Docker-Compose script for Cassandra + Zeppelin setup
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
assets
zepcass
README.md
compose1.yml
docker-cluster.yml
docker-compose.yml
gcloud-cass-zep.sh
gcloud-user.sh
gcloud-zeppelin.sh
import.sh
java+zeppelin.sh
script.sh
zepl-conf.sh
zeppelin.sh

README.md

Cassandra + Spark + Zeppelin

This is a repository for a couple of docker-compose scripts, one of which that creates two Docker containers - one with a Zeppelin instance and the other one with a Cassandra node, the other one starting 4 containers - one with Zeppelin and 3 with a Cassandra three node cluster

Configuration and Installation

Make sure to have a valid Docker and docker-compose Installation, running on a 64-bit system (either directly on a mac or Linux machine, or on a VirtualBox - or similar - VM running a 64-bit guest; this means that you'll end up running Docker inside a VM, this is fine for testing and learning purposes).

To install/configure Docker and/or Docker Compose follow the steps described at https://docs.docker.com/compose/install/ and https://docs.docker.com/engine/installation/linux/ubuntu/ (this is for Ubuntu based Linux systems)

As a last step, clone this repository (you might need to do first apt-get install git)

git clone https://github.com/academyofdata/cassandra-zeppelin

Starting a single node Cassandra + Zeppelin instance

Once the docker & docker-compose prerequisites are met and the repository is cloned (example below assumes it is cloned in a folder called cassandra-zeppelin), do the following

cd cassandra-zeppelin
docker-compose build
docker-compose up -d

Assuming that you haven't encountered problems during build or run phase, you can now test that the containers are running by issuing the following command

docker ps

which should have an output similar with the one below

CONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS              PORTS                                                                                                      NAMES
110e8f4b16b3        zeppelin_zeppelin   "bin/zeppelin.sh"        4 days ago          Up 3 days           0.0.0.0:4040->4040/tcp, 0.0.0.0:8080-8081->8080-8081/tcp                                                   zeppelin_zeppelin_1
bbb70c263987        cassandra:3.9       "/docker-entrypoint.s"   4 days ago          Up 3 days           0.0.0.0:7000-7001->7000-7001/tcp, 0.0.0.0:7199->7199/tcp, 0.0.0.0:9042->9042/tcp, 0.0.0.0:9160->9160/tcp   zeppelin_cassandra_1

(pay attention in special to the STATUS column - it should say Up and not Exited) Once the containers are running you can go to http://virtualmachineip:8080 (replace with your own VirtualBox or local machine IP) and you should see the Zeppelin interface

Starting a Zeppelin instance connected to a Cassandra cluster (with 3 nodes)

PLEASE NOTE If you've previously started other containers with Zeppelin (for instance the Zeppelin + a single Cassandra node as outlined above), make sure to stop them before starting the instance connected to the cluster. You can do that with

docker-compose stop

Otherwise there will be port conflicts when attempting to start the new cluster and the new Zeppelin instance.

Start with this more complex configuration by issuing the command below (in the same folder where you've cloned this git repository)

docker-compose -f docker-cluster.yml up -d

After starting check that the containers are running (docker ps -a), wait for a few seconds (20-30 should be enough), log into one of the cassandra nodes (docker exec -ti zeppelin_node01_1 bash) and check the cluster status (run this in the container)

nodetool status

If the cluster started correctly you should see back a few lines, three of them starting with UN, like this

Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
--  Address     Load       Tokens       Owns (effective)  Host ID                               Rack
UN  172.17.0.3  110.13 KiB  256          67.6%             5460abe0-cf14-4d87-bf11-04f4ccd3f14c  rack1
UN  172.17.0.2  108.46 KiB  256          62.0%             17d1e7cd-2ff6-4397-8495-a42c12a3807f  rack1
UN  172.17.0.4  103.09 KiB  256          70.4%             70d2d32c-d7cd-4662-9e98-906167b0e4b7  rack1

This means that all the nodes are up (U) and operating normally (N)

Bulk-Loading data in Cassandra

PLEASE NOTE If you already have a 'test' keyspace it's better to drop it before executing the steps below.

To load all the exercise data into a newly created "test" keyspace and creating all the required tables, run the following command inside the Cassandra container (if you have an existing "test" keyspace, drop it)

apt-get update && apt-get install -y wget && wget -qO- https://raw.githubusercontent.com/academyofdata/cassandra-zeppelin/master/script.sh | bash

(to log into the container run 'docker exec -ti containers_cassandra_1 bash' from your container host, after you check the exact name of your container with 'docker ps -a')

Connecting Zeppelin to Cassandra

To be able to run queries from Zeppelin against a cassandra cluster (or a single node) we need to instruct Zeppelin's interpreter for Cassandra to connect to the right host. Since when using docker-compose we've specified that the cassandra container (or, when using a cluster, one of the containers) is available as the host 'cassandra', we just need adjust a single configuration value. For this, click in the right top corner of Zeppelin the "Anonymous" button to open the menu with a few options, one of which is "Interpreter"

Once on that page scroll to the Cassandra section and edit the value for cassandra.hosts to read cassandra as shown below

NOTE We could configure Zeppelin to connect to any of the hosts when running in the cluster configuration. For this we would first need to ammend the docker-compose configuration to also link the other nodes into zeppelin (in "links" section) and then we could set the cassandra.hosts to the hostnames separated by comma (e.g. "cassandra,cassandra2,cassandra3")

Starting containers without docker-compose

Assuming that you already have a running Cassandra container, in order to connect a new zeppelin instance to it run the following

docker run -d -p 8080:8080 -p 8081:8081 -p 4040:4040 -e MASTER="local[*]" -e ZEPPELIN_PORT="8080" -e ZEPPELIN_JAVA_OPTS="-Dspark.driver.memory=1g -Dspark.executor.memory=2g" -e SPARK_SUBMIT_OPTS="--conf spark.driver.host=localhost --conf spark.driver.port=8081" --link <id_or_name_of_cassandra_container>:cassandra --name zeppelin dylanmei/zeppelin

after the container starts run

docker exec -ti `docker ps --format '{{.Names}}' | grep zeppelin` bash -c "/usr/zeppelin/bin/install-interpreter.sh --name cassandra"

Starting a Zeppelin only instance

Edit the docker-compose.yml file to read as below

zeppelin:
  image:  dylanmei/zeppelin
  environment:
    ZEPPELIN_PORT: 8080
    ZEPPELIN_JAVA_OPTS: >-
      -Dspark.driver.memory=1g
      -Dspark.executor.memory=2g
    SPARK_SUBMIT_OPTIONS: >-
      --conf spark.driver.host=localhost
      --conf spark.driver.port=8081
      
    MASTER: local[*]
  ports:
    - 8080:8080
    - 8081:8081
    - 4040:4040
  volumes:
    - ./znotebooks:/usr/zeppelin/notebook

and issue the same docker-compose up -d command

get_num_processes

If you get a get_num_processes() takes no keyword arguments error, get out of cqlsh (but stay in the container shell, not on the host system) and run

rm /usr/lib/pymodules/python2.7/cqlshlib/copyutil.so