准备工作:从hadoop集群找到core-site.xml,hdfs-site.xml,yarn-site.xml配置文件,放到项目对应资源配置目录下。这种方式属性配置全,自己配置容易遗漏配置,导致
一些无法预料的错误。
import org.apache.hadoop.conf.Configuration; import org.apache.hadoop.fs.FileSystem; import org.apache.hadoop.fs.Path; import org.apache.hadoop.io.IOUtils; import org.apache.hadoop.security.UserGroupInformation; import org.slf4j.Logger; import org.slf4j.LoggerFactory; import java.io.File; import java.io.FileOutputStream; import java.io.IOException; import java.io.InputStream; import java.io.OutputStream; public class FileSystemConnection {private static String KERBEROS_PRINC = "admin@TDH"; private static String KERBEROS_PRINC_KEYTAB_PATH = "D:\admin.keytab"; public static final Logger logger = LoggerFactory.getLogger(FileSystemConnection.class); public static FileSystem getHdfsConnection() { try { Configuration conf = new Configuration(); conf.addResource("core-site.xml");
conf.addResource("hdfs-site.xml");
conf.addResource("yarn-site.xml");
// 没开kerberos,注释下面两行 logger.info("kerberos principal: {} ,keytab path: {}", KERBEROS_PRINC, KERBEROS_PRINC_KEYTAB_PATH); UserGroupInformation.setConfiguration(conf); UserGroupInformation.loginUserFromKeytab(KERBEROS_PRINC, KERBEROS_PRINC_KEYTAB_PATH); FileSystem fs = FileSystem.get(conf); return fs; } catch (Exception e) { throw new RuntimeException("get hdfs connection failure", e); } } private static void downLoadFile() throws IOException {
FileSystem dfs=getHdfsConnection(); String src="/tmp/tdt"; String dest="D:\TestFs"; Path path = new Path(src); try { if (dfs.exists(path)) { File file = new File(dest + File.separator + path.getName()); file.createNewFile(); InputStream in = dfs.open(path); OutputStream out = new FileOutputStream(file); IOUtils.copyBytes(in, out, conf); in.close(); out.close(); } } catch (Exception e) { e.printStackTrace(); } } }