Public | Automated Build

Last pushed: 8 months ago
Short Description
Jupyter Notebook Python, Spark, Mesos Stack from
Full Description

Jupyter Notebook Python, Spark, Mesos Stack

What it Gives You

  • Jupyter Notebook 4.3.x
  • Conda Python 3.x and Python 2.7.x environments
  • pyspark, pandas, matplotlib, scipy, seaborn, scikit-learn pre-installed
  • Spark 2.1.0 with Hadoop 2.7 for use in local mode or to connect to a cluster of Spark workers
  • Mesos client 1.2.1 binary that can communicate with a Mesos master
  • Unprivileged user jovyan (uid=1000, configurable, see options) in group users (gid=100) with ownership over /home/jovyan and /opt/conda
  • tini as the container entrypoint and as the default command
  • A script useful for running a single-user instance of the Notebook server, as required by JupyterHub
  • A script useful for running alternative commands in the container (e.g. ipython, jupyter kernelgateway, jupyter lab)
  • Options for a self-signed HTTPS certificate and passwordless sudo

Basic Use

The following command starts a container with the Notebook server listening for HTTP connections on port 8888 with a randomly generated authentication token configured.

docker run -it --rm -p 8888:8888 jupyter/pyspark-notebook

Take note of the authentication token included in the notebook startup log messages. Include it in the URL you visit to access the Notebook server or enter it in the Notebook login form.

Using Spark Local Mode

This configuration is nice for using Spark on small, local data.

  1. Run the container as shown above.
  2. Open a Python 2 or 3 notebook.
  3. Create a SparkContext configured for local mode.

For example, the first few cells in the notebook might read:

import pyspark
sc = pyspark.SparkContext('local[*]')

# do something to prove it works
rdd = sc.parallelize(range(1000))
rdd.takeSample(False, 5)

Connecting to a Spark Cluster on Mesos

This configuration allows your compute cluster to scale with your data.

  1. Deploy Spark on Mesos.
  2. Configure each slave with the --no-switch_user flag or create the jovyan user on every slave node.
  3. Ensure Python 2.x and/or 3.x and any Python libraries you wish to use in your Spark lambda functions are installed on your Spark workers.
  4. Run the Docker container with --net=host in a location that is network addressable by all of your Spark workers. (This is a Spark networking requirement.)
  5. Open a Python 2 or 3 notebook.
  6. Create a SparkConf instance in a new notebook pointing to your Mesos master node (or Zookeeper instance) and Spark binary package location.
  7. Create a SparkContext using this configuration.

For example, the first few cells in a Python 3 notebook might read:

import os
# make sure pyspark tells workers to use python3 not 2 if both are installed
os.environ['PYSPARK_PYTHON'] = '/usr/bin/python3'

import pyspark
conf = pyspark.SparkConf()

# point to mesos master or zookeeper entry (e.g., zk://
# point to spark binary package in HDFS or on local filesystem on all slave
# nodes (e.g., file:///opt/spark/spark-2.0.2-bin-hadoop2.7.tgz)
conf.set("spark.executor.uri", "hdfs://")
# set other options as desired
conf.set("spark.executor.memory", "8g")
conf.set("spark.core.connection.ack.wait.timeout", "1200")

# create the context
sc = pyspark.SparkContext(conf=conf)

# do something to prove it works
rdd = sc.parallelize(range(100000000))

To use Python 2 in the notebook and on the workers, change the PYSPARK_PYTHON environment variable to point to the location of the Python 2.x interpreter binary. If you leave this environment variable unset, it defaults to python.

Of course, all of this can be hidden in an IPython kernel startup script, but "explicit is better than implicit." :)

Connecting to a Spark Cluster on Standalone Mode

Connection to Spark Cluster on Standalone Mode requires the following set of steps:

  1. Verify that the docker image (check the Dockerfile) and the Spark Cluster which is being deployed, run the same version of Spark.
  2. Deploy Spark on Standalone Mode.
  3. Run the Docker container with --net=host in a location that is network addressable by all of your Spark workers. (This is a Spark networking requirement.)
  4. The language specific instructions are almost same as mentioned above for Mesos, only the master url would now be something like spark://

Notebook Options

The Docker container executes a script script by default. The script handles the NB_UID and GRANT_SUDO features documented in the next section, and then executes the jupyter notebook.

You can pass Jupyter command line options through the script when launching the container. For example, to secure the Notebook server with a custom password hashed using IPython.lib.passwd() instead of the default token, run the following:

docker run -d -p 8888:8888 jupyter/pyspark-notebook --NotebookApp.password='sha1:74ba40f8a388:c913541b7ee99d15d5ed31d4226bf7838f83a50e'

For example, to set the base URL of the notebook server, run the following:

docker run -d -p 8888:8888 jupyter/pyspark-notebook --NotebookApp.base_url=/some/path

For example, to disable all authentication mechanisms (not a recommended practice):

docker run -d -p 8888:8888 jupyter/pyspark-notebook --NotebookApp.token=''

You can sidestep the script and run your own commands in the container. See the Alternative Commands section later in this document for more information.

Docker Options

You may customize the execution of the Docker container and the command it is running with the following optional arguments.

  • -e GEN_CERT=yes - Generates a self-signed SSL certificate and configures Jupyter Notebook to use it to accept encrypted HTTPS connections.
  • -e NB_UID=1000 - Specify the uid of the jovyan user. Useful to mount host volumes with specific file ownership. For this option to take effect, you must run the container with --user root. (The script will su jovyan after adjusting the user id.)
  • -e GRANT_SUDO=yes - Gives the jovyan user passwordless sudo capability. Useful for installing OS packages. For this option to take effect, you must run the container with --user root. (The script will su jovyan after adding jovyan to sudoers.) You should only enable sudo if you trust the user or if the container is running on an isolated host.
  • -v /some/host/folder/for/work:/home/jovyan/work - Host mounts the default working directory on the host to preserve work even when the container is destroyed and recreated (e.g., during an upgrade).

SSL Certificates

You may mount SSL key and certificate files into a container and configure Jupyter Notebook to use them to accept HTTPS connections. For example, to mount a host folder containing a notebook.key and notebook.crt:

docker run -d -p 8888:8888 \
    -v /some/host/folder:/etc/ssl/notebook \
    jupyter/pyspark-notebook \

Alternatively, you may mount a single PEM file containing both the key and certificate. For example:

docker run -d -p 8888:8888 \
    -v /some/host/folder/notebook.pem:/etc/ssl/notebook.pem \
    jupyter/pyspark-notebook \

In either case, Jupyter Notebook expects the key and certificate to be a base64 encoded text file. The certificate file or PEM may contain one or more certificates (e.g., server, intermediate, and root).

For additional information about using SSL, see the following:

Conda Environments

The default Python 3.x Conda environment resides in /opt/conda. A second Python 2.x Conda environment exists in /opt/conda/envs/python2. You can switch to the python2 environment in a shell by entering the following:

source activate python2

You can return to the default environment with this command:

source deactivate

The commands jupyter, ipython, python, pip, easy_install, and conda (among others) are available in both environments. For convenience, you can install packages into either environment regardless of what environment is currently active using commands like the following:

# install a package into the python2 environment
pip2 install some-package
conda install -n python2 some-package

# install a package into the default (python 3.x) environment
pip3 install some-package
conda install -n python3 some-package

Alternative Commands

JupyterHub requires a single-user instance of the Jupyter Notebook server per user. To use this stack with JupyterHub and DockerSpawner, you must specify the container image name and override the default container run command in your

# Spawn user containers from this image
c.DockerSpawner.container_image = 'jupyter/pyspark-notebook'

# Have the Spawner override the Docker run command
    'command': '/usr/local/bin/'

The script supports the same features as the default script (e.g., GRANT_SUDO), but allows you to specify an arbitrary command to execute. For example, to run the text-based ipython console in a container, do the following:

docker run -it --rm jupyter/pyspark-notebook ipython

This script is particularly useful when you derive a new Dockerfile from this image and install additional Jupyter applications with subcommands like jupyter console, jupyter kernelgateway, and jupyter lab.


You can bypass the provided scripts and specify your an arbitrary start command. If you do, keep in mind that certain features documented above will not function (e.g., GRANT_SUDO).

Docker Pull Command
Source Repository