安装Hadoop时Unable to load native-hadoop library for your platform报错原因

刚安装完hadoop后，出现了如下报错

[[email protected] ~]# /usr/local/hadoop/sbin/start-dfs.sh
20/06/04 08:22:37 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Starting namenodes on [localhost]
localhost: namenode running as process 19541. Stop it first.
localhost: datanode running as process 19667. Stop it first.
Starting secondary namenodes [0.0.0.0]
0.0.0.0: secondarynamenode running as process 19839. Stop it first.
20/06/04 08:22:40 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable

意思是：WARN级别的警告，无法为平台加载本地hadoop库…可使用合适的内置java类。

这个时候使用jps指令可以检查NameNode和DateNode是否正常启动。

[[email protected] ~]# jps
19667 DataNode
19541 NameNode
19097 Jps
19839 SecondaryNameNode

可能的原因:
Apache提供的hadoop本地库是32位的，而在64位的服务器上就会有问题，因此需要自己编译64位的版本。
解决方法
1、首先找到对应自己hadoop版本的64位的lib包，可以自己手动去编译
2、可以上网下载对应的编译版本
（如http://dl.bintray.com/sequenceiq/sequenceiq-bin/）
将准备好的64位的lib包解压到已经安装好的hadoop安装目录的lib/native 和 lib目录下

[[email protected] ~]# tar -xvf hadoop-native-64-2.7.0.tar -C /usr/local/hadoop/lib/native

[[email protected] ~]# tar -xvf hadoop-native-64-2.7.0.tar -C /usr/local/hadoop/lib

添加环境变量

[[email protected] ~]# vim /etc/profile
在/etc/profile中添加以下内容

export HADOOP_COMMON_LIB_NATIVE_DIR=$HADOOP_HOME/lib/native
 
export HADOOP_OPTS="-Djava.library.path=$HADOOP_HOME/lib"

让环境变量生效

[[email protected] ~]# source /etc/profile 

再尝试一次/usr/local/hadoop/sbin/start-dfs.sh

[[email protected] native]# /usr/local/hadoop/sbin/start-dfs.sh
Starting namenodes on [localhost]
localhost: namenode running as process 19541. Stop it first.
localhost: datanode running as process 19667. Stop it first.
Starting secondary namenodes [0.0.0.0]
0.0.0.0: secondarynamenode running as process 19839. Stop it first.

没有报错了

hadoop安装参考https://www.cnblogs.com/StarZhai/p/11712074.html