Please follow instructions below to download the hotfix: 

================================================================================

- Login to TIBCO Support Portal using your TIBCO Support Portal credentials.

- After logging-in select menu option “Downloads” -> “Hotfixes” , then navigate to /AvailableDownloads/BusinessEvents/EnterpriseEdition/6.3.0/hotfix-06 to download the Hotfix.

- Alternatively, you can use the direct portal link below to download 6.3.0 Hotfix-06. 

TIBCO BusinessEvents 6.3.0 Hotfix-06 Download Link

================================================================================

Closed Issues in 6.3.0_HF-006 (This Release)

BE-33123

The HTTP connection timeout is not reported after retry. 

But retry occurs for all connect exceptions.

For example: -HttpTimeoutException(request timeout)

We are restructuring retry handling for a request timeout. 

Now, the retry will be forInterruptedException, InterruptedIOException, 

UnknownHostException,ConnectException, SSLException, HttpConnectTimeoutException.

BE-33056

In TIBCO BusinessEvents, Apache Tomcat is upgraded to version 11.0.0-M21, 

10.1.25, or 9.0.90 to address the Apache Tomcat CVE-2024-34750 vulnerability.

BE-33039

The _annotations.idx file corrupts on concurrent access by two or more agents. 

As a result, an XPATH expression error occurs due to the corrupted index file 

"$BE_HOME/bin/_annotations.idx". 

Lock is applied to the _annotations.idx file to avoid concurrent access.

BE-33027

When connecting to an existing backingstore to generate an alter.sql script in 

Oracle, incorrect "ALTER TABLE" commands are created for existing CLOB columns. 

The be-storedeploy utility creates a CLOB column type for all the string columns 

exceeding the supported size. 

Consequently, the alter.sql script generates incorrect "ALTER TABLE" commands to 

add the existing CLOB column again. On execution, these commands fail with an error, 

"Column exist already". 

This issue also fails for keybase lookup projects, where an invalid SQL statement 

is created for the id$ column.

BE-33021

When migrating inbuild cache from ActiveSpaces 2.x to Apache Ignite, BQL query with 

DateTime field is not working properly in Ignite. This is because Ignite stores the 

DateTime field values as Java objects. 

Additionally, when a BQL query with DateTime is used in Ignite, it impacts the 

performance and is significantly slower as a native query.

So, when Shared Nothing data is migrated from Legacy ActiveSpaces to Ignite and if 

the DateTime field values is stored as Java objects in Ignite cache, then run the 

migration utility with option as -op "ignite_migration" to rectify the DateTime field 

value to be stored as TimeStamp. The command is as follows:

be-storedeploy -c <cdd-path> -op ignite_migration -u <processing-unit> -n <engine-name> <project_ear_path>

BE-32989

When the BE-TEA-Agent is restarted and BE running instances are stopped, there 

are errors in the log for each of the stopped instances. These errors are 

repeated every 15 seconds in the BE-TEA agent log until be-teagent is restarted. 

To stop these errors, it is required to restart the BE-TEA-Agent or restart 

the instances.

BE-32988

When a BE instance is manually killed in a terminal and restarted in BE-TEA-Agent GUI,

the running status of the instance is not updated correctly in the GUI. The instance

is running but displays as not running in BE-TEA-Agent GUI.

To return the correct status of running instance, it is required to restart 

the BE-TEA-Agent.

BE-32974

An exception occurs in Ignite Logs when delete instances and concept TTL 

coincides.

BE-32927

The Type attributes are printing for defaultJSONModel.cdd and the following 

CDD properties are ignored:

be.http.json.rootElement.ignore

be.http.json.childAttributes.ignore

BE-32882

When an index is created on the field with the same name but on two different 

concepts, an error message, "msg=Index already exists: " is displayed.

BE-32879

The contained child references are not deleted properly when TTL of the parent 

expires. The parent concepts evicts only from cache and not from store for 

concept TTL>0 in CDD.

Hence, the contained child concepts encounters error when it is assigned to the

new parent with the same extID. 

BE-32841

On performing the HTTP request, there is a FileDescriptor leak because of the 

continuous increase in the thread counts. The NO PROXY option of JDK HTTP client 

creates unwanted threads. 

To resolve this issue, the default proxy selector of JDK which has NO PROXY host

is set to NO PROXY.

BE-32838

There is no option to customize and override the default settings of 

clientConnectorConfig for a Ignite cache cluster. This setting of 

clientConnectorConfig allows to connect  to Ignite cache nodes from external 

SQL client tools, data access, running the SQLs, and so on.

To resolve this issue, now the custom port can be set as a CDD property 

"client-port" under IgniteStoreProvider section.

It should be set as <property name="client-port" value="<port>"/>. 

The default client port is 10800.

In a cluster of two nodes, if the client-port is configured for same CDD, then one

node uses the configured client-port and the next node takes the incremental port value.

For example, for same CDD, if the client-port is configured 10805, then one node

uses 10805 and the other node uses 10806. 

BE-32819

The QueryCache rule function works properly with ActiveSpaces but fails for 

Apache Ignite. While running the query on Query.Metadata.getColumnCount, the 

column count sometimes returns 0 which is incorrect.

BE-32798

Multiple times execution of a native query in a single transaction is causing 

the inference agent heap memory to grow. This ultimately results in Out-of-Memory.

================================================================================

Please refer to attached hotfix readme file for installation instructions, affected files and list of fixes in this release.