以下是在Ubuntu上对Hadoop数据进行备份的几种方法:
hdfs dfs -mkdir /source/path
和hdfs dfs -mkdir /destination/path
,然后执行hadoop distcp hdfs:///source/path hdfs:///destination/path
进行复制。hdfs dfsadmin -allowSnapshot /path/to/directory
,再创建快照hdfs dfs -createSnapshot /path/to/directory snapshotName
,恢复时用hdfs dfs -revertSnapshot /path/to/directory snapshotName
。sudo ambari - backup backup --target - dir=/backup/location
进行备份,恢复时用sudo ambari - backup restore --target - dir=/backup/location
。#!/bin/bash
,内容为rsync -avz --delete /source/path /backup/path
,再通过crontab -e
设置定时任务,如0 2 * * * /path/to/script.sh
实现定时备份。