HDFS C API接口介绍
功能简介
C语言应用开发代码样例中所涉及的文件操作主要包括创建文件、读写文件、追加文件、删除文件。完整和详细的接口请直接参考官网上的描述以了解其使用方法:
http://hadoop.apache.org/docs/r3.1.1/hadoop-project-dist/hadoop-hdfs/LibHdfs.html。
代码样例
下面代码片段仅为演示,具体代码请参见获取样例代码解压目录中“hdfs-c-example/hdfs_test.c”文件。
- 设置HDFS NameNode参数,建立HDFS文件系统连接。
hdfsFS fs = hdfsConnect("default", 0); fprintf(stderr, "hdfsConnect- SUCCESS!\n");
- 创建HDFS目录。
const char* dir = "/tmp/nativeTest"; int exitCode = hdfsCreateDirectory(fs, dir); if( exitCode == -1 ){ fprintf(stderr, "Failed to create directory %s \n", dir ); exit(-1); } fprintf(stderr, "hdfsCreateDirectory- SUCCESS! : %s\n", dir);
- 写文件。
const char* file = "/tmp/nativeTest/testfile.txt"; hdfsFile writeFile = openFile(fs, (char*)file, O_WRONLY |O_CREAT, 0, 0, 0); fprintf(stderr, "hdfsOpenFile- SUCCESS! for write : %s\n", file); if(!hdfsFileIsOpenForWrite(writeFile)){ fprintf(stderr, "Failed to open %s for writing.\n", file); exit(-1); } char* buffer = "Hadoop HDFS Native file write!"; hdfsWrite(fs, writeFile, (void*)buffer, strlen(buffer)+1); fprintf(stderr, "hdfsWrite- SUCCESS! : %s\n", file); printf("Flushing file data ....\n"); if (hdfsFlush(fs, writeFile)) { fprintf(stderr, "Failed to 'flush' %s\n", file); exit(-1); } hdfsCloseFile(fs, writeFile); fprintf(stderr, "hdfsCloseFile- SUCCESS! : %s\n", file);
- 读文件。
hdfsFile readFile = openFile(fs, (char*)file, O_RDONLY, 100, 0, 0); fprintf(stderr, "hdfsOpenFile- SUCCESS! for read : %s\n", file); if(!hdfsFileIsOpenForRead(readFile)){ fprintf(stderr, "Failed to open %s for reading.\n", file); exit(-1); } buffer = (char *) malloc(100); tSize num_read = hdfsRead(fs, readFile, (void*)buffer, 100); fprintf(stderr, "hdfsRead- SUCCESS!, Byte read : %d, File contant : %s \n", num_read ,buffer); hdfsCloseFile(fs, readFile);
- 指定位置开始读文件。
buffer = (char *) malloc(100); readFile = openFile(fs, file, O_RDONLY, 100, 0, 0); if (hdfsSeek(fs, readFile, 10)) { fprintf(stderr, "Failed to 'seek' %s\n", file); exit(-1); } num_read = hdfsRead(fs, readFile, (void*)buffer, 100); fprintf(stderr, "hdfsSeek- SUCCESS!, Byte read : %d, File seek contant : %s \n", num_read ,buffer); hdfsCloseFile(fs, readFile);
- 复制文件。
const char* destfile = "/tmp/nativeTest/testfile1.txt"; if (hdfsCopy(fs, file, fs, destfile)) { fprintf(stderr, "File copy failed, src : %s, des : %s \n", file, destfile); exit(-1); } fprintf(stderr, "hdfsCopy- SUCCESS!, File copied, src : %s, des : %s \n", file, destfile);
- 移动文件。
const char* mvfile = "/tmp/nativeTest/testfile2.txt"; if (hdfsMove(fs, destfile, fs, mvfile )) { fprintf(stderr, "File move failed, src : %s, des : %s \n", destfile , mvfile); exit(-1); } fprintf(stderr, "hdfsMove- SUCCESS!, File moved, src : %s, des : %s \n", destfile , mvfile);
- 重命名文件。
const char* renamefile = "/tmp/nativeTest/testfile3.txt"; if (hdfsRename(fs, mvfile, renamefile)) { fprintf(stderr, "File rename failed, Old name : %s, New name : %s \n", mvfile, renamefile); exit(-1); } fprintf(stderr, "hdfsRename- SUCCESS!, File renamed, Old name : %s, New name : %s \n", mvfile, renamefile);
- 删除文件。
if (hdfsDelete(fs, renamefile, 0)) { fprintf(stderr, "File delete failed : %s \n", renamefile); exit(-1); } fprintf(stderr, "hdfsDelete- SUCCESS!, File deleted : %s\n",renamefile);
- 设置副本数。
if (hdfsSetReplication(fs, file, 10)) { fprintf(stderr, "Failed to set replication : %s \n", file ); exit(-1); } fprintf(stderr, "hdfsSetReplication- SUCCESS!, Set replication 10 for %s\n",file);
- 设置用户、用户组。
if (hdfsChown(fs, file, "root", "root")) { fprintf(stderr, "Failed to set chown : %s \n", file ); exit(-1); } fprintf(stderr, "hdfsChown- SUCCESS!, Chown success for %s\n",file);
- 设置权限。
if (hdfsChmod(fs, file, S_IRWXU | S_IRWXG | S_IRWXO)) { fprintf(stderr, "Failed to set chmod: %s \n", file ); exit(-1); } fprintf(stderr, "hdfsChmod- SUCCESS!, Chmod success for %s\n",file);
- 设置文件时间。
struct timeval now; gettimeofday(&now, NULL); if (hdfsUtime(fs, file, now.tv_sec, now.tv_sec)) { fprintf(stderr, "Failed to set time: %s \n", file ); exit(-1); } fprintf(stderr, "hdfsUtime- SUCCESS!, Set time success for %s\n",file);
- 获取文件信息。
hdfsFileInfo *fileInfo = NULL; if((fileInfo = hdfsGetPathInfo(fs, file)) != NULL) { printFileInfo(fileInfo); hdfsFreeFileInfo(fileInfo, 1); fprintf(stderr, "hdfsGetPathInfo - SUCCESS!\n"); }
- 遍历目录。
hdfsFileInfo *fileList = 0; int numEntries = 0; if((fileList = hdfsListDirectory(fs, dir, &numEntries)) != NULL) { int i = 0; for(i=0; i < numEntries; ++i) { printFileInfo(fileList+i); } hdfsFreeFileInfo(fileList, numEntries); } fprintf(stderr, "hdfsListDirectory- SUCCESS!, %s\n", dir);
- stream builder接口。
buffer = (char *) malloc(100); struct hdfsStreamBuilder *builder= hdfsStreamBuilderAlloc(fs, (char*)file, O_RDONLY); hdfsStreamBuilderSetBufferSize(builder,100); hdfsStreamBuilderSetReplication(builder,20); hdfsStreamBuilderSetDefaultBlockSize(builder,10485760); readFile = hdfsStreamBuilderBuild(builder); num_read = hdfsRead(fs, readFile, (void*)buffer, 100); fprintf(stderr, "hdfsStreamBuilderBuild- SUCCESS! File read success. Byte read : %d, File contant : %s \n", num_read ,buffer); struct hdfsReadStatistics *stats = NULL; hdfsFileGetReadStatistics(readFile, &stats); fprintf(stderr, "hdfsFileGetReadStatistics- SUCCESS! totalBytesRead : %" PRId64 ", totalLocalBytesRead : %" PRId64 ", totalShortCircuitBytesRead : %" PRId64 ", totalZeroCopyBytesRead : %" PRId64 "\n", stats->totalBytesRead , stats->totalLocalBytesRead, stats->totalShortCircuitBytesRead, stats->totalZeroCopyBytesRead); hdfsFileFreeReadStatistics(stats); free(buffer);
- 断开HDFS文件系统连接。
hdfsDisconnect(fs);
准备运行环境
在节点上安装客户端,例如安装到“/opt/client”目录。
Linux中编译并运行程序
- 进入Linux客户端目录,运行如下命令导入公共环境变量:
sourcebigdata_env
- 在该目录下用hdfs用户进行命令行认证,用户密码请咨询集群管理员。
- 进入“/opt/client/HDFS/hadoop/hdfs-c-example”目录下,运行如下命令导入客户端环境变量。
cd/opt/client/HDFS/hadoop/hdfs-c-example
sourcecomponent_env_C_example
- 清除之前运行生成的目标文件和可执行文件,运行如下命令。
执行结果如下:
[root@10-120-85-2 hdfs-c-example]# make clean rm -f hdfs_test.o rm -f hdfs_test
- 编译生成新的目标和可执行文件,运行如下命令。
执行结果如下:
[root@10-120-85-2 hdfs-c-example]# make all cc -c -I/opt/client/HDFS/hadoop/include -Wall -o hdfs_test.o hdfs_test.c cc -o hdfs_test hdfs_test.o -lhdfs
- 运行文件以实现创建文件、读写追加文件和删除文件的功能,运行如下命令。
执行结果如下:
[root@10-120-85-2 hdfs-c-example]# make run ./hdfs_test hdfsConnect- SUCCESS! hdfsCreateDirectory- SUCCESS! : /tmp/nativeTest hdfsOpenFile- SUCCESS! for write : /tmp/nativeTest/testfile.txt hdfsWrite- SUCCESS! : /tmp/nativeTest/testfile.txt Flushing file data .... hdfsCloseFile- SUCCESS! : /tmp/nativeTest/testfile.txt hdfsOpenFile- SUCCESS! for read : /tmp/nativeTest/testfile.txt hdfsRead- SUCCESS!, Byte read : 31, File contant : Hadoop HDFS Native file write! hdfsSeek- SUCCESS!, Byte read : 21, File seek contant : S Native file write! hdfsPread- SUCCESS!, Byte read : 10, File pead contant : S Native f hdfsCopy- SUCCESS!, File copied, src : /tmp/nativeTest/testfile.txt, des : /tmp/nativeTest/testfile1.txt hdfsMove- SUCCESS!, File moved, src : /tmp/nativeTest/testfile1.txt, des : /tmp/nativeTest/testfile2.txt hdfsRename- SUCCESS!, File renamed, Old name : /tmp/nativeTest/testfile2.txt, New name : /tmp/nativeTest/testfile3.txt hdfsDelete- SUCCESS!, File deleted : /tmp/nativeTest/testfile3.txt hdfsSetReplication- SUCCESS!, Set replication 10 for /tmp/nativeTest/testfile.txt hdfsChown- SUCCESS!, Chown success for /tmp/nativeTest/testfile.txt hdfsChmod- SUCCESS!, Chmod success for /tmp/nativeTest/testfile.txt hdfsUtime- SUCCESS!, Set time success for /tmp/nativeTest/testfile.txt Name: hdfs://hacluster/tmp/nativeTest/testfile.txt, Type: F, Replication: 10, BlockSize: 134217728, Size: 31, LastMod: 1500345260, Owner: root, Group: root, Permissions: 511 (rwxrwxrwx) hdfsGetPathInfo - SUCCESS! Name: hdfs://hacluster/tmp/nativeTest/testfile.txt, Type: F, Replication: 10, BlockSize: 134217728, Size: 31, LastMod: 1500345260, Owner: root, Group: root, Permissions: 511 (rwxrwxrwx) hdfsListDirectory- SUCCESS!, /tmp/nativeTest hdfsTruncateFile- SUCCESS!, /tmp/nativeTest/testfile.txt Block Size : 134217728 hdfsGetDefaultBlockSize- SUCCESS! Block Size : 134217728 for file /tmp/nativeTest/testfile.txt hdfsGetDefaultBlockSizeAtPath- SUCCESS! HDFS Capacity : 102726873909 hdfsGetCapacity- SUCCESS! HDFS Used : 4767076324 hdfsGetCapacity- SUCCESS! hdfsExists- SUCCESS! /tmp/nativeTest/testfile.txt hdfsConfGetStr- SUCCESS : hdfs://hacluster hdfsStreamBuilderBuild- SUCCESS! File read success. Byte read : 31, File contant : Hadoop HDFS Native file write! hdfsFileGetReadStatistics- SUCCESS! totalBytesRead : 31, totalLocalBytesRead : 0, totalShortCircuitBytesRead : 0, totalZeroCopyBytesRead : 0
- 进入debug模式(可选)
执行结果如下:
[root@10-120-85-2 hdfs-c-example]# make gdb gdb hdfs_test GNU gdb (GDB) SUSE (7.5.1-0.7.29) Copyright (C) 2012 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-suse-linux". For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>... Reading symbols from /opt/client/HDFS/hadoop/hdfs-c-example/hdfs_test...done. (gdb)