MacOS平台上编译 hadoop 3.1.2 源码

1. 先从官方下载源码:源码下载地址:https://hadoop.apache.org/releases.html,下载 3.1.2 版本

2. 解压缩源码:tar xvf hadoop-3.1.2-src.tar.gz

3. 安装依赖(注:由于我本机环境不是从零开始安装的,部分步骤省略掉):

 3.1  安装并配置 java 环境

 3.2 安装并配置 cmake 环境, 可以参照:https://blog.csdn.net/baimafujinji/article/details/78588488

 3.3 安装并配置 protobuf 环境 ,可以参照:https://www.jianshu.com/p/0a70eef1bd4a

 3.4 安装并配置 openssl 环境:brew install openssl

4. 编译生成安装包:mvn package -Pdist,native -DskipTests=true -Dtar

在编码过程中遇到的问题总结:

问题1 :cmake 函数库变更引起的问题:

问题描述:

[WARNING] /Users/wyx/Downloads/hadoop-rel-release-3.1.1/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/impl/utils/docker-util.c:1227:12: error: no matching function for call to 'getgrouplist'
[WARNING]   int rc = getgrouplist(user, pw->pw_gid, groups, &ngroups);
[WARNING]            ^~~~~~~~~~~~
[WARNING] /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk/usr/include/unistd.h:653:6: note: candidate function not viable: no known conversion from 'gid_t *' (aka 'unsigned int *') to 'int *' for 3rd argument
[WARNING] int      getgrouplist(const char *, int, int *, int *);

修改 c 文件:

vim hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/impl/utils/docker-util.c 中 getgrouplist函数调用的 groups参数,强转成(int *)

具体如下所示:

1384   int rc = getgrouplist(user, pw->pw_gid, (int *)groups, &ngroups);
1385   if (rc < 0) {
1386     groups = (gid_t *) alloc_and_clear_memory(ngroups, sizeof(gid_t));
1387     if (groups == NULL) {
1388       fprintf(ERRORFILE, "Failed to allocate buffer for group lookup for user %s.
", user);
1389       exit(OUT_OF_MEMORY);
1390     }
1391     if (getgrouplist(user, pw->pw_gid, (int * )groups, &ngroups) == -1) {
1392       fprintf(ERRORFILE, "Fail to lookup groups for user %s.
", user);
1393       ret = 2;
1394     }
1395   }

问题2:由于 openssl 环境变量引起的

问题描述:

[WARNING] CMake Error at /Applications/CMake.app/Contents/share/cmake-3.15/Modules/FindPackageHandleStandardArgs.cmake:137 (message):
[WARNING]   Could NOT find OpenSSL, try to set the path to OpenSSL root folder in the
[WARNING]   system variable OPENSSL_ROOT_DIR (missing: OPENSSL_INCLUDE_DIR)
[WARNING] Call Stack (most recent call first):
[WARNING]   /Applications/CMake.app/Contents/share/cmake-3.15/Modules/FindPackageHandleStandardArgs.cmake:378 (_FPHSA_FAILURE_MESSAGE)
[WARNING]   /Applications/CMake.app/Contents/share/cmake-3.15/Modules/FindOpenSSL.cmake:413 (find_package_handle_standard_args)
[WARNING]   CMakeLists.txt:24 (find_package)
[WARNING] 
[WARNING] 
[WARNING] -- Configuring incomplete, errors occurred!
[WARNING] See also "/Users/xxx/local/hadoop-3.1.2-src/hadoop-tools/hadoop-pipes/target/native/CMakeFiles/CMakeOutput.log".

  

由于我本机之前装过 openssl,但配置不完全,解决方案如下:

1. 先查看 openssl 的安装目录:

解决方案:

(base) bogon:hadoop-3.1.2-src johnnybai$ brew list openssl
/usr/local/Cellar/openssl/1.0.2o_1/.bottle/etc/ (8 files)
/usr/local/Cellar/openssl/1.0.2o_1/bin/c_rehash
/usr/local/Cellar/openssl/1.0.2o_1/bin/openssl
/usr/local/Cellar/openssl/1.0.2o_1/include/openssl/ (74 files)
/usr/local/Cellar/openssl/1.0.2o_1/lib/libcrypto.1.0.0.dylib
/usr/local/Cellar/openssl/1.0.2o_1/lib/libssl.1.0.0.dylib
/usr/local/Cellar/openssl/1.0.2o_1/lib/engines/ (12 files)
/usr/local/Cellar/openssl/1.0.2o_1/lib/pkgconfig/ (3 files)
/usr/local/Cellar/openssl/1.0.2o_1/lib/ (4 other files)
/usr/local/Cellar/openssl/1.0.2o_1/share/man/ (1680 files)

2. 配置 openssl 环境变量

sudo vim /etc/profile
添加如下环境变量:
# openssl
export OPENSSL_ROOT_DIR="/usr/local/Cellar/openssl/1.0.2o_1"
export OPENSSL_INCLUDE_DIR="/usr/local/Cellar/openssl/1.0.2o_1/include"

编译成功之后,可以在 hadoop-3.1.2-src/hadoop-dist/target/ 目录下找到安装部署binary 文件: hadoop-3.1.2.tar.gz

注: 编译步骤,部分参照了 https://github.com/shawfdong/hyades/wiki/Building-Hadoop-from-source 的做法。

原文地址:https://www.cnblogs.com/johnny666888/p/11054253.html