A robust and legendary analytics system of giant proportions said to have dwelt off the coasts of Norway and Iceland.

You'll need Java in at least version 6. For Ubuntu-like systems, you can use the instructions from the next item. For other distributions, please look up in your distribution's manual on how to install Java 6.

Add these two lines at the end of your /etc/apt/sources.list

sudo add-apt-repository ppa:webupd8team/java
sudo apt-get update
sudo apt-get install oracle-java6-installer

Now run java -version and make sure it says that you're running java 6.

If for some reason you're still running some other version do

sudo update-java-alternatives -s java-6-oracle

And finally for environment variables

sudo apt-get install oracle-java7-set-default

Run the following command to create a kraken subdirectory in the current folder and clone kraken into

git clone https://github.com/wikimedia/kraken.git

If your an a Debian system, you can probably just grab the most recent .deb files of libdclass0 and libdclass-jni for your architecture from Wikimedia's apt repository and install them using dpkg and can proceed to the next step.

If your not on a Debian system, you can try getting the needed files by hand or compiling dclass by yourself.

For an amd64 system with the most recent version being 2.2.0, that would be:

mkdir dclass
cd dclass

mkdir libdclass0
cd libdclass0
wget http://apt.wikimedia.org/wikimedia/pool/main/d/dclass/libdclass0_2.2.2-1_amd64.deb
ar x libdclass0_2.2.2-1_amd64.deb
tar -xzvf data.tar.gz
sudo mv usr/lib/x86_64-linux-gnu/libdclass.so.0.0.0 /usr/lib/libdclass.so.0
sudo chmod +x /usr/lib/libdclass.so.0
cd ..

mkdir libdclass-jni
cd libdclass-jni
wget http://apt.wikimedia.org/wikimedia/pool/main/d/dclass/libdclass-jni_2.2.2-1_amd64.deb
ar x libdclass-jni_2.2.2-1_amd64.deb
tar -xzvf data.tar.gz
sudo mv usr/lib/x86_64-linux-gnu/jni/libdclassjni.so.0.0.0 /usr/lib/libdclassjni.so
sudo chmod +x /usr/lib/libdclassjni.so
cd ..

cd ..

You can test whether or not it's working by doing

ldd /usr/lib/libdclassjni.so /usr/lib/libdclass.so.0

to see if the dynamic linker caught the files. You should get an output like:

/usr/lib/libdclassjni.so:
        linux-vdso.so.1 (0x00007f43caf4b000)
        libdclass.so.0 => /usr/lib64/libdclass.so.0 (0x00007f43cab05000)
        libc.so.6 => /lib64/libc.so.6 (0x00007f43ca75a000)
        librt.so.1 => /lib64/librt.so.1 (0x00007f43ca551000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f43caf4c000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f43ca334000)
/usr/lib/libdclass.so.0:
        linux-vdso.so.1 (0x00007fffdb3ff000)
        librt.so.1 => /lib64/librt.so.1 (0x00007f6948ba3000)
        libc.so.6 => /lib64/libc.so.6 (0x00007f69487f9000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f69485db000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f6948fef000)

(If numbers changed, that's not a problem. Only „Not Found” lines are a problem)

We do no longer need the dclass folder. You can remove the dclass folder again.

Instructions on how to compile dclass by yourself are at:

https://github.com/wikimedia/dClass/blob/package/README

Get a openddr.dtree file (version 1.13) from an Analytics Team member, and store it as

/usr/share/libdclass/openddr.dtree

Get the

• GeoIPCity.dat,
• GeoIP.dat,
• GeoIPRegion.dat, and
• GeoIPv6.dat

files from an Analytics Team member and store them in /usr/share/GeoIP.

If noone is available and you have access to the stat1 machine, you can find one GeoIPCity.dat in /home/spetrea/maxmind_archive/GeoIP_City*.gz.

Get the

• testdata_desktop.csv,
• testdata_mobile.csv, and
• testdata_session.csv

files from an Analytics Team member and store them in kraken-pig/src/test/resources in your clone of kraken.

To test whether things worked out, run

mvn test

If you want to start coding around Kraken, the Analytics team mostly uses IDEA from IntelliJ.

Get the IDE called IDEA from IntelliJ here.

The Community edition suffices.

Start IDEA. It will ask you which version of JDK you want to use, you will tell it you want jdk6 (not openjdk6, but jdk6 from Oracle).

Now import the Kraken project from where you cloned it.

Now you can run individual tests easily.

Further information about building kraken is available in the build.md here.