System Integration & Performance Testing

This directory contains Kafka system integration and performance tests. ducktape is used to run the tests.
(ducktape is a distributed testing framework which provides test runner, result reporter and utilities to pull up and tear down services.)

Running tests using docker

Docker containers can be used for running kafka system tests locally.

• Requirements
  • Docker 1.12.3 (or higher) is installed and running on the machine.
  • Test requires that Kafka, including system test libs, is built. This can be done by running

./gradlew clean systemTestLibs

• Run all tests

bash tests/docker/run_tests.sh

• Run all tests with debug on (warning will produce log of logs)

_DUCKTAPE_OPTIONS="--debug" bash tests/docker/run_tests.sh | tee debug_logs.txt

• Run a subset of tests

TC_PATHS="tests/kafkatest/tests/streams tests/kafkatest/tests/tools" bash tests/docker/run_tests.sh

• Run a specific tests file

TC_PATHS="tests/kafkatest/tests/client/pluggable_test.py" bash tests/docker/run_tests.sh

• Run multiple test files

TC_PATHS="tests/kafkatest/tests/client/pluggable_test.py tests/kafkatest/services/console_consumer.py" bash tests/docker/run_tests.sh

• Run a specific test class

TC_PATHS="tests/kafkatest/tests/client/pluggable_test.py::PluggableConsumerTest" bash tests/docker/run_tests.sh

• Run a specific test method

TC_PATHS="tests/kafkatest/tests/client/pluggable_test.py::PluggableConsumerTest.test_start_stop" bash tests/docker/run_tests.sh

• Run a specific test method with specific parameters

TC_PATHS="tests/kafkatest/tests/streams/streams_upgrade_test.py::StreamsUpgradeTest.test_metadata_upgrade" _DUCKTAPE_OPTIONS='--parameters '\''{"from_version":"0.10.1.1","to_version":"2.6.0-SNAPSHOT"}'\' bash tests/docker/run_tests.sh

• Run tests with a specific image name

image_name="ducker-ak-openjdk:17-buster" bash tests/docker/run_tests.sh

• Run tests with a different JVM

bash tests/docker/ducker-ak up -j 'openjdk:17-buster'; tests/docker/run_tests.sh

• Remove ducker-ak containers

bash tests/docker/ducker-ak down -f

• Rebuild first and then run tests

REBUILD="t" bash tests/docker/run_tests.sh

• Run tests with Kafka in native mode

  • To run tests with Kafka in native mode, pass {"kafka_mode": "native"} to the ducktape globals. This will bring up ducker nodes with the native Kafka binary inside them and use it to start Kafka while running the tests.

    _DUCKTAPE_OPTIONS="--globals '{\"kafka_mode\":\"native\"}'" TC_PATHS="tests/kafkatest/tests/"  bash tests/docker/run_tests.sh
  • To only bring up ducker nodes with kafka native binary inside it.

    bash tests/docker/ducker-ak up -m native
  • To run tests with Kafka in native mode using ducker-ak test(Make sure that the ducker nodes are already up with kafka native binary inside it).

    tests/docker/ducker-ak test tests/kafkatest/tests/client/compression_test.py -- --globals '{\"kafka_mode\":\"native\"}'

• Debug tests in VS Code:

  • Run test with --debug flag (can be before or after file name):

    tests/docker/ducker-ak up; tests/docker/ducker-ak test tests/kafkatest/tests/core/security_test.py --debug
  • Test will run in debug mode and wait for a debugger to attach.
  • Launch VS Code debugger with "attach" request - here's an example:

    {
    "version": "0.2.0",
    "configurations": [
        {
            "name": "Python: Attach to Ducker",
            "type": "python",
            "request": "attach",
            "connect": {
                "host": "localhost",
                "port": 5678
            },
            "justMyCode": false,
            "pathMappings": [
                {
                    "localRoot": "${workspaceFolder}",
                    "remoteRoot": "."
                }
            ]
        }
      ]
    }
  • To pass --debug flag to ducktape itself, use --:

    tests/docker/ducker-ak test tests/kafkatest/tests/core/security_test.py --debug -- --debug

• Notes

  • The scripts to run tests creates and destroys docker network named knw. This network can't be used for any other purpose.
  • The docker containers are named knode01, knode02 etc. These nodes can't be used for any other purpose.

• Exposing ports using --expose-ports option of ducker-ak up command

  If --expose-ports is specified then we will expose those ports to random ephemeral ports on the host. The argument can be a single port (like 5005), a port range like (5005-5009) or a combination of port/port-range separated by comma (like 2181,9092 or 2181,5005-5008). By default no port is exposed.

  The exposed port mapping can be seen by executing docker ps command. The PORT column of the output shows the mapping like this (maps port 33891 on host to port 2182 in container):

  0.0.0.0:33891->2182/tcp

  Behind the scene Docker is setting up a DNAT rule for the mapping and it is visible in the DOCKER section of iptables command (sudo iptables -t nat -L -n), something like:

  DNAT       tcp  --  0.0.0.0/0      0.0.0.0/0      tcp       dpt:33882       to:172.22.0.2:9092

  The exposed port(s) are useful to attach a remote debugger to the process running in the docker image. For example if port 5005 was exposed and is mapped to an ephemeral port (say 33891), then a debugger attaching to port 33891 on host will be connecting to a debug session started at port 5005 in the docker image. As an example, for above port numbers, run following commands in the docker image (say by ssh using ./docker/ducker-ak ssh ducker02):

  $ export KAFKA_OPTS="-agentlib:jdwp=transport=dt_socket,server=y,suspend=y,address=5005"

  $ /opt/kafka-dev/bin/kafka-topics.sh --bootstrap-server ducker03:9095 --topic __consumer_offsets --describe

  This will run the TopicCommand to describe the __consumer-offset topic. The java process will stop and wait for debugger to attach as suspend=y option was specified. Now starting a debugger on host with host localhost and following parameter as JVM setting:

  -agentlib:jdwp=transport=dt_socket,server=y,suspend=n,address=33891

  will attach it to the TopicCommand process running in the docker image.

Local Quickstart

This quickstart will help you run the Kafka system tests on your local machine. Note this requires bringing up a cluster of virtual machines on your local computer, which is memory intensive; it currently requires around 10G RAM. For a tutorial on how to setup and run the Kafka system tests, see https://cwiki.apache.org/confluence/display/KAFKA/tutorial+-+set+up+and+run+Kafka+system+tests+with+ducktape

• Install Virtual Box from https://www.virtualbox.org/ (run $ vboxmanage --version to check if it's installed).

• Install Vagrant >= 1.6.4 from https://www.vagrantup.com/ (run vagrant --version to check if it's installed).

• Install system test dependencies, including ducktape, a command-line tool and library for testing distributed systems. We recommend to use virtual env for system test development

    $ cd kafka/tests
    $ virtualenv -p python3 venv
    $ . ./venv/bin/activate
    $ python3 -m pip install --editable .
    $ cd ..  # back to base kafka directory

• Run the bootstrap script to set up Vagrant for testing

    $ tests/bootstrap-test-env.sh

• Bring up the test cluster

    $ vagrant/vagrant-up.sh
    $ # When using Virtualbox, it also works to run: vagrant up

• Build the desired branch of Kafka

    $ git checkout $BRANCH
    $ gradle  # (only if necessary)
    $ ./gradlew systemTestLibs

• Run the system tests using ducktape:

    $ ducktape tests/kafkatest/tests

EC2 Quickstart

This quickstart will help you run the Kafka system tests on EC2. In this setup, all logic is run on EC2 and none on your local machine.

There are a lot of steps here, but the basic goals are to create one distinguished EC2 instance that will be our "test driver", and to set up the security groups and iam role so that the test driver can create, destroy, and run ssh commands on any number of "workers".

As a convention, we'll use "kafkatest" in most names, but you can use whatever name you want.

Preparation

In these steps, we will create an IAM role which has permission to create and destroy EC2 instances, set up a keypair used for ssh access to the test driver and worker machines, and create a security group to allow the test driver and workers to all communicate via TCP.

• Create an IAM role. We'll give this role the ability to launch or kill additional EC2 machines.

• Create role "kafkatest-master"
• Role type: Amazon EC2
• Attach policy: AmazonEC2FullAccess (this will allow our test-driver to create and destroy EC2 instances)

• If you haven't already, set up a keypair to use for SSH access. For the purpose of this quickstart, let's say the keypair name is kafkatest, and you've saved the private key in kafktest.pem

• Next, create a EC2 security group called "kafkatest".

• After creating the group, inbound rules: allow SSH on port 22 from anywhere; also, allow access on all ports (0-65535) from other machines in the kafkatest group.

Create the Test Driver

• Launch a new test driver machine

• OS: Ubuntu server is recommended
• Instance type: t2.medium is easily enough since this machine is just a driver
• Instance details: Most defaults are fine.
• IAM role -> kafkatest-master
• Tagging the instance with a useful name is recommended.
• Security group -> 'kafkatest'

• Once the machine is started, upload the SSH key to your test driver:

    $ scp -i /path/to/kafkatest.pem \
        /path/to/kafkatest.pem ubuntu@public.hostname.amazonaws.com:kafkatest.pem

• Grab the public hostname/IP (available for example by navigating to your EC2 dashboard and viewing running instances) of your test driver and SSH into it:

    $ ssh -i /path/to/kafkatest.pem ubuntu@public.hostname.amazonaws.com

Set Up the Test Driver

The following steps assume you have ssh'd into the test driver machine.

• Start by making sure you're up to date, and install git and ducktape:

    $ sudo apt-get update && sudo apt-get -y upgrade && sudo apt-get install -y python3-pip git
    $ pip install ducktape

• Get Kafka:

    $ git clone https://git-wip-us.apache.org/repos/asf/kafka.git kafka

• Update your AWS credentials:

    export AWS_IAM_ROLE=$(curl -s http://169.254.169.254/latest/meta-data/iam/info | grep InstanceProfileArn | cut -d '"' -f 4 | cut -d '/' -f 2)
    export AWS_ACCESS_KEY=$(curl -s http://169.254.169.254/latest/meta-data/iam/security-credentials/$AWS_IAM_ROLE | grep AccessKeyId | awk -F\" '{ print $4 }')
    export AWS_SECRET_KEY=$(curl -s http://169.254.169.254/latest/meta-data/iam/security-credentials/$AWS_IAM_ROLE | grep SecretAccessKey | awk -F\" '{ print $4 }')
    export AWS_SESSION_TOKEN=$(curl -s http://169.254.169.254/latest/meta-data/iam/security-credentials/$AWS_IAM_ROLE | grep Token | awk -F\" '{ print $4 }')

• Install some dependencies:

    $ cd kafka
    $ ./vagrant/aws/aws-init.sh
    $ . ~/.bashrc

• An example Vagrantfile.local has been created by aws-init.sh which looks something like:

    # Vagrantfile.local
    ec2_instance_type = "..." # Pick something appropriate for your
                              # test. Note that the default m3.medium has
                              # a small disk.
    ec2_spot_max_price = "0.123"  # On-demand price for instance type
    enable_hostmanager = false
    num_zookeepers = 0
    num_kafka = 0
    num_workers = 9
    ec2_keypair_name = 'kafkatest'
    ec2_keypair_file = '/home/ubuntu/kafkatest.pem'
    ec2_security_groups = ['kafkatest']
    ec2_region = 'us-west-2'
    ec2_ami = "ami-29ebb519"

• Start up the instances:

    # This will brink up worker machines in small parallel batches
    $ vagrant/vagrant-up.sh --aws

• Now you should be able to run tests:

    $ cd kafka/tests
    $ ducktape kafkatest/tests

• Update Worker VM

If you change code in a branch on your driver VM, you need to update your worker VM to pick up this change:

    $ ./gradlew systemTestLibs
    $ vagrant rsync

• To halt your workers without destroying persistent state, run vagrant halt. Run vagrant destroy -f to destroy all traces of your workers.

Unit Tests

The system tests have unit tests! The various services in the python kafkatest module are reasonably complex, and intended to be reusable. Hence we have unit tests for the system service classes.

Where are the unit tests?

• The kafkatest unit tests are located under kafka/tests/unit

How do I run the unit tests?

$ cd kafka/tests # The base system test directory
$ python3 setup.py test

How can I add a unit test?

• Follow the naming conventions - module name starts with "check", class name begins with "Check", test method name begins with "check"
• These naming conventions are defined in "setup.cfg". We use "check" to distinguish unit tests from system tests, which use "test" in the various names.