1 前言

一个问题，一篇文章，一出故事。
笔者生产环境有一套Hadoop平台需要自动平衡数据节点的数据，于是整理此文。

2 最佳实践

2.1 手动执行平衡

2.1.1 获取命令帮助

sudo -u hdfs hdfs balancer -h

可见如下显示，

Usage: hdfs balancer
        [-policy ]      the balancing policy: datanode or blockpool
        [-threshold ]        Percentage of disk capacity
        [-exclude [-f  | ]]  Excludes the specified datanodes.
        [-include [-f  | ]]  Includes only the specified datanodes.
        [-idleiterations ]      Number of consecutive idle iterations (-1 for Infinite) before exit.
        [-runDuringUpgrade]     Whether to run the balancer during an ongoing HDFS upgrade.This is usually not desired since it will not affect used space on over-utilized machines.

Generic options supported are
-conf      specify an application configuration file
-D             use value for given property
-fs       specify a namenode
-jt     specify a ResourceManager
-files     specify comma separated files to be copied to the map reduce cluster
-libjars     specify comma separated jar files to include in the classpath.
-archives     specify comma separated archives to be unarchived on the compute machines.

The general command line syntax is
bin/hadoop command [genericOptions] [commandOptions]

[root@HD01 ~]# sudo -u hdfs hdfs balancer -h
Usage: hdfs balancer
        [-policy ]      the balancing policy: datanode or blockpool
        [-threshold ]        Percentage of disk capacity
        [-exclude [-f  | ]]  Excludes the specified datanodes.
        [-include [-f  | ]]  Includes only the specified datanodes.
        [-idleiterations ]      Number of consecutive idle iterations (-1 for Infinite) before exit.
        [-runDuringUpgrade]     Whether to run the balancer during an ongoing HDFS upgrade.This is usually not desired since it will not affect used space on over-utilized machines.

Generic options supported are
-conf      specify an application configuration file
-D             use value for given property
-fs       specify a namenode
-jt     specify a ResourceManager
-files     specify comma separated files to be copied to the map reduce cluster
-libjars     specify comma separated jar files to include in the classpath.
-archives     specify comma separated archives to be unarchived on the compute machines.

The general command line syntax is
bin/hadoop command [genericOptions] [commandOptions]

2.1.2 查询当前的集群数据节点

sudo -u hdfs hdfs dfsadmin -printTopology

可见如下显示，

Rack: /default
   10.168.0.102:50010 (hd07.cmdschool.org)
   10.168.0.105:50010 (hd22.cmdschool.org)
   10.168.0.153:50010 (hd10.cmdschool.org)
   10.168.0.19:50010 (hd23.cmdschool.org)
   10.168.0.215:50010 (hd06.cmdschool.org)
   10.168.0.22:50010 (hd08.cmdschool.org)
   10.168.0.23:50010 (hd09.cmdschool.org)
   10.168.0.24:50010 (hd01.cmdschool.org)
   10.168.0.25:50010 (hd02.cmdschool.org)
   10.168.0.26:50010 (hd03.cmdschool.org)

2.1.3 使用命令平衡集群数据节点

sudo -u hdfs hdfs balancer -threshold 5.0 -policy DataNode -include hd01.cmdschool.org,hd02.cmdschool.org,hd03.cmdschool.org,hd06.cmdschool.org,hd07.cmdschool.org,hd08.cmdschool.org,hd09.cmdschool.org,hd10.cmdschool.org,hd22.cmdschool.org,hd23.cmdschool.org

2.2 自动执行平衡

2.2.1 创建执行脚本

vim ~/scripts/hdfs-balancer.sh

加入如下代码，

#!/bin/bash
hdfsBalancerLog="/var/log/hdfsBalancerLog.log"

if [ `ps -ef | grep balancer | egrep -v "$0|grep" | wc -l` -gt "1" ]; then
        echo `date +'%Y-%m-%d %H:%M:%S'`" Already running!" >> "$hdfsBalancerLog"
        exit 1
fi
hdfsNodes=`sudo -u hdfs hdfs dfsadmin -printTopology | grep ":50010" | awk -F' ' '{print $2}' | sed -e 's/(//g' -e 's/)//g'`

if [ `echo "$hdfsNodes" | wc -l` -lt "3" ]; then
        echo `date +'%Y-%m-%d %H:%M:%S'`" Number is abnormal of nodes!" >> "$hdfsBalancerLog"
        exit 1
fi
balancerNodes=`echo $hdfsNodes | sed -e 's/ /,/g'`
echo `date +'%Y-%m-%d %H:%M:%S'`" Execution Started!" >> "$hdfsBalancerLog"
sudo -u hdfs hdfs balancer -threshold 5.0 -policy DataNode -include "$balancerNodes"
echo `date +'%Y-%m-%d %H:%M:%S'`" Execution Ended!" >> "$hdfsBalancerLog"

2.2.2 设置脚本触发

crontab -e

加入如下配置，

0 0 * * * sh ~/scripts/hdfs-balancer.sh

参阅文档
========================
https://hadoop.apache.org/docs/r2.7.3/hadoop-project-dist/hadoop-hdfs/HDFSCommands.html