Adoptable Cookbooks List

Looking for a cookbook to adopt? You can now see a list of cookbooks available for adoption!
List of Adoptable Cookbooks

Supermarket Belongs to the Community

Supermarket belongs to the community. While Chef has the responsibility to keep it running and be stewards of its functionality, what it does and how it works is driven by the community. The chef/supermarket repository will continue to be where development of the Supermarket application takes place. Come be part of shaping the direction of Supermarket by opening issues and pull requests or by joining us on the Chef Mailing List.

Select Badges

Select Supported Platforms

Select Status


cerner_kafka (12) Versions 2.1.0

Installs and configures a Kafka

cookbook 'cerner_kafka', '= 2.1.0', :supermarket
cookbook 'cerner_kafka', '= 2.1.0'
knife supermarket install cerner_kafka
knife supermarket download cerner_kafka
Quality 100%

Cerner Kafka

Cookbook Version
Build Status

A Chef cookbook to install Apache's Kafka.

A recipe to install the KafkaOffsetMonitor application is
also included.

View the [Change Log]( to see what has changed.

What this will setup

This will install Kafka at node["kafka"]["install_dir"]. In order to handle upgrades appropriately we use symbolic links and land the real installations

The true installations will land,

node["kafka"]["base_dir"] (defaults to /opt)
| - kafka_1.2.3
| | - config
| | - bin
| | - kafka_1.2.3.jar
| | - ...

While we provide a symbolic link to the convenient location,

node["kafka"]["install_dir"] (defaults to /opt/kafka)
| - config
| - bin
| - kafka_1.2.3.jar
| - ...

It will also create/setup a service which can be used to start/stop/restart kafka,

service kafka [start|stop|restart|status]

We also link kafka's log directory to /var/log/kafka to make it easier to find kafka's logs.


These are the requirements that need to be satisfied in order to get the cookbook to run.

  • Have filled out the following required attributes
    • node["kafka"]["brokers"] or node["kafka"][""][""]
    • node["kafka"]["zookeepers"] or node["kafka"][""]["zookeeper.connect"]

If you don't fulfill the requirements the recipe will error out telling you what was missing.

Additionally if you want to use a Zookeeper chroot with your kafka installation you can provide it by setting

The attributes,

  • node["kafka"]["brokers"]
  • node["kafka"]["zookeepers"]

Actually map to '' settings,

  • node["kafka"][""][""] : The id of broker running on the server
  • node["kafka"][""]["zookeeper.connect"] : The Kafka configuration used to connect to Zookeeper

We do this mapping for you when you provide the node["kafka"]["brokers"] and node["kafka"]["zookeepers"]

For the node["kafka"]["brokers"] => node["kafka"][""][""] mapping to work properly
all Chef nodes running the kafka recipe (and are part of the same Kafka cluster) must have the same list of
node["kafka"]["brokers"] and all broker hostnames must be in the same order. We determine the
node["kafka"][""][""] by using the index of Chef node's fqdn/hostname/ip in the
list as the node["kafka"][""][""].

Using node["kafka"]["brokers"], node["kafka"]["zookeepers"] and node["kafka"]["zookeeper_chroot"] attribtues are
the recommended way to setup your kafka cluster in Chef.

Once all that is done you should be able to run the recipe without any problem.

Updating from 1.X and 2.X of the Cookbook

There were some non-passive changes made during the upgrade to the 2.X version of
the cookbook. Specifically,

  • Removed a number of default kafka configs (See here)
  • Updated defaults to install Kafka (Scala 2.11)

Make sure to make the appropriate attribute changes if needed. Otherwise the
cookbook should work just as it did before.

Additionally in Kafka 0.9 broker's can be auto assigned broker ids. The cookbook
supports this feature. Make sure to keep the broker ids for existing nodes otherwise
they may drop their data.

How to specify where to download kafka from and which version to install

This cookbook supports Kafka version and above. The default attributes currently will install version from
''. This is configured using a number of different attributes in order to make it easier for you.

NOTE: If you are upgrading from 0.8.X to 0.9.X there are some additional steps to handle a rolling upgrade.

There are basically two ways to configure these settings. The first way is via 3 different attributes,

  • node["kafka"]["scala_version"] : The scala version number associated with the kafka installation (default="2.11")
  • node["kafka"]["version"] : The version number associated with the kafka installation (default="")
  • node["kafka"]["download_url"] : The base url used to download Kafka (default="")

With these 3 attributes we build the full url of the form

This makes it easy to specify just a single change while still maintaining the rest of the URL.

The other option is to just provide the full URL itself,

  • node["kafka"]["binary_url"] : The full url used to download Kafka

NOTE : If you specify the node["kafka"]["binary_url"] a valid and up to date node["kafka"]["version"] must also be provided as this is what we use to determine
if a new version of kafka is specified.

Configuring java for the server/broker

Currently the cookbook defaults to use the same java settings that Linkedin recommends.

Kafka uses different environment variables to configure the java settings for the server/broker,

  • KAFKA_JMX_OPTS : The JMX settings (default="$JMX_PORT")
  • JMX_PORT : The port to run Kafka's JMX on (default=9999)
  • KAFKA_LOG4J_OPTS : The log4j settings (default="-Dlog4j.configuration=file:$base_dir/../config/")
  • KAFKA_HEAP_OPTS : The options used to control Kafka's Heap (default="-Xmx4G -Xms4G")
  • KAFKA_JVM_PERFORMANCE_OPTS : The options used to control JVM performance settings (default="-XX:PermSize=48m -XX:MaxPermSize=48m -XX:+UseG1GC -XX:MaxGCPauseMillis=20 -XX:InitiatingHeapOccupancyPercent=35")
  • KAFKA_GC_LOG_OPTS : The options used to control GC logs (default="-Xloggc:$LOG_DIR/$GC_LOG_FILE_NAME -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCDateStamps -XX:+PrintGCTimeStamps")
  • KAFKA_OPTS : Used for any generic JVM settings (default="")

You can customize these environment variables (as well as any environment variable for the kafka user) using the

  • node["kafka"]["env_vars"] : A hash of environment variable names to their values to be set for the kafka user

Consumer Offset Monitor

The kafka::offset_monitor recipe will install the Kafka Consumer Offset Monitor application, which provides a
web UI for monitoring various aspects of your Kafka cluster, including consumer processing lag.

This recipe shares several attributes with the default recipe:

  • node["kafka"]["zookeepers"] (required)
  • node["kafka"]["user"]
  • node["kafka"]["group"]
  • node["kafka"]["base_dir"]
  • node["kafka"]["log_dir"]

The offset monitor application is installed to node["kafka"]["offset_monitor"]["install_dir"]
(defaults to node["kafka]["base_dir"]/kafka-offset-monitor). The application download URL is controlled by
the node["kafka"]["offset_monitor"]["url"] attribute.

The offset monitor maintains an SQLite database comparing the latest Kafka broker offset for each
topic partition to the latest offset persisted in Zookeeper for each consumer group. The database file
(node["kafka"]["offset_monitor"]["db_name"].db) is written to the kafka user's home directory. The monitoring
refresh interval and retention period are configurable by the attributes node["kafka"]["offset_monitor"]["refresh"]
and node["kafka"]["offset_monitor"]["retain"] using scala.concurrent.duration.Duration syntax (for example
"60.seconds" or "30.days").

The offset monitor web application listens on port 8080 by default. The port is configurable by setting the
node["kafka"]["offset_monitor"]["port"] attribute in the event that a conflicting service is already using port

This recipe is not included by the default recipe. It can be added to any or all of the Kafka broker nodes,
or a separate node or VM. Each instance of the offset monitor will collect an identical complete data set
for the entire Kafka cluster regardless of where it is installed, other than minor variances due to refresh interval

This recipe creates a service which can be used to start/stop/restart the offset monitor java process,

service kafka-offset-monitor [start|stop|restart|status]

Log files are written to kafka-offset-monitor.log in node["kafka"]["log_dir"] (defaults to /var/log/kafka).


  • node["kafka"]["brokers"] : An array of the list of brokers in the Kafka cluster. This should even include the node running the recipe. (default=[])
  • node["kafka"]["zookeepers"] : An array of the list of Zookeepers that Kafka uses. (default=[])
  • node["kafka"]["zookeeper_chroot"] : A string representing the Zookeeper chroot to use for Kafka. (default=nil)
  • node["kafka"]["user"] : The name of the user used to run Kafka (default="kafka")
  • node["kafka"]["group"] : The name of the group the user running Kafka is associated with (default="kafka")
  • node["kafka"]["openFileLimit"] : The open file limit for the user running the Kafka service (default=32768)
  • node["kafka"]["maxProcesses"] : The max processes allowed for the user running the Kafka service (default=1024)
  • node["kafka"]["scala_version"] : The scala version number associated with the kafka installation (default="2.9.2")
  • node["kafka"]["version"] : The version number associated with the kafka installation (default="")
  • node["kafka"]["download_url"] : The base url used to download Kafka. Uses this and node["kafka"]["scala_version"] as well as node["kafka"]["version"] to build the full url. (default="")
  • node["kafka"]["binary_url"] : The full url used to download Kafka. (default=#{node["kafka"]["download_url"]}/#{node["kafka"]["version"]}/kafka_#{node["kafka"]["scala_version"]}-#{node["kafka"]["version"]}.tgz)
  • node["kafka"]["base_dir"] : This is the directory that contains the current installation as well as every other installation (default="/opt")
  • node["kafka"]["install_dir"] : This is the directory of the current installation (default=node["kafka"]["base_dir"]/kafka)
  • node["kafka"]["log_dir"] : The directory of the log files for Kafka. Not Kafka's message/log data but debug logs from the server. (default="/var/log/kafka")
  • node["kafka"]["shutdown_timeout"] : The init.d script shutdown timeout in seconds. Adjust as needed based on cluster size (in terms of partitions) and required shutdown time. (default=30)
  • node["kafka"]["env_vars"] : A hash of environment variable names to their values to be set for the kafka user. This can be used to customize the server memory settings. (default={})
  • node["kafka"]["lib_jars"] : A list of URLs to install a jar in #{node["kafka"]["install_dir"]}/libs. (default=[])
  • node["kafka"][""][*] : A key/value that will be set in server's properties file. Used to customize the broker configuration. (default={} See Kafka doc for Kafka defaults)
  • node["kafka"][""][*] : A key/value that will be set in the server's file. (for defaults see attributes file)
  • node["kafka"]["offset_monitor"]["url"] The download url for the offset monitor (default = "")
  • node["kafka"]["offset_monitor"]["install_dir"] : The installation directory for the offset monitor (default = node["kafka]["base_dir"]/kafka-offset-monitor)
  • node["kafka"]["offset_monitor"]["main_class"] : The main class for the offset monitor (default = "com.quantifind.kafka.offsetapp.OffsetGetterWeb")
  • node["kafka"]["offset_monitor"]["port"] = The port used by the offset monitor web application (default = 8080)
  • node["kafka"]["offset_monitor"]["refresh"] : How often the offset monitor refreshes and stores a point in the DB, in value.unit format (default = "15.minutes")
  • node["kafka"]["offset_monitor"]["retain"] : How long the offset monitoring data is kept in the DB, in value.unit format (default = "7.days")
  • node["kafka"]["offset_monitor"]["db_name"] : The base file name for the offset monitoring database file written into the kafka user's home directory (default = "offset_monitor")
  • node["kafka"]["service"]["stdout"] : The file to keep std output of kafka init service (default = "/dev/null")
  • node["kafka"]["service"]["stderr"] : The file to keep std error of kafka init service (default = "/dev/null")


We have provided some simple integration tests for testing the cookbook.

How to run tests

To run the tests for this cookbook you must install ChefDK.

The unit tests are written with rspec and chefspec.
They can be run with rspec.

The lint testing uses Foodcritic and can be run with foodcritic . -f any.

The integration tests are written with test-kitchen and serverspec.
They can be run with kitchen test.


This project is licensed under the Apache License, Version 2.0.

When contributing to the project please add your name to the CONTRIBUTORS.txt file. Adding your name to the CONTRIBUTORS.txt file
signifies agreement to all rights and reservations provided by the License.

To contribute to the project execute a pull request through github. The pull request will be reviewed by the community and merged
by the project committers. Please attempt to conform to the test, code conventions, and code formatting standards if any
are specified by the project before submitting a pull request.


Releases should happen regularly after most changes. Feel free to request a release by logging an issue.


For information related to being a committer check [here](


Copyright 2013 Cerner Innovation, Inc.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Dependent cookbooks

java >= 0.0.0
ulimit >= 0.0.0
logrotate >= 0.0.0

Contingent cookbooks

There are no cookbooks that are contingent upon this one.

Change Log

2.1.0 - 04-25-2016

  • [Enhancement] Issue-39 : separated user and group logic to another recipe

2.0.0 - 02-12-2016

  • [Enhancement] Issue-33 : Default to version 0.9.0 and scala 2.11
  • [Bug] Issue-29 : kafka stdout logs same thing as server.log
  • [Enhancement] Issue-28 : add libraries to find broker id and zookeepers
  • [Enhancement] Issue-14 : Remove broker config defaults

1.1.0 - 08-14-2015

  • [Enhancement] Issue-24 : I've encountered some issues on the shell scripts with this cookbook. Here are my proposed fix and improvements
  • [Enhancement] Issue-22 : lock file cleanup, make kafka user uid configureable
  • [Enhancement] Issue-18 : fix adding nil if hostname isn't set up first
  • [Feature] Issue-12 : Support zookeeper chroot in a first class manner
  • [Bug] Issue-10 : Error executing action stop on resource 'service[kafka]'
  • [Enhancement] Issue-9 : Setup repo to use travis ci
  • [Enhancement] Issue-8 : Updated README with release info, created COMMITTERS file for commiter i...
  • [Enhancement] Issue-7 : Cleaned up code to add stderr/stdout logging for kafka start/stop comman...
  • [Enhancement] Issue-6 : Add kafka service stdout and stderr
  • [Enhancement] Issue-3 : Restrict overwriting to < 0.8.2

Foodcritic Metric

2.1.0 passed this metric