- 下載snappy源碼包
http://pkgs.fedoraproject.org/repo/pkgs/snappy/ - 編譯
對jdk1.8.0_191/jre/lib/amd64/server/libjvm.so二進(jìn)制包軟鏈到/usr/local/lib/libjvm.so
進(jìn)入對應(yīng)snappy包內(nèi)
./autogen.sh
./configure
make -j4 && make install - 編譯hadoop對應(yīng)的snappy包
下載
https://code.google.com/archive/p/hadoop-snappy/source/default/source
編譯
cd trunk
mvn clean package -DskipTests -Dsnappy.prefix=/usr/local/ - 復(fù)制hadoop的snappy二進(jìn)制包到hadoop的lib目錄
hadoop-snappy-0.0.1-SNAPSHOT.tar.gz解壓
cp -r hadoop-snappy-0.0.1-SNAPSHOT/lib/* <HADOOP_HOME>/lib - 修改環(huán)境變量
hadoop-env.sh
export LD_LIBRARY_PATH=LD_LIBRARY_PATH
export JAVA_LIBRARY_PATH=${HADOOP_HOME}/lib/native/Linux-amd64-64/
測試
/data/hadoop/bin/hadoop checknative -a
- 對hbase新增snappy壓縮
將hadoop-snappy-0.0.1-SNAPSHOT.jar復(fù)制到{HADOOP_HOME}/lib/native
LD_LIBRARY_PATH:
HBASE_LIBRARY_PATH:$HBASE_HOME/lib/native/
測試
bin/hbase --config /data/hbase/conf org.apache.hadoop.util.NativeLibraryChecker
