Skip to content

Instantly share code, notes, and snippets.

@massenz
Created Feb 16, 2015
Embed
What would you like to do?
This will install all binaries and libraries for the SciPy Python environment, Cassandra and Spark on an AWS Ubuntu AMI instance. This assumes you are running this script as described in the README documentation.

HOW-TO Build SparkLab

Author:
  1. Massenzio
Date: 2015-01-22
Version: 1.0.0
Scope:Personal Use
License:Apache 2

https://go-shields.herokuapp.com/license-apache2-blue.png

Prerequisite

Create an AWS instance and install your SSH key on the machine so you can ssh into it (use ubuntu user) making sure that the security groups allow access to the instance's port 22 (SSH) from anywhere (0.0.0.0/32) or, at least from the subnet you will be accessing it from.

You may also want to place the instance on a private subnet and open a number of ports from the subnet's private LAN (eg, 10.10.0.0/16) so you can connect to Cassandra, Spark, iPython, etc.

Copy the files to that instance:

$ scp build-sparklab.sh ubuntu@ipaddress:./
$ scp sparklab-requirements.txt ubuntu@ipaddress:./

then execute it:

$ ssh ubuntu@ipaddress
...
$ ./build-sparklab.sh

Notes

It may be easier to scp a more complete .bashrc to the AWS instance before starting; also .gitconfig and .gitignore may be useful (optional) [1]

One issue (particularly for accessing Cassandra) is around the hostname; eventually, I've found that it's best to use Route53 and make the hostname DNS resolvable.

See Cassandra documentation as to how to make the server reachable from clients other than localhost.

Starting Spark master:

$SPARK_HOME/sbin/start-master.sh

Sparker worker:

$SPARK_HOME/sbin/spark-class org.apache.spark.deploy.worker.Worker spark://ip-xx-xx-xxx-xxx:xxxx

$SPARK_HOME/sbin and $SPARK_HOME/bin should already be in PATH so you can execute the above by just running the command.

Spark logs:

$SPARK_HOME/logs/spark-ubuntu-org.apache.spark.deploy.master.Master-1-ip-xx-xx-xxx-xxx.out

(still TODO: change spark logs to somewhere like /var/log/spark)

Starting cassandra:

sudo service cassandra start

cassandra logs:

/var/log/cassandra/system.log
/var/log/cassandra/output.log

IPython notebook

Start with:

workon scipy3
ipython notebook --pylab=inline --profile=nbserver

see documentation on securing ipython.

NOTE Currently SSL connection uses a self-signed cert saved in ~/.ssh/ipy-cert.pem this will cause a warning in the browser - it is safe to ignore.

git

In order to clone projects from github to any newly created SparkLab instance, you need to generate a key-pair and then add the public part to github's SSH keys for the account.

If you don't know how to do this, maybe you should not be reading this guide after all.

[1]Rename them from edit.* to their .* equivalent
#!/bin/bash
#
# SparkLab build script
#
# Created by M. Massenzio, 2015-01-22
#
# ssh ubuntu@host:port
declare -r VENVS="/data/virtualenvs"
declare -r SCIPY_VENV="scipy3"
declare -r SCIPY_DIR="${VENVS}/${SCIPY_VENV}"
echo "[INFO] This will install all binaries and libraries for the SciPy Python environment,
[INFO] Cassandra and Spark on an AWS Ubuntu AMI instance.
[INFO] This assumes you are running this script as described in the README documentation.
"
read -p "Press a key when ready or Ctrl-C to abort"
sudo apt-get update && sudo apt-get -y upgrade
# There are a bunch of packages that will be needed later
# Install them all now:
sudo apt-get install -y pkg-config gcc g++ libreadline-dev git git-sh maven
# Creating a /data directory for DB files and other assorted data files
# Making it also a bit more friendly (if less secure) to use
sudo adduser ${USER} users
# Attach EBS block device
# https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ebs-using-volumes.html
read -p "If this is being configured on a AWS new AMI instance, you may need
to attach, format and mount an external EBS disk to the instance.
Do you wish to do this? [y/N] " ebs
if [[ $ebs == 'y' ]]; then
lsblk
#
# Should emit something like the below:
#
# NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
# xvda 202:0 0 8G 0 disk
# └─xvda1 202:1 0 8G 0 part /
# xvdb 202:16 0 100G 0 disk
echo "[INFO] The listing above should show at least an entry of the form
xvdb 202:16 0 100G 0 disk"
read -p "Please enter the name of the EBS device you wish to format and mount on this instance
(or enter None to skip): " disk
# Verify it's an empty, unformatted device:
IS_EMPTY=$(sudo file -s /dev/$disk | egrep "/dev/$disk: data")
#
# /dev/xvdb: data
SKIP='n'
if [[ -z ${IS_EMPTY} ]]; then
sudo file -s /dev/$disk
echo "[WARN] It seems that the device is already formatted, are you sure you want to progress."
echo "[WARN] All data on the device will be lost; you've been warned!"
echo "[WARN] To force re-format, please enter 'n' at the next prompt."
SKIP='y'
read -p "Do you want to skip this step (Ctrl-C to abort) [Y/n]? " SKIP
fi
# Create an ext4 filesystem on it
if [[ ${SKIP} == 'n' ]]; then
sudo mkfs -t ext4 /dev/$disk
sudo mkdir -p /mnt/data
# Add the mount point to fstab
cp /etc/fstab ./
echo "/dev/$disk /mnt/data ext4 defaults,auto,nofail,nobootwait 0 0" >>fstab
sudo cp ./fstab /etc/fstab
sudo mount -a
fi
echo "[INFO] /dev/$disk should now be mounted on /mnt/data"
IS_MOUNTED=$(mount |grep $disk)
if [[ -z "${IS_MOUNTED}" ]]; then
echo "[WARN] If seems that the secondary EBS disk is not mounted:"
mount
read -p "Press a key to continue or Ctrl-C to abort"
fi
sudo ln -snf /mnt/data /data
else
sudo mkdir -p /data
fi
if [[ ! -d "/data" ]]; then
echo "[ERROR] Missing data directory (/data) please correct and run this script again"
exit 1
fi
sudo chown ${USER}:users /data
sudo chmod 2775 /data
# Setup the Python environment
#
# Install matplotlib (pip install matplotlib will fail)
# and the Python dev package (the headers are required by numpy and friends)
# Make sure the python versions match the one installed
# libzmq is required by pyzmq (or its install will fail)
#
echo "[INFO] Installing Python 3, virtual environment support and SciPy packages"
read -p "Press a key when ready or Ctrl-C to abort"
sudo apt-get install -y python-setuptools python3-matplotlib python3.4-dev libzmq-dev
sudo easy_install pip
sudo pip install virtualenvwrapper
if [[ $? != 0 ]]; then
echo "[ERROR] Could not install virtualenvwrapper, please check pip logs"
exit 1
fi
mkdir -p "${VENVS}"
# Add the virtualenvwrapper env for .bashrc
#
echo "export WORKON_HOME=\"${VENVS}\"
source /usr/local/bin/virtualenvwrapper.sh" >> .bashrc
source .bashrc
# Create the virtual environment
# The use of --system-site-packages is necessary to make matplotlib work in Python 3
#
if [[ ! -d "${SCIPY_DIR}" ]]; then
mkvirtualenv -p `which python3` --system-site-packages ${SCIPY_VENV}
if [[ $? != 0 ]]; then
echo "[ERROR] Could not create a virtualenv, please check the error message, if any"
exit 1
fi
workon ${SCIPY_VENV}
fi
IS_PY3=$(python --version | egrep "3\.[0-9]+")
if [[ -z "${IS_PY3}" ]]; then
echo "[WARN] You may not have Python 3.x enabled, double check the virtualenv"
echo "[WARN] Python 3 is here: [`which python3`]; your default python is: [`which python`]"
python --version
fi
# Add all the necessary packages (see the SparkLab-requirements.txt file)
# NOTE - this will take a long time to run, with no output to stdout
if [[ ! -f SparkLab-requirements.txt ]]; then
echo "[ERROR] Missing SparkLab-requirements.txt file, please copy it to the ubuntu user home dir"
exit 1
fi
echo "[INFO] Installing SciPy packages, this will take forever: go grab a book..."
read -p "Press a key when ready or Ctrl-C to abort"
pip install -r SparkLab-requirements.txt
if [[ $? != 0 ]]; then
echo "[ERROR] Your virtual environment may miss critical packages"
fi
read -p "Press a key when ready or Ctrl-C to abort"
## Add Java JDK 8 & Install Cassandra
# See: http://tecadmin.net/install-oracle-java-8-jdk-8-ubuntu-via-ppa/
sudo add-apt-repository ppa:webupd8team/java
sudo apt-get update && sudo apt-get install -y oracle-java8-installer
IS_JAVA8=$(java -version 2>&1 | egrep -e "1\.8\.0_[0-9]+")
if [[ -z ${IS_JAVA8} ]]; then
echo "[WARN] Your installed java may not support the latest features:"
java -version
echo "[WARN] Make sure to fix this after setup completes. This is not critical for progress"
read -p "Press a key when ready"
fi
## Install Cassandra
# See: http://www.datastax.com/documentation/cassandra/1.2/cassandra/install/installDeb_t.html
# This would FAIL (with some unreasonable permission error):
# sudo echo "deb http://debian.datastax.com/community stable main" > /etc/apt/sources.list.d/cassandra.sources.list
#
# We need to manually edit the file and then `sudo cp` to the correct directory:
echo "deb http://debian.datastax.com/community stable main" > ~/cassandra.sources.list
sudo cp cassandra.sources.list /etc/apt/sources.list.d/
curl -L http://debian.datastax.com/debian/repo_key | sudo apt-key add -
sudo apt-get update && sudo apt-get install -y dsc20=2.0.11-1 cassandra=2.0.11
# Cassandra configuration is fiddly and does not work well with AWS EC2 instances
#
# 1. you need to edit the -Xss Java option (stack size)
# See: http://stackoverflow.com/questions/22470628/cassandrathe-stack-size-specified-is-too-small-specify-at-least-228k
# The only solution is to manually edit the environment settings and change the following line:
#
# JVM_OPTS="$JVM_OPTS -Xss180k"
# to
# JVM_OPTS="$JVM_OPTS -Xss256k"
#
echo "Change the line JVM_OPTS=\"\$JVM_OPTS -Xss180k\" to a value >= 256k"
read -p "Press a key when ready"
sudo vim /etc/cassandra/cassandra-env.sh
# 2. The host location and data directories need to be configured.
# While you are at it, add the host IP address to the list of `seeds`
# in /etc/cassandra/cassandra.yaml
#
# Look for a line that says `- seeds` and change it to something like:
#
# - seeds: "ip1,ip2"
#
# and the directories where Cassandra should store data on disk. Cassandra
# will spread data evenly across them, subject to the granularity of
# the configured compaction strategy.
# data_file_directories:
# - /data/cassandra
#
sudo mkdir -p /data/cassandra
HOST_IP=$(ifconfig | grep -Po "(?<=inet addr:)\s*[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}" | egrep -v "127\.0\.0\.1")
echo "Change the 'data_file_directories' to point to /data/cassandra,"
echo "the 'seeds' and 'listen_address' to point to $HOST_IP, at the very least"
#
# These are the manual changes necessary to make this all work:
# seed_provider:
# - seeds: "ip1,ip2"
# rpc_address: 0.0.0.0
# listen_address: ip1
# data_file_directories:
# - /data/cassandra
#
read -p "Press a key when ready"
sudo vim /etc/cassandra/cassandra.yaml
# 3. In a VPC with no public nodes, your hostname won't resolve and Cassandra
# will terminate the node
#
cp /etc/hosts ./
echo "$HOST_IP `hostname`" >> hosts
sudo cp ./hosts /etc/
echo "[INFO] Please verify that the last line added to your /etc/hosts file is correct"
cat /etc/hosts
read -p "Press a key when ready"
# The above MAY fail; the alternative is to change the
# listen_address: ip1
# to whatever the IP address of the host is.
# Verify that you can run Cassandra: the line below starts it in foreground mode (so you can see
# the logs); then open another shell and run `cqlsh` to start a cassandra shell.
# You will have to terminate this session via Ctrl-C
# sudo cassandra -f
sudo chown -R cassandra: /var/{lib,log}/cassandra /data/cassandra
sudo service cassandra start
sudo service cassandra status
echo "[INFO] Cassandra installation complete, please verify that you can connect
to the server using: cqlsh"
read -p "Press a key when ready"
## Install Apache Spark
# See: http://spark.apache.org/docs/1.2.0/spark-standalone.html
#
echo "[INFO] Installing Apache Spark"
read -p "Press a key when ready"
#
# Ubuntu repositories only have Scala 2.9 - we need to get and install 2.11 manually
#
wget http://downloads.typesafe.com/scala/2.11.5/scala-2.11.5.deb
sudo dpkg -i scala-2.11.5.deb
rm scala*
IS_SCALA=$(scalac -version | grep "2\.11")
if [[ -z ${IS_SCALA} ]]; then
echo "[ERROR] Scala 2.11 did not get properly installed, the installed (if any) scalac is: $(scalac -version)"
read -p "Hit Ctrl-C to quit now or any key to continue "
fi
wget http://apache.petsads.us/spark/spark-1.2.0/spark-1.2.0-bin-hadoop2.4.tgz
tar xfz spark-1.2.0-bin-hadoop2.4.tgz
sudo mv spark-1.2.0-bin-hadoop2.4 /usr/local/
sudo ln -snf spark-1.2.0-bin-hadoop2.4 /usr/local/spark
echo "export SPARK_HOME=/usr/local/spark
export PATH=\${PATH}:\${SPARK_HOME}/sbin:\${SPARK_HOME}/bin
" >> .bashrc
echo "[INFO] Apache Spark installed at ${SPARK_HOME}"
echo "[INFO] To start Spark Master, run: sudo start-master.sh"
# Adding some useful aliases
echo "alias mkvenv3='mkvirtualenv -p `which python3` --system-site-packages'
alias upgrade='sudo apt-get update && sudo apt-get -y upgrade'
alias notebook='ipython notebook --profile=nbserver --pylab=inline'" >>.bash_aliases
source .bashrc
# Finally create a "durable" development environment
sudo mkdir -p /data/dev
sudo chown -R :users /data/dev
sudo chmod -R 2775 /data/dev
echo "[SUCCESS] Done."
# SparkLab Python dev requirements
Jinja2==2.7.3
MarkupSafe==0.23
Pygments==2.0.2
blist==1.3.6
cassandra-driver==2.1.3
certifi==14.05.14
futures==2.2.0
ipython==2.3.1
# TODO: matplotlib fails if installed from here
# matplotlib==1.4.2
nose==1.3.4
numpy==1.9.1
pandas==0.15.2
pymongo==2.7.2
pyparsing==2.0.3
python-dateutil==2.4.0
pytz==2014.10
pyzmq==14.5.0
six==1.9.0
tornado==4.0.2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment