ftl-fsk-tools

Tools for configuring and migrating to TIBCO FTL(R) Service for Kafka (FSK) persistence clusters.

Tool	Language	Description
tibfsk/importconfig	Go	Translates Apache Kafka KRaft server.properties into FSK artifacts (ftlserver.json, tibftlserver-cluster.yaml, etc.) — start with getting-started.md
tibfsk/importdata	Java	Replicates records from a source Apache Kafka cluster to a target FSK cluster

Coming from an existing Apache Kafka cluster? Start with tibfsk/importconfig/getting-started.md, which converts the server.properties you already run into a working FSK configuration across twelve worked scenarios, from a single plaintext broker up to a secured three-node cluster. It is the only guide you need: the flag-by-flag reference, the output file details and the worked examples are sections at the end of it. importdata is a separate job — moving the records over after the configuration is in place.

The migration tool also ships a runnable demo scenario — see tibfsk/importdata/README.md for the four end-to-end paths.

---

Prebuilt artifacts

The repository checks in a built copy of each tool, so a clone runs with no Go toolchain and no javac:

Artifact	Built from	Used by
tibfsk/importconfig/bin/tibftlimportconfig	the Go sources, for linux/amd64	run it directly
tibfsk/importdata/classes/	tibfsk/importdata/src/main/java	run-apachekafka-to-fsk.sh
tibfsk/importdata/demo/classes/	tibfsk/importdata/demo/InsuranceDataProducer.java	demo/populate-kafka.sh

Both scripts prefer classes/ and fall back to compiling into build/ only when it is missing. Set FSK_FORCE_REBUILD=1 to compile from source anyway after editing the Java.

The binary is linux/amd64, statically linked with CGO_ENABLED=0. Build from source (below) for any other platform, or pass GOOS/GOARCH to the script. The class files are platform-independent.

To refresh the artifacts after changing a source file:

export KAFKA_HOME=/path/to/kafka      # or set KAFKA_CLASSPATH directly
./build-artifacts.sh

The checked-in class files were compiled with JDK 25 targeting --release 17, so they load on the Java 17 listed below and on anything newer. GO, GOOS, GOARCH, JAVAC, JAVA_RELEASE and KAFKA_CLASSPATH override the defaults. Nothing refreshes these artifacts automatically — re-run the script and commit the result whenever the sources change.

CMake ignores the checked-in artifacts and builds from source into its own binary directory.

---

Building

Building is optional. The documentation throughout this repository refers to the checked-in artifacts above — tibfsk/importconfig/bin/tibftlimportconfig and tibfsk/importdata/classes/ — which is what a clone runs by default. Build from source when you need a platform other than linux/amd64, or when you are changing the tools themselves.

Prerequisites

Requirement	Minimum	Notes
CMake	3.15
Go	1.25	For tibftlimportconfig. Set GO_EXECUTABLE if the right version is not first on PATH.
Java (JDK)	17	For tibftlfskimportdata. The checked-in classes were built with JDK 25.
Apache Kafka client jars	—	kafka-clients.jar + slf4j-api.jar; see below.

Build (both tools)

# Create a build directory outside the source tree.
mkdir build && cd build

# Configure — provide KAFKA_CLASSPATH for the Java tool.
cmake .. \
  -DKAFKA_CLASSPATH="/path/kafka-clients.jar:/path/slf4j-api.jar"

# Build.
cmake --build .

List the jars explicitly, colon-separated. A libs/* wildcard does not work here — the shell expands it inside the javac rule and every jar after the first is passed as a flag (error: invalid flag: .../activation-1.1.1.jar). The wildcard form is only valid for the runtime KAFKA_CLASSPATH read by run-apachekafka-to-fsk.sh.

Outputs:

• build/tibftlimportconfig — ready-to-run binary
• build/tibfsk/importdata/classes/ — compiled Java class files (run via run-apachekafka-to-fsk.sh)

javac runs only when the classes are missing or a .java source has changed, so repeat builds are no-ops. Because the compiled classes are installed next to run-apachekafka-to-fsk.sh, the script finds them and skips compiling entirely at run time — set FSK_FORCE_REBUILD=1 if you edit the sources and want the script to recompile.

Build tibftlimportconfig only (no Java required)

mkdir build && cd build
cmake ..   # omit -DKAFKA_CLASSPATH; Java tool is skipped with a warning
cmake --build . --target tibftlimportconfig_build

Omitting -DKAFKA_CLASSPATH is not an error: the Java tool's compile step is skipped and the sources ship as-is, so run-apachekafka-to-fsk.sh compiles them into build/ on first use.

Build the Java tool only

cmake --build . --target tibftlfskimportdata

Selecting the Go toolchain

If Go 1.25+ is not on your PATH but is installed elsewhere:

cmake .. -DGO_EXECUTABLE=/usr/local/go/bin/go

Obtaining Apache Kafka client jars

The Apache Kafka client jars are not bundled in this repository. Common sources:

• Apache Kafka installation: $KAFKA_HOME/libs/kafka-clients-*.jar and $KAFKA_HOME/libs/slf4j-api-*.jar
• Maven Central: org.apache.kafka:kafka-clients and org.slf4j:slf4j-api

At runtime, run-apachekafka-to-fsk.sh also needs a logging binding such as slf4j-simple.jar.

Install

cmake --install . --prefix /usr/local

Installs:

• /usr/local/bin/tibftlimportconfig
• /usr/local/share/tibftlfskimportdata/ (classes, conf, run script)

---

License

BSD 3-Clause. See LICENSE for the full text and NOTICE for attribution.