# checkout the LZO-related code from Google Code<\/span><\/div>svn checkout http://hadoop-gpl-compression.googlecode.com/svn/trunk/ hadoop-gpl-compression<\/div>
<\/div>
# build the Java code<\/span><\/div>cd <\/span>hadoop-gpl-compression<\/div>ant<\/div>
<\/div>
# move the jar file to $HADOOP_HOME/lib<\/span><\/div>mv build/hadoop-gpl-compression-0.1.0-dev.jar
$HADOOP_HOME<\/span>/lib/<\/div> <\/div>
# build the LZO native C code<\/span><\/div>cd <\/span>src/native<\/div>sudo apt-get install liblzo2-dev<\/div>
# indicate if your JVM is 32-bit or 64-bit<\/span><\/div>export <\/span>JVM_DATA_MODEL<\/span>=<\/span>32<\/div>./configure<\/div>
make<\/div>
cp -rf .libs/libgplcompression.*
$HADOOP_HOME<\/span>/lib/Linux-i386-32/<\/div> <\/div>
# now you're good to go. launch a test job to verify if LZO compression is working fine.<\/span><\/div>$HADOOP_HOME<\/span>/bin/hadoop jar $HADOOP_HOME<\/span>/hadoop-0.20.0-examples.jar randomtextwriter -Dtest.randomtextwrite.bytes_per_map=<\/span>104857600 -Dtest.randomtextwrite.maps_per_host=<\/span>3 -Dmapred.output.compress=<\/span>true<\/span> -Dmapred.output.compression.type=<\/span>BLOCK -Dmapred.output.compression.codec=<\/span>com.hadoop.compression.lzo.LzopCodec randomtext_BLOCK_LzopCodec<\/div><\/pre><\/div>\n \n <\/div>\n\n