complete. The workaround is to run the Search Update Agent again. (Issue #58560)
• No search results are shown for 'Reference' property type even if it is set to be searchable. (Issue #58933)
• The Search Update Agent adds errors to the job log about trying to update an object with ClassID 514. This object referred to is a MyPage object that is not intended to be indexed and can be ignored. Depending on the environment, this message may be written many times in each job log, potentially filling up the job log table. A workaround is to change the Logging Level for the Search Update Agent job to Silent. (Issue #62413)

• The Known Issues section has been updated with a new issue, #62413. September 4, 2007

 

 

 

 

 

 

 

 

 

 

 

 

Refer to the Interoperability page in the Product Center at support.plumtree.com for the latest information on supported operating systems, application servers, databases, and browsers.

AquaLogic Interaction 6.1 MP1 supports the following:

• Operating Systems: AIX 5.3 (on POWER3, POWER4, POWER5); HP-UX 11i V2 (on Itanium); RedHat Enterprise Linux 4 update 3 (on x86); Solaris 8, 9, and 10 (on SPARC); SUSE Enterprise Linux 9 (on x86); Windows Server 2003 SP1 or R2, SP2 (on x86)
  
  Solaris 8 requires these patches:
  • 108434-21 (32-Bit Shared library patch for C++)
  • 108435-21 (64-Bit Shared library patch for C++)
  
  Solaris 9 requires these patches:
  • 111711-15 (32-bit Shared library patch for C++)
  • 111712-15 (64-Bit Shared library patch for C++)
  
  
• Databases: Microsoft SQL Server 2000 SP3a or SP4, Microsoft SQL Server 2005 (with 2000 Compatibility Level), Microsoft SQL Server 2005 SP2 (with 2000 Compatibility Level), Oracle 9i (9.2.0.4 and above), Oracle 10g (10.1.0.3 and above), Oracle 10g R2 (10.2.0.1 and above)
  
• Application Servers:
  • AIX:
    • IBM WebSphere 6.0.1 with IBM (32-bit) SDK 1.4.2 SR6, IBM WebSphere 6.1 with IBM (32bit) SDK 5.0 SR2
  • HP-UX:
    • BEA WebLogic 9.2 with HP-UX JDK 5.0.03 with Java HotSpot Server VM (32-bit), BEA WebLogic 9.2 MP1 with HP-UX JDK 5.0.04 (32-bit) with Java HotSpot Server VM
  • Red Hat Enterprise Linux:
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with BEA JRockit 5.0 (R26.0.0) JDK (32-bit); BEA WebLogic 9.2 MP1 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit), BEA WebLogic 10 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit) and all later 5.0 releases of JRockit
  • Solaris:
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06; BEA WebLogic 9.2 MP1 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06; BEA WebLogic 10 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06
    • IBM WebSphere 6.1 with IBM (32bit) SDK 5.0 SR2
  • SUSE Enterprise Linux:
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with BEA JRockit 5.0 (R26.0.0-189_CR269406) JDK (32-bit); BEA WebLogic 9.2 MP1 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit), BEA WebLogic 10 with BEA JRockit 5.0 (R26.4.0 + CR302700) JDK (32-bit) and all later 5.0 releases of JRockit
  • Windows:
    • Microsoft IIS 6.0 with .NET Framework 1.1 SP1
    • Apache Tomcat 5.0.28 with Sun SDK 1.4.2_11
    • BEA WebLogic 8.1 SP5 with BEA JRockit 1.4.2_11; BEA WebLogic 9.1, 9.2 with Sun Java 2 JDK 5.0 (32-bit) with Java HotSpot Client VM, version 1.5.0_06; BEA WebLogic 9.2 MP1 with Sun Java 2 JDK 5.0 (32-bit) with Java HotSpot Client VM, version 1.5.0_06; BEA WebLogic 10 with Sun Java 2 JDK 5.0 (32-bit) with the Java HotSpot Client and Server VMs, version 1.5.0_06
• Browsers: Administrative Use - Firefox 2.0, Internet Explorer 6.0, 7.0; Browsing Use - Firefox 1.5, 2.0; Internet Explorer 6.0, 7.0; Netscape 7.2, 8.0; Safari 2.0 (Mac only)

Note: AquaLogic Interaction 6.1 MP1 does not support IPv6. If a system is incorrectly configured to enable IPv6, errors may occur with your system. You must ensure that IPv6 is not enabled on your system before installing AquaLogic Interaction 6.1 MP1. For help verifying whether IPv6 is enabled, please contact Customer Support.

Refer to the Interoperability page in the Product Center at support.plumtree.com for the latest information on supported BEA products.

At the time of release, AquaLogic Interaction 6.1 MP1 supports the following:

• AquaLogic Interaction Collaboration 4.2 MP1, Publisher 6.4, Studio 2.2 MP1, and Analytics 2.1
• AquaLogic Interaction Content Service - Windows Files 6.1, Identity Service - Active Directory 6.3, and Identity Service - LDAP 2.2
• For information on additional products, refer to the Interoperability page.

As more platforms are supported, additional upgrade paths to the new supported platforms will be included.

AquaLogic Interaction 6.1 MP1 Patch 1 only supports upgrades from AquaLogic Interaction 6.1 MP1.

AquaLogic Interaction 6.1 MP1 currently supports upgrade directly from the following previous releases:

• AquaLogic Interaction 6.1.0.1 and 6.1.
• Plumtree Foundation 6.0 and 6.0 SP1.
• Plumtree Corporate Portal 5.0.2, 5.0.3, 5.0.4, and 5.0.5 on Windows.
• Plumtree Corporate Portal 5.0.4J and 5.0.5J Solaris.
• Plumtree Corporate Portal 4.5WS SP3 on Windows.

The following product documentation is available for download at edocs.bea.com/alui/ali/docs61:

Document Title and File Name	Document Description
Installation and Upgrade Guide for BEA AquaLogic Interaction 6.1 install.pdf	Describes how to install and upgrade the components of AquaLogic Interaction 6.1.
Administrator Guide for BEA AquaLogic Interaction 6.1 admin.pdf	Describes how to use, manage, and administer AquaLogic Interaction 6.1.
Online Help	Describes how to use the user interface features delivered with AquaLogic Interaction 6.1.
Installation Worksheet for AquaLogic Interaction 6.1 - Unix Worksheet_ALI_Unix.pdf	Print this document and use it to gather the host computer, portal, database, and additional information that you need to install and configure all of the various AquaLogic Interaction components.
Installation Worksheet for AquaLogic Interaction 6.1 - Windows Worksheet_ALI_Windows.pdf	Print this document and use it to gather the host computer, portal, database, and additional information that you need to install and configure all of the various AquaLogic Interaction components.

The release package for AquaLogic Interaction no longer contains a "ThirdParty" folder with a copy of Apache Tomcat and Apache HTTP Server. The correct locations to obtain a copy of Apache Tomcat and Apache HTTP Server for your OS are:

• Apache Tomcat: http://tomcat.apache.org
• Apache HTTP Server: http://httpd.apache.org

Additionally, the Sun Java Runtime Environment can be obtained from http://java.sun.com.

• It is now possible to clear a reference property value from an object. (Issue #49362)
• When using the File Upload to submit PDF files on a .NET portal, certain date attributes from the PDF can now be mapped to the document properties. (Issue #52252)
• Document title and description changes made from the Approve Directory Content utility are now saved. (Issue #53015)
• The Audit Manager will delete messages after selecting the 'Delete Messages and Archives when settings are saved?' checkbox and saving the Audit Manager settings. (Issue #53238)
• Incorrect physical memory size is no longer reported in the Search Server Cluster Status page. (Issue #54100)
• From the Administration directory, editing a group with a large number of members (16,000 or more) will no longer cause a high amount of CPU utilization. (Issue #57861)
• On Windows, clicking the 'Finish' button in Cluster Manager no longer returns you back to the Home Page. (Issue #58040)

• The following error is no longer displayed in the job log and Logging Spy when refreshing documents: Error in function PTCard.UnlockObject (): -2147205112 - Object not in locked state. (-2147205112) (Issue #52859)
• Crawl jobs will no longer hang when a crawler has a large number of Knowledge Directory folders with filters assigned. (Issue #53106)
• Knowledge Directory documents are now automatically approved when the target document is a linked card in the Documentum Content Server. (Issue #53324)
• Crawler and Document Refresh Agent operations will now remove Knowledge Directory documents that have been deleted from their original repositories. (Issue #53520)
• Two concurrent instances of the same mirror crawler will no longer delete each others cards. (Issue #54819)
• Documentum and Windows File Crawlers will no longer re-import all Knowledge Directory cards if 'Refresh ACLS' is selected. (Issue #56225)

• JavaScript with line breaks after a plus symbol (+) are now transformed properly when gatewayed. (Issue #53221)
• In the OnAfterLogin PEI, when trying to construct the logout URL redirect, the current user will be logged off properly. (Issue #53270)
• Using the adaptive tag <pt:ptui.loginauthsource pt:defaultselection='' pt:hidden='true'/> now selects the ALUI User Database as the authentication source. (Issue #54220)
• When using Firefo, when gatewaying an HTML form with inline-refresh enabled, if the form is posted empty, it will no longer generate an error. (Issue #54300)
• There is now a UICI for the AIX version of AquaLogic Interaction. (Issue #55945)
• With SSO enabled, if you write any portlet or login tag with the <pt:standard.loginlink> tag, it will now redirect to the SSO login page instead of the standard login page. (Issue #58441)

• The LocalMachineOnly setting in serverconfig.xml functions correctly now. If it is set to 'true', Logging messages will not be sent over the network. (Issue #36806)
• Using a Remote Server with a hostname that contains a dash (-) and a number (e.g. host-01) will no longer cause an error. (Issue #46871)
• The Guest user can no longer use the site map to list all users. (Issue #48648)
• Logging information from Search is now fully enabled on AIX or Solaris. (Issue #49527)
• Users will no longer be prompted for logins multiple times when using Security Mode 1 and accessing Collaboration or Publisher Web Folders. (Issue #50181)
• Mandatory Community tabs now highlight the entire button on mouse over. (Issue #52965)
• The navigation dropdown menus have been fixed so that certain HTML SELECT elements will no longer display through the background. (Issue #54689)
• Activity Server portlets are now able to generate Image Server and JS base URLs when the PT-Imageserver-URI string is a relative URL. (Issue #55592)
• Add Portlets and Create New Page icons are now displayed correctly in the My Communities dropdown. Previously the icons were mismatched. (Issue #57217)

• The Oracle Global Deployment Scripts should not be run on the AquaLogic Interaction database as it is not supported (Issue #27272)
• The portalconfig.xml file now correctly describes how to order Authentication Sources. (Issue #51484)
• Uninstalling AquaLogic Interaction before uninstalling the Logging Utilities will no longer cause the uninstaller for the Logging Utilities to fail. (Issue #51543)
• When using Microsoft SQL Server, it is now possible to use a database password that contains the equals '=' or semicolon ';' characters. (Issue #54497)
• The UICI java installer now works on Apache Tomcat 5.5. (Issue #54739)
• The ALI 6.1 MP1 Installer now lists WLS 8.1, 9.1, and 9.2 for auto-deployment. (Issue #55862)
• In the Install guide, under the 'Windows User and Group Requirements' section, the 'Dedicated Windows User' reference has been removed. (Issue #56144)
• The Install Guide now mentions the required import of SearchClusterAdminUI.pte in both the Install and Upgrade sections. (Issue #56405)
• Documentation no longer incorrectly states SQL Server 2005 requires SQL Server 2000 Compatibility 'Mode' instead of 'Level'. (Issue #56523)

• Uploaded or crawled files in the Knowledge Directory that contain '#' in the file name will now have the '#' symbol escaped properly to prevent browsers from assuming that it is an HTML reference. (Issue #50424)
• The Knowledge Directory sort order is now correct when a localized name is assigned to the folder properties. (Issue #52931)
• Users who are explicitly granted Edit (or above) permissions on a Knowledge Directory folder and a document therein will now be able to delete the document without having to belong to a group with Edit (or above) rights. (Issue #53764)
• Folder Sorting in Edit Mode is now consistent with Browse Mode. (Issue #54344)

• Clicking on 'more . . .' link in a Snapshot Query Portlet when editing portlet preferences or previewing the portlet will no longer return an error to the user. (Issue #30485)
• Snapshot Query Portlets will no longer generate an error when users do not have Select permissions on any of the custom properties used for sorting. (Issue #53075)

• When running examinearchive on a large Search Server index, the process will no longer hang or core dump. (Issue #50454)
• Cluster nodes no longer need to be reset after aborting a checkpoint restore (either from the UI or command line). (Issue #55021)
• Search Server cluster checkpoint creation during a time of heavy search indexing will no longer lead to a deadlock on one or more nodes. (Issue #56537)
• Community pages are now searchable after a Search Server rebuild. (Issue #56664)
• On Solaris 10, with the kernel patch 118833-24, the Search Server will no longer report SOCKET_TIMEOUT errors. (Issue #56708)
• If a request transaction delta file is missing in folder where Search Server expect it, indexing will no longer fail and Search Server will no longer to switch into 'read-only' mode as a result. (Issue #58677)

• Users in the Korean locale will not be able to adjust AM/PM when editing jobs. The job will always be saved as AM. (Issue #56275)
• Audit log messages are not deleted the 'Message types to delete' is set to 'All' and the 'Delete Messages and Archives when settings are saved' option is selected. (Issue #57276)
• When indexing a large document set, the Search Cluster UI may display somewhat misleading information in the 'Document Load' bar. Under some circumstances, the 'Document Load' value may be displayed over 100%. This should not be considered an indication of any problem with the Search Cluster. (Issue #58042)
• The Top Best Bet icon will not be displayed even when adding the HTML code for it. (Issue #58216)

• The Portal Diagnostic Tool will not work in Java when you create and register a custom tag library. (Issue #53063)

• The Content Upload Service requires the Document Repository Service to be running before starting. (Issue #56575)
• On the Linux operating system: when using non-English locales, you must run 'export LANG=en_US' before running ptspy.sh to launch the Logging Spy graphical utility. (Issue #58101)

• If the Search component is missing from the AquaLogic Interaction license key, both the diagnostic tool and the application server log will display a warning stating "Missing, invalid, or expired license for AquaLogic Interaction Grid Search-6.1-Search. The search functionality has been disabled." This message can be ignored. If the Search component is missing from the AquaLogic Interaction key, search will be limited to 2 nodes. (Issue #60468)
• The installer may garble the names of the Server Side Includes and Intern Data Connector web extensions. You can use MetaEdit to change the name back to the original names. (Issue #54147)
• The installer may take longer to complete the installation than indicated by the status bar. (Issue #57235)
• On HP-UX: if you set the location where the installer unpacks files from /tmp to another location (this can be done by setting IATEMPDIR environment variable to a directory) the installer will still use approximately 90MB of space from /tmp. Clear out at least 90MB of space from /tmp and shut down any processes on the server that may also be using space on /tmp before running installer. (Issue #58874)
• On HP-UX: Re-installations of all components of the ALI installer occasionally cause the installer to hang. The hang always occurs when the installer is performing the post-installation step of deploying the Image Service. To locate the hanging process, run 'ps -eaf | grep jre142'. This should show at least one process running jre142/bin/IA64N/java in your installation directory. Stop this process and let the installer run to completion. The installer will install everything properly except the Image Service. Run the installer again and choose only the 'Image Service' for installation. The installer should run to completion without hanging and properly install the Image Service. (Issue #58954)
• Updates to the BEA ALI license file will only take effect only after your application server is restarted (Issue #59412)

• Copying a saved query portlet will unlink it from the original snapshot query. (Issue #58282)

• On HP-UX: The JVM on HPUX-IA64 tends to agressively consume and fragment memory. At the same time, the default kernel parameters limit the data segment size to 1GB by default. The combination of these factors can lead to out-of-memory errors. This can affect any Java process, particularly those which use the search client, which include Portal, Automation Server, Collaboration Server and Publisher. We recommend increasing the maximum data segment size to at least 2GB. This can be done using the kctune or sam tools. For instance: /usr/sbin/kctune maxdsiz=2147483648 If memory usage continues to be a problem, it may be necessary to tune the Java heap parameters. In order to minimize heap fragmentation, it is best to set the initial and maximum heap parameters to the same value. For instance -Xms 320m -Xmx 320m (Issue #57256)
• If a search node has been offline for some period of time and several checkpoints have been made while it was offline, the node may be unable to transition into RUN state when started. Because the checkpoints have deleted the old requests, the search server logs will mention that it is looking for an index queue segment file which no longer exists. The server should still come up and stay in standby mode. The node needs to be issued a reset command through the cadmin tool (cadmin reset <node-name>). This reset will clear out the contents of the node’s index, and the node will restore to the most recent checkpoint and catch up from there. (Issue #57464)
• On HP-UX: adding more than one crawler operation to a job will cause both crawlers to fail. The workaround is to associate only one crawler operation with each job. (Issue #57589)
• Search Server index directory and cluster request directory should be excluded from virus scanning, as indicated in product documentation. Under certain circumstances a node may crash if a virus scanner interferes with file I/O. (Issue #57858)
• The Search Cluster Manager allows the administrator to schedule cluster checkpoint operations. Once a checkpoint schedule has been established, you may change the schedule using the same user interface, however there is no facility in the user interface to completely turn off scheduled checkpoints. To completely remove a scheduled checkpoint, you may edit the cluster.ini file, found in the cluster home directory. Find the line with the parameter RF_CHECKPOINT_SCHEDULE and comment it out with a '#' character or delete it completely. (Issue #58443)
• On HP-UX: The Search Update Agent may occasionally fail to complete. The workaround is to run the Search Update Agent again. (Issue #58560)
• No search results are shown for 'Reference' property type even if it is set to be searchable. (Issue #58933)
• The Search Update Agent adds errors to the job log about trying to update an object with ClassID 514. This object referred to is a MyPage object that is not intended to be indexed and can be ignored. Depending on the environment, this message may be written many times in each job log, potentially filling up the job log table. A workaround is to change the Logging Level for the Search Update Agent job to Silent. (Issue #62413)

• The Known Issues section has been updated with a new issue, #62413. September 4, 2007

• The installer doesn't create all the necessary search links on AIX. (Issue #55825)
• Expensive query at the beginning of a Search Repair or Search Update can cause a pulse timeout and/or stall the operation. The operation fails with the message "operation supporting pulse is unresponsive". (Issue #55485)
• In some cases, the indexing back-end can deadlock and cause entire cluster to go into stall, with the exception of the deadlocked node. (Issue #55708)
• Search server heartbeats are mistakenly logged as incomplete in some cases. (Issue #55322)
• The Automation server does not gracefully call the shutdown to the DTCMCWS thread when the Card Refresh Agent is run. (Issue #54522)
• Repeatedly calling the Approve Directory Content significantly impacts the performance of the editor. (Issue #53012)

• AquaLogic Interaction 6.1 now includes support for additional operating systems, databases, and application servers.
• AquaLogic Interaction 6.1 now includes clustered search for performance, fault-tolerance, and scalability.
• AquaLogic Interaction 6.1 now includes alphabetic sorting available in all supported locales.
• AquaLogic Interaction 6.1 now includes server-side page compression for all portal pages.

• Non-ASCII characters in URLs are properly handled. (Issue #54184)
• Experience definitions can now be modified with relative image server URLs. (Issue #47416)
• If the experience definition editor is not closed gracefully (for example, other than with the finish or cancel button), the portal will no longer experience timeouts when attempting to access other experience definitions. (Issue #51426)
• Objects with IDs greater than 1,000,000 can be edited with the multiple object security matrix editor. (Issue #51452)
• Caching can now be set beyond 15 minutes. (Issue #52937)
• Document refresh settings are properly maintained after editing the document properties. (Issue #52291)
• If a portlet is created from the Add Portlets page of a community, that portlet is now automatically added to that community page. (Issue #49912)
• Clicking "Join This Community" from the dropdown menu properly adds the community to the membership list. (Issue #51667)
• Snapshot queries are now properly migrated. (Issue #52364)
• Unchanged source documents will no longer be refreshed when the refresh agent is run. (Issue #53903)

• The Create New Page option no longer appears in 508 mode. (Issue #47764)
• In 508 mode, the Edit tool tip includes the name of the portlet. (Issue #45620)
• In 508 mode, the Go Back link now properly directs users to the last page. (Issue #48109)
• In 508 mode, the federated search help link directs to the proper help page.
• In 508 mode, the "How to join a community", "How to navigate inside a community", and "How to search for a community" help links direct to the proper help pages. (Issue #48197)
• In 508 mode, the "My Home, My Account, Logging Off, Help, and Searching" help link directs to the proper help page. (Issue #48203)
• In 508 mode, the "Access Keys" help link directs to the proper help page. (Issue #48217)
• In 508 mode, the "Browse Directory" help link directs to the proper help page. (Issue #48301)
• In 508 mode, the federated search key stroke is now ALT+F rather than ALT+S. (Issue #49345)
• In 508 mode, custom document icons/UUIDs get the correct alt text. (Issue #51214)
• JavaScript is no longer added in 508 mode. (Issue #53014)

• Web crawls can accept pages with invalid charsets. (Issue #53010)
• Gatewayed web crawls of sites protected by basic authentication properly display the names and properties of the individual pages. (Issue #51675)
• More PDF files are properly indexed with the new search system. (Issue #51653)

• HTTP Post method properly passes parameters to non-gatewayed pages. (Issue #52224)
• Javascript it now properly transformed. (Issue #52878)
• The portal properly handles portlets with HREF attributes that contain a space. (Issue #53013)
• Downloads of Collaboration templates now function properly in all locales. (Issue #51829)
• Javascript is properly handled when stack trace errors are returned. (Issue #47534)
• The HTTP load balancer properly handles re-adding servers to a cluster of remote servers. (Issue #50135)

• Column customizations via CSS are now fully functional. (Issue #50030)
• Standalone upgrades of the Automation Service function properly. (Issue #52266)
• The Configuration Manager works properly with Oracle RAC systems. (Issue #47694)
• The proxy server bypass settings now function properly. (Issue #50012)
• N_NARROW_COL_WIDTH and N_PORTLET_PADDING are now CSS styles that can be configured using the cssmill. (Issue #52836)
• Running SQL script on Oracle 10g R2 reports ORA-30041 error when granting quota on PLUMTEMP. (Issue #51234)
• Running SQL script on Oracle 10g reports ORA-01442: column to be modified to NOT NULL is already NOT NULL. (Issue #52298)

• Online help in Japanese is now properly translated. (Issues #54149 and #54750)

• AVI and other media files download more reliably from the Knowledge Directory. (Issue #50732)

• The Community Links portlet does not display a "More..." link in the portlet when there are more links in the community knowledge directory folder than the "Set Maximum number of links to display" setting. (Issue #38463)

• If an AWS is disabled and has an appropriate disabled message, the message will not be displayed when a user tries to log in using that AWS. The user will receive a generic error message instead. Also, the Administrator will not be able to edit the Authentication Source whose corresponding Web service is disabled. If a Web service object for a crawler is disabled, administrators may not be able to edit corresponding data sources and crawlers. (Issue #15755)
• Mandatory portlets are not listed for users in 508 mode. (Issue #42952)
• When editing "User Information - Property Map", if two properties are mapped to a same User Information Attribute, a vague error message will be displayed when Finish is clicked. The message is sometimes "Unable to store the UserInf map each user information element to one and only one property." The administrator must find the duplicate and resolve that before saving the Map. (Issue #16082)
• When an administrative folder containing users is copied, all objects other than users will be copied. There is no warning or error message. Users cannot be copied. (Issue #20116)
• If you create a new community group in a subcommunity, it will appear on the Security page of the subcommunity with Read access. If you then change the settings on the Security page of the parent community, these changes will be propagated down to the subcommunity (if the user chooses to propagate the changes). Therefore, if the community group was not on the Security page of the parent community, it will no longer appear on the Security page of the subcommunity. The community group must therefore be added back manually to the Security page of the subcommunity. (Issue #33288)
• The operation may fail when trying to delete a large number of objects at once. (Issue #42399)
• Dynamic Membership statements with values that contain certain characters such as a semicolon may cause javascript errors on the page. (Issue #46462)
• The system health monitor may erroneously report that there is an issue with Process even if there are no issues. (Issue #49704)
• Submitting community preferences will not automatically refresh the page. (Issue #51435)
• It is not possible to edit a document's security with the security matrix if the document's ID is greater than 999,999. It is still possible to edit the security of the document in the document's editor itself. (Issue #51452)

• Reference type Properties are read-only in the User Profile Editor in Assistive Technology mode. (Issue #36042)
• In Assistive Technology mode, clicking "Go Back" on the Help page from the second page of the Document Submission editor takes you to your default MyPage, rather than to the first page of the editor. (Issue #37080)
• The General Information Portlet in users' profiles may not be visible in 508 mode. (Issue #44996)
• Search results always click through in the main window and not in a new window regardless of the Document Display Settings when in 508 mode. (Issue #51252)

• The Web crawler is unable to read the value of a META tag if it is the very last tag in an HTML document. This is unusual since META tags generally occur in the HEAD section of the document and are followed by the content of the document in the BODY section. (Issue #37143)
• If an owner of a crawler does not have sufficient access to a folder in the target taxonomy, he/she will only receive an "insufficient privilege" message in the job log if that is the only folder you can go into. (Issue #43896)
• Job logs fail to download on Solaris systems. (Issue #49252)
• Mirrored file crawlers do not properly handle NT folders that have a space as the 63rd character in the name. The name will get truncated. (Issue #50730)
• If all of the search nodes are offline, it could take several minutes to open a crawler editor. Ensure that at least one search node is online when editing crawlers. (Issue #54973)
• You may need to modify your external operations shell scripts for Unix-based deployments of AquaLogic Interaction. (Issue #55061)
  
  Previous to 6.1, external operations ran in the same environment as the Automation Server process. In 6.1, a separate Unix shell environment is spawned for each external operation. As a result, some environment variable will need to be set in each external operation shell script. For example, if your script utilizes 'dirname', you should ensure that 'dirname' is in the PATH.

• The portal stores session states in cookies. If you store many session states, this may go over the maximum number of session cookies allowed in IE and may cause some cookies such as the login cookie to be removed from the session. When developing on the portal, be careful to minimize the number of session states stored on a page. (Issue #47642)
• You cannot set or get session preferences that contain unusual characters such as "?". (Issue #47708)
• Reading a string field from an IPortalSearchResult converts a full width number to a half width number. The value is stored correctly in the DB, and is returned correctly using IDocumentProperties. (Issue #51243)

• If the maximum upload size specified in ALI Collaboration is higher than the maximum upload size specified in AquaLogic Interaction, documents larger than the maximum size specified in AquaLogic Interaction may never complete. (Issue #52082)
• Some versions of IE6 on Windows 2000 have a bug where they do not render compressed pages properly. End users should upgrade to the latest version of IE6 if compression is enabled. (Issue #54230)
• Some versions of IE6 (6.0.2800.1106 Update Version SP1; Q833989) produce a javascript error when the recommended portlet functionality is used. End users should upgrade to the latest version of IE6. (Issue #54379)
• You may experience issues downloading Excel spreadsheets from Analytics reports using the Microsoft IE6 browser. (Issue #54827)
  
  This problem is directly related to one or more of the following bugs in IE6 SP1 resolved by the KB articles listed below. Please require users to either upgrade their browsers by installing IE6 SP2 or patching with the specific downloads listed in the KB articles.
  
  http://support.microsoft.com/kb/871205
  http://www.microsoft.com/downloads/details.aspx?FamilyID=0c7055c3-0bec-4042-b6aa-3443739c66ab&DisplayLang=en
  http://support.microsoft.com/?kbid=823386
  
  The other workaround is to specifically disable compression for Excel spreadsheets in the httpcompression.xml setting file, located in %PT_HOME%/settings/portal.

• On Windows servers, the AquaLogic Interaction portal or Administrative Portal Server should not be installed on the same server as an Oracle database. This is because both AquaLogic Interaction and Oracle install OutsideIn libraries and reference them via the PATH environment variable. The AquaLogic Interaction portal requires its version of the OutsideIn libraries. If Oracle has been installed first, its reference to its OutsideIn libraries will be listed first which would cause the AquaLogic Interaction portal to fail.
  
  If the AquaLogic Interaction portal or Administrative Portal Server move AquaLogic Interaction's reference to its OutsideIn libraries in front of Oracle's in the PATH environment variable. In order to do this, edit the PATH environment variable and move AquaLogic Interaction's reference to the its OutsideIn libraries, %OUTSIDEIN_PATH%, to the front of the PATH environment variable. Note, doing this may disable some features of Oracle that depend on its version of the OutsideIn libraries. (Issue #52885)
• The Content Upload service must be started AFTER the Document Repository Service. (Issue #49725)
• The ALI Logging Utilities product is not properly uninstalled when AquaLogic Interaction is uninstalled. (Issue #53950)
• The installer may garble the names of the "Server Side Includes" and "Intern Data Connector" web extensions. You can use MetaEdit to change the name back to the original names. (Issue #54147)
• When you run the Portal Configuration Manager to test the DB connection on an upgraded portal installation, you must enter the DB Login password.
• The portal upgrade does not properly upgrade the DB password for upgraded Solaris systems. You must run the Portal Configuration Manager and re-enter the DB password before starting the upgraded portal on Solaris systems. (Issue #49509)
• Values entered in the Portal Configuration Manager will always be saved if you test the DB connection. This is true even if you click Cancel after testing the connection. (Issue #47736)
• Some of the CSS classes have changed between Plumtree Corporate Portal 5.0 and AquaLogic Interaction 6.x. This could cause some customizations or portlets that depend on those classes to behave differently in AquaLogic Interaction 6.x than in Plumtree Corporate Portal 5.0. (Issue #47594)
• The Administration link may not work correctly in Security Mode 3. If there are security issues (page cannot be accessed), a workaround is to set the AdminSiteBaseURL in x_config.xml (j_config.xml for Java portals, n_config.xml for .NET portals) to be an absolute URL (i.e. http://mycomputer.mycompany.com/portal/server.pt). (Issue #28121)
• AquaLogic Interaction 6.x no longer supports Windows Authentication to connect to MS SQL Server databases.
• The Oracle Global Deployment Scripts should not be run on the AquaLogic Interaction database as it is not supported. (Issue #27272)
• When using a non-existent parameters file for the silent install a stack trace is shown instead of a simple error message. The specific error is an InvocationTargetException. To solve this problem type in the correct, full path to the silent properties file you wish to use. (Issue #47469)
• For the UICI, you must manually specify the default install path for Linux deployments because the installer does not utilize the Linux HOME environment variable. (Issue #48499)
• When the portal is auto-deployed onto WebSphere, the application name is "portal". However, when the portal is manually deployed onto WebSphere, the application name is "PlumtreePortal". (Issue #40641)
• When running .NET in IIS 6.0, the AppPool Config Options should be disabled when running AquaLogic Interaction. (Issue #47702)
  
  To disable the AppPool Config Options:
  1. Open IIS Manager.
  2. Expand the Local Computer and Application Pools.
  3. Right-click on the DefaultAppPool and select Properties.
  4. On the Recycling Tab make sure that the Recycle Worker Process options are disabled.
  5. On the Performance Tab, make sure the Idle Timeout option is disabled.
  6. Click OK to save changes.
     
• The port number displayed in the "API Service URL" panel of the installer always defaults to 11905 regardless of the port specified in the "API Service - Application Port" panel when installing the API Service. You must manually enter the correct port number in the "API Service URL" panel. (Issue #47756)
  
  Note: You will only see the "API Service URL" panel when installing the API Service and you will only see the "API Service - Application Port" panel when installing a portal or Administrative Portal.
  
  If you did not change the port number in the "API Service URL" panel to match the port in the "API Service - Application Port" panel, you can change it after installing the portal and running the DB scripts by going to the "Portal Settings" utility and changing the port specified in the "SOAP Server URL".
• When upgrading to AquaLogic Interaction 6.x from a non-English version of a previous version of the portal, the default names of many objects will revert to English. (Issue #47877)
• The objects that make up the Content Upload service are named "File Upload..." when imported into the portal. You can change the names as needed. (Issue #47711)
• If you are utilizing the File Upload Web Service in the 5.0.x version of the portal, the portal installer does not properly upgrade the URL of the remote server object of the File Upload Web Service to the 6.1 Content Upload Service. You must manually update the URL. In order to resolve this issue:
  1. Install the new 6.1 Content Upload Service.
     • Record the URL and port number.
     • Do NOT import the migration package to the portal. The portal objects created by this package will not be able to access previously uploaded documents because those documents will still be associated with the data source based on the old File Upload Web Service. The goal is to switch the underlying version of the upload Web service that this data source uses and not replace the data source itself with a new one.
  2. Locate the old File Upload Remote Server object in the administration hierarchy of Plumtree Foundation 6.0.
  3. Edit the remote server URL to match the URL and port number of the newly installed 6.0 Content Upload Service. This will switch the version of the upload service that the original data source object utilizes while maintaining the data source's association with previously uploaded documents.
• For the UICI, ANT_OPS and JAVA_OPTS are not set in unixvars.sh. (Issue #48519)
• For the UICI, the portaluiinfrastructure project has an invalid Default output folder for the Java Build Path. (Issue #48573)
• There is a general issue that affects all applications that are deployed on an IBM WebSphere 6.0 instance that has been upgraded to 6.0.1. Upon migrating to WebSphere 6.0.1 with the IBM HTTP Server and installing the plugin, applications are not displayed on the administrative console. Restarting IBM WebSphere will start those applications even if those applications are not displayed. IBM plans a fix for this in fixpack 6.0.1.2. Please visit the IBM support site for more details.
• For the UICI, custom .jar files should be manually put in the portal.war/ear if ALI Analytics is also installed. (Issue #51678)
• For the UICI, on Solaris configurations, the PATH environment variable may need to be moved when editing UI customizations. This is because Solaris comes with its own version of Java. In order to resolve the situation:
  • After the UICI has completed, edit plutree_ui_source/unixvars.sh in a text editor.
  • In the line with the export PATH command, move ${PATH} to the end of the line and save the file. For example, the revised command might look like this: export PATH=/opt/apache-ant/bin:/opt/j2sdk/bin:${PATH}
  • Source the modified unixvars.sh as you normally would when editing UI customizations.
• Auto-deployment fails on any version of WebLogic Server more recent than 8.1. It is recommended that administrators manually deploy the .EAR (not the .WAR) in order to complete installation on this application server. (Issue #55102)
• Some products install conflicting versions of icu 3.0 into \Windows\system32. When conflicting versions of icu are loaded before the version distributed with the search client, the portal will fail to start up and ptverify will report errors. The problem can be resolved by searching for and removing icu*30.dll from \Windows\system32. Products with known conflicts include Crystal Reports XI, WebEx, and possibly other web client software. (Issue #56359)

• In an exported Excel file, the time format follows the remote server's system locale instead of the client's locale. (Issue #47883)

• Security changes are not properly propagated for changes to the Knowledge Directory because propagation jobs are not being created. (Issue #49694)
• With Symantec Client Firewall installed, cannot view PDFs in the Knowledge Directory. (Issue #51038)

• When creating a page layout, you may receive an error if you collapse the tips after adding a portlet. (Issue #34611)
• You cannot edit the access level of a group to a subcommunity if the group has Admin access to the parent community (regardless of the group's access level on the subcommunity). To workaround this, edit the group's access level on the parent community and propagate the changes to the objects in that community. (Issue #37466)

• No unresolved dependencies are reported after deleting all dependencies after uploading a package when some of the dependencies don't exist on the importing portal. (Issue #42402)
• Migrating a community links portlet does not migrate the folder of links correctly. (Issue #27145)
• The migration import utility allows the selection of a web service of the incorrect type when resolving dependencies. (Issue #42401)
• When migrating activity rights, there may be a FolderID Exception in the migration log. This can be ignored. (Issue #46079)
• The migration utility only stores two digits for year values. Therefore some dates may be migrated improperly if they are in the distant past or future. (Issue #46941)
• When exporting a folder with export dependencies = true and export parent folders = false, the entire folder hierarchy is exported. It is recommended that the export parent option not be used when migrating folders or communities. (Issue #47920)
• When exporting communities with export dependencies = true, all objects stored in the community folder will be exported recursively down the object hierarchy. (Issue #47953)

• The online help menu is not expanded when accessing the help in Safari browsers. (Issue #47705)

• When expanding a list of users on the administration site, if there are a large number of users (for example, 1000), the portal can respond very slowly or even time out. This can usually be solved by optimizing the database. (Issue #27566)

• If a mandatory reference property is mapped to a document folder, the Properties&Names page does not properly enforce that the mandatory property has a value selected. Therefore when a user is creating or editing an object with a mandatory reference property, if no value has been set for that property, no error notification will be displayed when closing the editor. (Issue #26210)
• There are no scroll bars in the help file page of AWS/PWS wizards. The workarounds are as follows:
  o If you highlight the text, you can drag the mouse down and then the rest of the text will show.
  o You can press CTRL-N to open the window in a new window, which will have scroll bars. (Issue #27920)
• If you open two editors for the same portlet at the same time in the same session, changes made in one editor overwrite changes made in the other. (Issue #34922)
• In the content source editor, some of the information, for example the state, is not maintained when switching between pages of the editor. (Issue #43436)
• CSS files will be transformed by Web services even if the "Transform CSS Files" option is not checked. (Issue #50528)

• After importing the Search Cluster Manager migration package, you must log off and log back in to the portal in order to see the utility. (Issue #54901)
• If you abort a checkpoint restore either by the UI or command line, you should restart the Search Service. (Issue #55035)
• After aborting a named restore completely (see previous release note), the cluster should end up with all nodes in RUN mode, the cluster state is "complete", and all nodes report a docset TID equal to the TID of the restored checkpoint. You can confirm the detailed cluster status by running "cadmin status --verbose". If the nodes are not in RUN mode, or the cluster status is not "complete", you should restart the nodes and confirm the correct status. (Issue #55035)
• If network communication errors happen between search nodes during certain phases of the repartitioning process, the nodes may be left in an inconsistent state. The expected state after a successful repartition is that all nodes transition into RUN mode on the new topology. The expected state after a failed repartition is that all nodes transition back into RUN mode on the previous (non-repartitioned) topology. If the nodes end up in a persistent STALL, READONLY, or STANDBY state, you will need to revert to the previous topology with the following steps:
  
  1) Check that the cluster.nodes file reflects the previous (non-repartitioned) topology. If it does not, edit it manually so that it does.
  
  2) Issue a SOFTRESET to the cluster using the CAdmin tool. > cadmin softreset -cluster_home. (Issue #55082)
• The indexing pipeline can become deadlocked when multiple nodes are installed on the same windows box. Redeploying the nodes on separate boxes should resolve the issue. Further details below.

Symptom: Crawls succeed but documents never become searchable.

Diagnosis:

• Crawl job logs contain messages like the following: "FullySerializeCards caught an exception indexing cards in the property search collection: com.plumtree.server.marshalers.PTException: -2147204349 - Error in function PTIndexer.CommitAll (): -2147204349 - Failover failed -- no servers left to accept request."
• Search Cluster Manager shows all nodes in "RUN", "STALL", or "RECOVER" state; not "OFFLINE" or "READONLY".
• Search Server logs contain <warning> tags with the following text: "Index connection queue backlog exceeds maximum."

Analysis: The search cluster is in a deadlocked state where it cannot accept indexing requests. Queries are unaffected. The deadlock is due to contention over a lock file which one of the nodes is holding and not releasing.

Resolution: The problem can be temporarily resolved by restarting the search nodes one at a time. When the node holding the lock is restarted, the lock will be released and indexing requests can be accepted once more. Following the node recovery, run Search Repair to resubmit any indexing requests that may have failed during the outage. (Issue #56305)

• Pressing Enter on the Advanced Search page refreshes the page, but does not run a new search. Click 'Search' to perform a new search. (Issue #13773)
• The "More From Each..." link may appear at the bottom of the Federated Search results page when some of the Federated Searches do not have more results to be displayed. This is a bug. Clicking the "More From Each..." link under these circumstances may result in improper result count text, for example "Showing results 6-6 of 1", and/or it may result in the original results being repeated in the "More From Each..." page. There is no workaround for this. (Issue #28063)
• Certain Federated Search errors will result in an incorrect or misleading error message being displayed to the end user initiating the search. For example, a search for "*ra*" should cause the message "Specify at least three characters when wildcards appear at both the beginning and end of a query" to be displayed. Instead, the message "The Search Server for the Portal being searched could not be contacted" is shown, which is incorrect.
• You can choose to disable objects that cannot be disabled if a user object is part of the search result set. However, you are not notified that the object will not be disabled. (Issue #42546)
• When a Federated Search request to an external search Web service takes too long to complete (times out), the user sees this error message: "An error was returned by the search service: 'An Error occurred while conducting this search.'". Note that this generic message does not specify that the timeout was the cause of the failure. The first incoming Federated Search request to a newly-started WS Server is often slow, so your first attempt to test Federated Search may result in this generic error. Try repeating the request; subsequent attempts should be faster. (Issue #35290)
• Snapshot Queries using "In the Last" comparison operator includes future dates in the search results. (Issue #46105)
• Snapshot query portlets are sometimes referred to as "content snapshots" in online help and snapshot query portlet tips. (Issue #47687)

• If AllowGuestAccess is set to 0, guest users can access unsecured documents and objects. In order to properly protect data, remove the guest user from the ACL for objects. (Issue #45898)
• For portals that use Netegrity SSO, users may receive a warning when trying to join communities with an apostrophe (') in the name. (Issue #49552)
• Clickthroughs to gatewayed portlet URLs will succeed for unauthenticated users if Guest users are on the ACL of the portlet even if Guest access is disabled in the configuration of the portal in general. (Issue #50756)

• JSPortal dropdown menus cannot be opened until the page completes loading (until you see 'Done' at the bottom of the browser). Wait until the page finishes loading before using the menus or use another navigation scheme. (Issue #29933)
• The markup tag PT:TreeLink does not properly handle empty selections. You will receive an error if you deselect all in a tree node. (Issue #33730)

• Username/loginname can become corrupt during user import synchronization if there is a user that is missing their login name attribute. (Issue #28614)
• Editing a user profile portlet configuration and closing the browser can cause a profile page to be locked. It cannot be unlocked with the Release Item Locks utility. This can be repaired by connecting to the portal database. In the PTOBJECTLOCKS table, find and delete the row(s) with CLASSID 59. (Issue #47713)

• The Supported Operating Systems, Application Servers, Databases, and Browsers section has been updated to included support for BEA WebLogic 10, Windows Server 2003 SP2, and Microsoft SQL Server 2005 SP2 (with 2000 Compatibility Level). August 23, 2007