Hadoop集群节点的动态增加与删除

http://heylinux.com/archives/2047.html

Hadoop集群节点的动态增加与删除

Hadoop集群节点的动态增加
1. 安装配置节点
具体过程参考 《Hadoop集群实践 之 (1) Hadoop(HDFS)搭建》

2. 在配置过程中需要在所有的Hadoop服务器上更新以下三项配置
$ sudo vim /etc/hadoop/conf/slaves

1 hadoop-node-1
2 hadoop-node-2
3 hadoop-node-3

$ sudo vim /etc/hosts

1 10.6.1.150 hadoop-master
2 10.6.1.151 hadoop-node-1
3 10.6.1.152 hadoop-node-2
4 10.6.1.153 hadoop-node-3

$ sudo vim /etc/hadoop/conf/hdfs-site.xml

01 <?xml version="1.0"?>
02 <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
03
04 <configuration>
05 <property>
06   <name>dfs.data.dir</name>
07   <value>/data/hdfs</value>
08 </property>
09 <property>
10   <name>dfs.replication</name>
11   <value>3</value>
12 </property>
13 <property>
14   <name>dfs.datanode.max.xcievers</name>
15   <value>4096</value>
16 </property>
17 </configuration>

3. 启动datanode与tasktracker
dongguo@hadoop-node-3:~$ sudo /etc/init.d/hadoop-0.20-datanode start
dongguo@hadoop-node-3:~$ sudo /etc/init.d/hadoop-0.20-tasktracker start

4. 检查新增节点是否已经Live
通过WEB管理界面查看

http://10.6.1.150:50070/dfsnodelist.jsp?whatNodes=LIVE

可以看到hadoop-node-3已经被动态添加到了Hadoop集群中

5.应用新的备份系数dfs.replication

5.1 检查目前的备份系数
dongguo@hadoop-master:~$ sudo -u hdfs hadoop fs -lsr /dongguo
-rw-r--r-- 2 hdfs supergroup 33 2012-10-07 22:02 /dongguo/hello.txt

结果行中的第2列是备份系数（注：文件夹信息存储在namenode节点上，没有备份，故文件夹的备份系数是横杠-）
目前文件的备份系数仍是之前设置的参数2，Hadoop不会自动的按照新的备份系数进行调整。

dongguo@hadoop-master:~$ sudo -u hdfs hadoop fsck /

01 12/10/10 21:18:32 INFO security.UserGroupInformation: JAAS Configuration already setup for Hadoop, not re-installing.
02 FSCK started by hdfs (auth:SIMPLE) from /10.6.1.150 for path / at Wed Oct 10 21:18:33 CST 2012
03 .................Status: HEALTHY
04  Total size:    7786 B
05  Total dirs:    46
06  Total files:   17
07  Total blocks (validated):  17 (avg. block size 458 B)
08  Minimally replicated blocks:   17 (100.0 %)
09  Over-replicated blocks:    0 (0.0 %)
10  Under-replicated blocks:   0 (0.0 %)
11  Mis-replicated blocks:     0 (0.0 %)
12  Default replication factor:    2
13  Average block replication: 2.0
14  Corrupt blocks:        0
15  Missing replicas:      0 (0.0 %)
16  Number of data-nodes:      4
17  Number of racks:       1
18 FSCK ended at Wed Oct 10 21:18:33 CST 2012 in 48 milliseconds
19 The filesystem under path '/' is HEALTHY

通过 hadoop fsck / 也可以方便的看到Average block replication的值仍然为旧值2，该值我们可以手动的进行动态修改。
而Default replication factor则需要重启整个Hadoop集群才能修改，但实际影响系统的还是Average block replication的值，因此并非一定要修改默认值。

5.2 修改hdfs文件备份系数，把/ 目录下所有文件备份系数设置为3
dongguo@hadoop-master:~$ sudo -u hdfs hadoop dfs -setrep -w 3 -R /

01 12/10/10 21:22:35 INFO security.UserGroupInformation: JAAS Configuration already setup for Hadoop, not re-installing.
02 Replication 3 set: hdfs://hadoop-master/dongguo/hello.txt
03 Replication 3 set: hdfs://hadoop-master/hbase/-ROOT-/70236052/.oldlogs/hlog.1349695889266
04 Replication 3 set: hdfs://hadoop-master/hbase/-ROOT-/70236052/.regioninfo
05 Replication 3 set: hdfs://hadoop-master/hbase/-ROOT-/70236052/info/7670471048629837399
06 Replication 3 set: hdfs://hadoop-master/hbase/.META./1028785192/.oldlogs/hlog.1349695889753
07 Replication 3 set: hdfs://hadoop-master/hbase/.META./1028785192/.regioninfo
08 Replication 3 set: hdfs://hadoop-master/hbase/.META./1028785192/info/7438047560768966146
09 Waiting for hdfs://hadoop-master/dongguo/hello.txt .... done
10 Waiting for hdfs://hadoop-master/hbase/-ROOT-/70236052/.oldlogs/hlog.1349695889266... done
11 Waiting for hdfs://hadoop-master/hbase/-ROOT-/70236052/.regioninfo ... done
12 Waiting for hdfs://hadoop-master/hbase/-ROOT-/70236052/info/7670471048629837399 ...done
13 Waiting for hdfs://hadoop-master/hbase/.META./1028785192/.oldlogs/hlog.1349695889753... done
14 Waiting for hdfs://hadoop-master/hbase/.META./1028785192/.regioninfo ... done
15 Waiting for hdfs://hadoop-master/hbase/.META./1028785192/info/7438047560768966146 ...done
16 ...

可以看到Hadoop对所有文件的备份系数进行了刷新

5.3 再次检查备份系数的情况
dongguo@hadoop-master:~$ sudo -u hdfs hadoop fsck /

01 12/10/10 21:23:26 INFO security.UserGroupInformation: JAAS Configuration already setup for Hadoop, not re-installing.
02 FSCK started by hdfs (auth:SIMPLE) from /10.6.1.150 for path / at Wed Oct 10 21:23:27 CST 2012
03 .................Status: HEALTHY
04  Total size:    7786 B
05  Total dirs:    46
06  Total files:   17
07  Total blocks (validated):  17 (avg. block size 458 B)
08  Minimally replicated blocks:   17 (100.0 %)
09  Over-replicated blocks:    0 (0.0 %)
10  Under-replicated blocks:   0 (0.0 %)
11  Mis-replicated blocks:     0 (0.0 %)
12  Default replication factor:    2
13  Average block replication: 3.0
14  Corrupt blocks:        0
15  Missing replicas:      0 (0.0 %)
16  Number of data-nodes:      4
17  Number of racks:       1
18 FSCK ended at Wed Oct 10 21:23:27 CST 2012 in 11 milliseconds
19 The filesystem under path '/' is HEALTHY

可以看到已经变成了新的备份系数"3"

5.4 测试一下创建新的文件时是否能集成新的备份系数
dongguo@hadoop-master:~$ sudo -u hdfs hadoop fs -copyFromLocal mysql-connector-java-5.1.22.tar.gz /dongguo
dongguo@hadoop-master:~$ sudo -u hdfs hadoop fs -lsr /dongguo

1 -rw-r--r--   3 hdfs supergroup         33 2012-10-07 22:02 /dongguo/hello.txt
2 -rw-r--r--   3 hdfs supergroup    4028047 2012-10-10 21:28 /dongguo/mysql-connector-java-5.1.22.tar.gz

可以看到新上传的文件的备份系数是"3"

6 对HDFS中的文件进行负载均衡
dongguo@hadoop-master:~$ sudo -u hdfs hadoop balancer

1 Time Stamp               Iteration#  Bytes Already Moved  Bytes Left To Move  Bytes Being Moved
2 12/10/10 21:30:25 INFO net.NetworkTopology: Adding a new node: /default-rack/10.6.1.153:50010
3 12/10/10 21:30:25 INFO net.NetworkTopology: Adding a new node: /default-rack/10.6.1.150:50010
4 12/10/10 21:30:25 INFO net.NetworkTopology: Adding a new node: /default-rack/10.6.1.152:50010
5 12/10/10 21:30:25 INFO net.NetworkTopology: Adding a new node: /default-rack/10.6.1.151:50010
6 12/10/10 21:30:25 INFO balancer.Balancer: 0 over utilized nodes:
7 12/10/10 21:30:25 INFO balancer.Balancer: 0 under utilized nodes:
8 The cluster is balanced. Exiting...
9 Balancing took 1.006 seconds

至此，Hadoop集群的动态增加就已经完成了。
下面，我开始对Hadoop集群的节点进行动态的删除。

Hadoop集群节点的动态删除
1. 使用新增的节点
尽可能的在HDFS中产生一些测试数据，并通过Hive执行一些Job以便使新的节点也执行MapReduce并行计算。
这样做的原因是尽可能的模拟线上的环境，因为线上环境在进行删除节点之前肯定是有很多数据和Job执行过的。

2. 修改core-site.xml
dongguo@hadoop-master:~$ sudo vim /etc/hadoop/conf/core-site.xml

1 <property>
2   <name>dfs.hosts.exclude</name>
3   <value>/etc/hadoop/conf/exclude</value>
4   <description>Names a file that contains a list of hosts that are
5   not permitted to connect to the namenode.  The full pathname of the
6   file must be specified.  If the value is empty, no hosts are
7   excluded.</description>
8 </property>

3. 修改hdfs-site.xml
dongguo@hadoop-master:~$ sudo vim /etc/hadoop/conf/hdfs-site.xml

1 <property>
2   <name>dfs.hosts.exclude</name>
3   <value>/etc/hadoop/conf/exclude</value>
4   <description>Names a file that contains a list of hosts that are
5   not permitted to connect to the namenode.  The full pathname of the
6   file must be specified.  If the value is empty, no hosts are
7   excluded.</description>
8 </property>

4. 创建/etc/hadoop/conf/exclude
dongguo@hadoop-master:~$ sudo vim /etc/hadoop/conf/exclude

1 hadoop-node-3

在文件中增加需要删除的节点，一行一个，我这里仅需要写入新增的hadoop-node-3做测试。

5. 降低备份系数
在我的测试环境中，目前节点为4台，备份系数为3，如果去掉一台的话备份系数就与节点数相同了，而Hadoop是不允许的。
通常备份系数不需要太高，可以是服务器总量的1/3左右即可，Hadoop默认的数值是3。

下面，我们将备份系数从3降低到2

5.1 在所有的Hadoop服务器上更新以下配置
$ sudo vim /etc/hadoop/conf/hdfs-site.xml

01 <?xml version="1.0"?>
02 <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
03
04 <configuration>
05 <property>
06   <name>dfs.data.dir</name>
07   <value>/data/hdfs</value>
08 </property>
09 <property>
10   <name>dfs.replication</name>
11   <value>2</value>
12 </property>
13 <property>
14   <name>dfs.datanode.max.xcievers</name>
15   <value>4096</value>
16 </property>
17 </configuration>

5.2 修改hdfs文件备份系数，把/ 目录下所有文件备份系数设置为2
dongguo@hadoop-master:~$ sudo -u hdfs hadoop dfs -setrep -w 2 -R /

遇到的疑问：
在进行文件备份系数的降低时，能够很快的进行Replication set，但是在Waiting for的过程中却很长时间没有完成。
最终只能手动Ctrl+C中断，个人猜测在这个过程中HDFS正视图对数据文件进行操作，在删除一个副本容量的数据。
因此，我们应该对dfs.replication的数值做出很好的规划，尽量避免需要降低该数值的情况出现。

6. 动态刷新配置
dongguo@hadoop-master:~$ sudo -u hdfs hadoop dfsadmin -refreshNodes

7. 检查节点的处理状态
通过WEB管理界面查看
Decommissioning(退役中)

http://10.6.1.150:50070/dfsnodelist.jsp?whatNodes=DECOMMISSIONING

Dead(已经下线)

http://10.6.1.150:50070/dfsnodelist.jsp?whatNodes=DEAD

可以看到，节点已经经历了退役的过程并成功的下线了。

需要注意的是：
在删除节点时一定要停止所有Hadoop的Job，否则程序还会向要删除的节点同步数据，这样也会导致Decommission的过程一直无法完成。

8. 检查进程状态
这时我们查看进程状态，可以发现datanode进程已经被自动中止了
dongguo@hadoop-node-3:~$ sudo /etc/init.d/hadoop-0.20-datanode status
hadoop-0.20-datanode is not running.

而Tasktracker进程还在，需要我们手动中止
dongguo@hadoop-node-3:~$ sudo /etc/init.d/hadoop-0.20-tasktracker status
hadoop-0.20-tasktracker is running
dongguo@hadoop-node-3:~$ sudo /etc/init.d/hadoop-0.20-tasktracker stop
Stopping Hadoop tasktracker daemon: stopping tasktracker
hadoop-0.20-tasktracker.

此时，即使我们手动启动datanode，也是不能成功的，日志中会显示UnregisteredDatanodeException的错误。
dongguo@hadoop-node-3:~$ sudo /etc/init.d/hadoop-0.20-datanode start

1 Starting Hadoop datanode daemon: starting datanode, logging to /usr/lib/hadoop-0.20/logs/hadoop-hadoop-datanode-hadoop-node-3.out
2 ERROR. Could not start Hadoop datanode daemon

dongguo@hadoop-node-3:~$ tailf /var/log/hadoop/hadoop-hadoop-datanode-hadoop-node-3.log

01 2012-10-11 19:33:22,084 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: org.apache.hadoop.ipc.RemoteException: org.apache.hadoop.hdfs.protocol.UnregisteredDatanodeException: Data node hadoop-node-3:50010 is attempting to report storage ID DS-500645823-10.6.1.153-50010-1349941031723. Node 10.6.1.153:50010 is expected to serve this storage.
02     at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.getDatanode(FSNamesystem.java:4547)
03     at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.verifyNodeRegistration(FSNamesystem.java:4512)
04     at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.registerDatanode(FSNamesystem.java:2355)
05     at org.apache.hadoop.hdfs.server.namenode.NameNode.register(NameNode.java:932)
06     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
07     at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
08     at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
09     at java.lang.reflect.Method.invoke(Method.java:597)
10     at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:557)
11     at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1434)
12     at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1430)
13     at java.security.AccessController.doPrivileged(Native Method)
14     at javax.security.auth.Subject.doAs(Subject.java:396)
15     at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1157)
16     at org.apache.hadoop.ipc.Server$Handler.run(Server.java:1428)
17         at org.apache.hadoop.ipc.Client.call(Client.java:1107)
18     at org.apache.hadoop.ipc.RPC$Invoker.invoke(RPC.java:226)
19     at $Proxy4.register(Unknown Source)
20     at org.apache.hadoop.hdfs.server.datanode.DataNode.register(DataNode.java:717)
21     at org.apache.hadoop.hdfs.server.datanode.DataNode.runDatanodeDaemon(DataNode.java:1519)
22     at org.apache.hadoop.hdfs.server.datanode.DataNode.createDataNode(DataNode.java:1586)
23     at org.apache.hadoop.hdfs.server.datanode.DataNode.secureMain(DataNode.java:1711)
24     at org.apache.hadoop.hdfs.server.datanode.DataNode.main(DataNode.java:1728)
25 2012-10-11 19:33:22,097 INFO org.apache.hadoop.hdfs.server.datanode.DataNode: SHUTDOWN_MSG:
26 /************************************************************
27 SHUTDOWN_MSG: Shutting down DataNode at hadoop-node-3/10.6.1.153
28 ************************************************************/

至此，对Hadoop集群节点的动态删除也已经成功完成了。