百般波折的Hadoop编译

终于不能忍了，每次使用hadoop都出native库警告，于是决定编译hadoop，也是走了不少弯路。唯一参考了官方文档，当初居然没想到去百度一篇别人编译的过程，先填坑，哎~ 说多了都是泪。

准备

首先从官方下载hadoop2.6.4的源码包（Hadoop分布式集群安装一文有介绍），同时查看了官方文档对于Native Bulid的需求描述：

• C compiler (e.g. GNU C Compiler)
• GNU Autools Chain: autoconf, automake, libtool
• zlib-development package (stable version >= 1.2.0)
• openssl-development package(e.g. libssl-dev)
• 先查询系统是否有这些环境库：

rpm -qa | grep -E "gcc|autoconf|automake|libtool|libssl-dev"

zlib是有的，版本1.2.3，达到要求，下面开始安装没有的gcc、cmake、autoconf和automake，以及libssl-dev

yum install openssl-devel gcc cmake autoconf automake

libssl-dev

由于libssl-dev在yum安装时叫openssl-devel，pkgs上的rpm包版本是1.0.2i，但使用命令openssl version查看系统当前openssl的版本是1.0.1e，与要安装的libssl-dev包不符，需要升级openssl，下载openssl源码包编译

wget https://www.openssl.org/source/openssl-1.0.2i.tar.gz

解压，进入目录运行config可执行文件，报错

c_zlib.c:25:19: error: zlib.h: No such file or directory

zlib.h不存在，那就是缺少zlib-dev的包，yum install zlib-dev，没有这个包？yum search zlib，原来叫zlib-devel，安装完成后重新运行config，通过，然后make和make install

完成后查看版本，openssl version，版本居然还是1.0.1e，哪里不对，翻上去看看make install时的输出，有下面一段：

created directory `/usr/local/ssl/bin’
created directory `/usr/local/ssl/lib’
created directory `/usr/local/ssl/lib/engines’
created directory `/usr/local/ssl/lib/pkgconfig’
created directory `/usr/local/ssl/include’
created directory `/usr/local/ssl/include/openssl’
created directory `/usr/local/ssl/misc’
created directory `/usr/local/ssl/certs’
created directory `/usr/local/ssl/private’

哦，原来是安装在/usr/local/ssl目录下了，运行/usr/local/ssl/bin/openssl version，版本是1.0.2i，那好办

mv /usr/bin/openssl /usr/bin/openssl.old
ln -s /usr/local/ssl/bin/openssl /usr/bin/openssl
ln -s /usr/local/ssl/include/openssl /usr/include/openssl
echo '/usr/local/ssl/lib'>> /etc/ld.so.conf
ldconfig -v

上述命令依次备份了旧版本的openssl，但由于/usr/include下没有openssl的头文件库，就不去管它，接着在/usr/bin下创建名为openssl的软连接到/usr/local/ssl/bin/openssl以代替原来的旧版本，
同理，头文件库也是。最后别忘了动态共享库，把openssl新的lib路径加到Id.so.conf中，执行Idconfig命令生成缓存，现在再openssl version一下，版本变1.0.2i了。

编译

按照官方文档说的，输入以下命令执行编译：

mvn package -Pdist,native -DskipTests -Dtar

一看mvn命令，maven环境的安装也就不多说了。

再次运行编译。

[ERROR] Failed to execute goal org.apache.hadoop:hadoop-maven-plugins:2.6.4:protoc (compile-protoc) on project hadoop-common: org.apache.maven.plugin.MojoExecutionException: ‘protoc –version’ did not return a version -> [Help 1]

没有protocol Buffer compiler，下载一个，在官方github上，我选择了2.6.1版本的https://github.com/google/protobuf/tree/v2.6.1

下载解压后运行./autogen.sh

Google Test not present.  Fetching gtest-1.5.0 from the web…
curl: (7) Failed to connect to 2404:6800:4008:c01::52: Network is unreachable
bzip2: Compressed file ends unexpectedly;
perhaps it is corrupted?  *Possible* reason follows.
bzip2: Invalid argument
Input file = (stdin), output file = (stdout)
It is possible that the compressed file(s) have become corrupted.
You can use the -tvv option to test integrity of such files.
You can use the `bzip2recover’ program to attempt to recover
data from undamaged sections of corrupted files.
tar: Child returned status 2
tar: Error is not recoverable: exiting now

打开脚本文件看看是怎么回事。

if test ! -e gtest; then
  echo "Google Test not present.  Fetching gtest-1.5.0 from the web..."
  curl http://googletest.googlecode.com/files/gtest-1.5.0.tar.bz2 | tar jx
  mv gtest-1.5.0 gtest
fi

浏览器打开这个链接果然是404（挂了代理的），看样子是要下载一个gtest-1.5.0的tar包，上github找找，原来是google test，https://github.com/google/googletest/tree/release-1.5.0

下载这个版本的tar包，解压后将文件夹重命名为gtest（注意是在当前protobuf的目录下），进入make目录输入make编译。

g++ -I../include -g -Wall -Wextra -c ../samples/sample1.cc
make: g++: Command not found
make: *** [sample1.o] Error 127

好吧，缺c++的编译环境，yum install gcc-c++ ，再次运行编译，成功。返回protobuf的目录，重新运行autogen.sh

+ exit 0

看到结束状态为0，成功。在目录下多了一个configure可执行文件，运行它。

config.status: creating Makefile
config.status: creating scripts/gtest-config
config.status: creating build-aux/config.h
config.status: executing depfiles commands
config.status: executing libtool commands

最后这几行输出很重要，如果没有可能是前面某环节出错了，此时就可以make了，编译开始，输入make，经过几分钟等待，编译完成。make check下，看到一大堆的OK和PASSED就说明没问题了，
现在安装它，make install，完成，返回hadoop的编译，重新执行命令mvn package -Pdist,native -DskipTests -Dtar

[ERROR] Failed to execute goal org.apache.hadoop:hadoop-maven-plugins:2.6.4:protoc (compile-protoc) on project hadoop-common: org.apache.maven.plugin.MojoExecutionException: protoc version is ‘libprotoc 2.6.1’, expected version is ‘2.5.0’ -> [Help 1]

WTF？你官方文档也没说啊！

先去protobuf目录make uninstall，再去下载2.5.0版本的https://github.com/google/protobuf/archive/v2.5.0.zip

好在2.6.1编译时的坑都填了，解压后把原来2.6.1版本的目录下的gtest文件夹拷到2.5.0版本的目录下，依旧是按步骤执行autogen.sh 、configure、 make 、make check、 make install

一切OK，没有任何问题，还是确认一下版本，输入命令protoc --version

libprotoc 2.5.0

没问题，再次回到hadoop编译，运行mvn package的命令。

[ERROR] /home/bigdata/hadoop-2.6.4-src/hadoop-common-project/hadoop-nfs/src/main/java/org/apache/hadoop/nfs/nfs3/Nfs3Interface.java:42: warning: no @return
[ERROR] public NFS3Response access(XDR xdr, RpcInfo info);
[ERROR] ^
[ERROR] /home/bigdata/hadoop-2.6.4-src/hadoop-common-project/hadoop-nfs/src/main/java/org/apache/hadoop/nfs/nfs3/Nfs3Interface.java:45: warning: no @param for xdr
[ERROR] public NFS3Response readlink(XDR xdr, RpcInfo info);
[ERROR] ^
[ERROR]
[ERROR] Command line was: /usr/local/jdk1.8.0_101/jre/../bin/javadoc @options @packages
[ERROR]
[ERROR] Refer to the generated Javadoc files in ‘/home/bigdata/hadoop-2.6.4-src/hadoop-common-project/hadoop-nfs/target’ dir.
[ERROR] -> [Help 1]
[ERROR]
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MojoExecutionException
[ERROR]
[ERROR] After correcting the problems, you can resume the build with the command
[ERROR]   mvn <goals> -rf :hadoop-nfs

javadoc生成出错？不要了，跳过它，加-Dmaven.javadoc.skip=true 跳过javadoc生成，重新运行编译。

[INFO] ————————————————————————
[INFO] BUILD SUCCESS
[INFO] ————————————————————————
[INFO] Total time: 03:24 h
[INFO] Finished at: 2016-09-25T14:55:23+08:00
[INFO] Final Memory: 94M/206M
[INFO] ————————————————————————

看到这段，激动不已，终于成功了，到hadoop-dist/target/hadoop-2.6.4/lib/native目录下拷走native库到hadoop的native库目录下，然后终于没有警告了，感觉世界清静了呢~