src-kafka/vagrant

Apache Kafka

Using Vagrant to get up and running.

1. Install Virtual Box https://www.virtualbox.org/
2. Install Vagrant >= 1.6.4 https://www.vagrantup.com/
3. Install Vagrant Plugins:

$ vagrant plugin install vagrant-hostmanager

# Optional, to caches & shares package downloads across VMs
$ vagrant plugin install vagrant-cachier

In the main Kafka folder, do a normal Kafka build:

$ gradle
$ ./gradlew jar

You can override default settings in Vagrantfile.local, which is a Ruby file that is ignored by git and imported into the Vagrantfile. One setting you likely want to enable in Vagrantfile.local is enable_dns = true to put hostnames in the host's /etc/hosts file. You probably want this to avoid having to use IP addresses when addressing the cluster from outside the VMs, e.g. if you run a client on the host. It's disabled by default since it requires sudo access, mucks with your system state, and breaks with naming conflicts if you try to run multiple clusters concurrently.

Now bring up the cluster:

$ vagrant/vagrant-up.sh     
$ # If on aws, run: vagrant/vagrant-up.sh --aws

(This essentially runs vagrant up --no-provision && vagrant hostmanager && vagrant provision)

We separate out the steps (bringing up the base VMs, mapping hostnames, and configuring the VMs) due to current limitations in ZooKeeper (ZOOKEEPER-1506) that require us to collect IPs for all nodes before starting ZooKeeper nodes. Breaking into multiple steps also allows us to bring machines up in parallel on AWS.

Once this completes:

• Zookeeper will be running on 192.168.50.11 (and zk1 if you used enable_dns)
• Broker 1 on 192.168.50.51 (and broker1 if you used enable_dns)
• Broker 2 on 192.168.50.52 (and broker2 if you used enable_dns)
• Broker 3 on 192.168.50.53 (and broker3 if you used enable_dns)

To log into one of the machines:

vagrant ssh <machineName>

You can access the brokers by their IP or hostname, e.g.

# Specify brokers by their hostnames: broker1, broker2, broker3 (or just one of them)
bin/kafka-topics.sh --create --bootstrap-server broker1:9092 --replication-factor 3 --partitions 1 --topic sandbox
bin/kafka-console-producer.sh --bootstrap-server broker1:9092,broker2:9092,broker3:9092 --topic sandbox

# Specify brokers by their IP: 192.168.50.51, 192.168.50.52, 192.168.50.53
bin/kafka-console-consumer.sh --bootstrap-server 192.168.50.51:9092,192.168.50.52:9092,192.168.50.53:9092 --topic sandbox --from-beginning

If you need to update the running cluster, you can re-run the provisioner (the step that installs software and configures services):

vagrant provision

Note that this doesn't currently ensure a fresh start -- old cluster state will still remain intact after everything restarts. This can be useful for updating the cluster to your most recent development version.

Finally, you can clean up the cluster by destroying all the VMs:

vagrant destroy -f

Configuration

You can override some default settings by specifying the values in Vagrantfile.local. It is interpreted as a Ruby file, although you'll probably only ever need to change a few simple configuration variables. Some values you might want to override:

• enable_hostmanager - true by default; override to false if on AWS to allow parallel cluster bringup.
• enable_dns - Register each VM with a hostname in /etc/hosts on the hosts. Hostnames are always set in the /etc/hosts in the VMs, so this is only necessary if you want to address them conveniently from the host for tasks that aren't provided by Vagrant.
• enable_jmx - Whether to enable JMX ports on 800x and 900x for Zookeeper and the Brokers respectively where x is the nodes of each respectively. For example, the zk1 machine would have JMX exposed on 8001, ZK2 would be on 8002, etc.
• num_workers - Generic workers that get the code (from this project), but don't start any services (no brokers, no zookeepers, etc). Useful for starting clients. Each worker will have an IP address of 192.168.50.10x where x starts at 1 and increments for each worker.
• num_zookeepers - Size of zookeeper cluster
• num_brokers - Number of broker instances to run
• ram_megabytes - The size of each virtual machine's RAM; default to 1200MB

Using Other Providers

EC2

Install the vagrant-aws plugin to provide EC2 support:

$ vagrant plugin install vagrant-aws

Next, configure parameters in Vagrantfile.local. A few are required: enable_hostmanager, enable_dns, ec2_access_key, ec2_secret_key, ec2_keypair_name, ec2_keypair_file, and ec2_security_groups. A couple of important notes:

1. You definitely want to use enable_dns if you plan to run clients outside of the cluster (e.g. from your local host). If you don't, you'll need to go lookup vagrant ssh-config.

2. You'll have to setup a reasonable security group yourself. You'll need to open ports for Zookeeper (2888 & 3888 between ZK nodes, 2181 for clients) and Kafka (9092). Beware that opening these ports to all sources (e.g. so you can run producers/consumers locally) will allow anyone to access your Kafka cluster. All other settings have reasonable defaults for setting up an Ubuntu-based cluster, but you may want to customize instance type, region, AMI, etc.

3. ec2_access_key and ec2_secret_key will use the environment variables AWS_ACCESS_KEY and AWS_SECRET_KEY respectively if they are set and not overridden in Vagrantfile.local.

4. If you're launching into a VPC, you must specify ec2_subnet_id (the subnet in which to launch the nodes) and ec2_security_groups must be a list of security group IDs instead of names, e.g. sg-34fd3551 instead of kafka-test-cluster.

Now start things up, but specify the aws provider:

$ vagrant/vagrant-up.sh --aws

Your instances should get tagged with a name including your hostname to make them identifiable and make it easier to track instances in the AWS management console.