Haoop 3.1.2 HA High Availability Buildings

Posted by ONiX on Thu, 05 Sep 2019 09:08:19 +0200

Official Documents

https://hadoop.apache.org/docs/r3.1.1/hadoop-project-dist/hadoop-hdfs/HDFSHighAvailabilityWithQJM.html#Configuration_details

1. Resource Role Planning

2. Configuration

2.1 Backup of fully distributed cluster. If you want to use fully distributed cluster in the future, change the name of hadoop-full to hadoop, and the program will read it.

cd /var/opt/hadoop-3.1.2/etc
cp -r  hadoop hadoop-full

2.2 Modify the hadoop-env.sh file

Enter the etc in the hadoop directory (add two role configurations)

vi  + hadoop-env.sh

Comment out the line SECONDARYNAMENODE
Add the following two lines

export HDFS_ZKFC_USER=root
export HDFS_JOURNALNODE_USER=root

2.3 Modify the core-site.xml file

vi + core-site.xml

<property>
        <name>fs.defaultFS</name>        
        <value>hdfs://mycluster</value>      
    </property>
<property>
        <name>hadoop.tmp.dir</name>
        <value>/var/cjc/hadoop/ha</value>     #Set up the storage directory of hadoop files
<description>A base for other temporary directories.</description>   #Without this, NameNode won't start.
    </property>                                                      #The pit I trampled on
<property>
   <name>hadoop.http.staticuser.user</name>
   <value>root</value>
   </property>
<property>
   <name>ha.zookeeper.quorum</name>
   <value>slave2:2181,slave3:2181,slave4:2181</value>
</property>
<property>
  <name>ipc.client.connect.max.retries</name>
  <value>100</value>
  <description>Indicates the number of retries a client will make to establish
      a server connection.
  </description>
</property>
<property>
  <name>ipc.client.connect.retry.interval</name>
  <value>10000</value>
</property>

2.4 Modify the hdfs-site.xml file

vi + hdfs-site.xml

property>
        <name>dfs.replication</name>
        <value>2</value>
    </property>
<property>
       <name>dfs.nameservices</name>    #set name
       <value>mycluster</value>
    </property>
<property>
  <name>dfs.ha.namenodes.mycluster</name>   #Set the number of NameNode s, two of which are set here.
  <value>nn1,nn2</value>
</property>
<property>
  <name>dfs.namenode.rpc-address.mycluster.nn1</name>  #Call of remote service 
  <value>master:8020</value>
</property>
<property>
  <name>dfs.namenode.rpc-address.mycluster.nn2</name>
  <value>slave2:8020</value>
</property>
<property>
  <name>dfs.namenode.http-address.mycluster.nn1</name>  #Browsers access the IP and address of namenode configuration 
  <value>master:50070</value>
</property>
<property>
  <name>dfs.namenode.http-address.mycluster.nn2</name> #Two
  <value>slave2:50070</value>
</property>
<property>
  <name>dfs.namenode.shared.edits.dir</name>
  <value>qjournal://Master: 8485; slave 2: 8485; slave 3: 8485 / mycluster < / value > # journalnode is set to three nodes
</property>
<property>
  <name>dfs.client.failover.proxy.provider.mycluster</name>
  <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>  #The hdfs client finds the proxy class of NameNode, activity class name
</property>
    <property>
      <name>dfs.ha.fencing.methods</name>
      <value>sshfence</value>
    </property>
    <property>
      <name>dfs.ha.fencing.ssh.private-key-files</name>
      <value>/root/.ssh/id_rsa</value>
    </property>
<property>
  <name>dfs.journalnode.edits.dir</name>
  <value>/var/cjc/hadoop/ha/journalnodeDiary</value>
</property>
<property>
# Configuration of automatic failover
   <name>dfs.ha.automatic-failover.enabled</name>
   <value>true</value>
</property>

Save out
Run this command in the hadoop directory of etc in hadoop 3.1.2

scp core-site.xml hdfs-site.xml hadoop-env.sh slave2:pwd
Change 2 to 3,4 and do it twice.
pwd needs to be wrapped in `, and CSDN editors wrapped in `will become blocks of code
Distribution of configuration files to nodes 2, 3, 4 without knocking one by one

3. Install zookeeper

3.1 Unzip and configure zookeeper

According to the resource allocation, zookeeper is installed in node 2, 3, 4.

tar xf zookeeper-3.4.6.tar.gz -C /var/opt/  #decompression

vi + /etc/profile

Add these two lines

export ZOOKEEPER_HOME=/var/opt/zookeeper-3.4.6
PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$ZOOKEEPER_HOME/bin

Scp/etc/profile slave3:/etc/# distributed to slave3 nodes

Scp/etc/profile slave4:/etc/# is distributed to slave4 nodes

source /etc/profile  #Make the configuration file effective

Enter the conf directory of zookeeper

cp zoo_sample.cfg zoo.cfg   #Rename

vi zoo.cfg  #Edit zoo.cfg file, mainly change data directory and configure zookeeper startup node

dataDir=/var/opt/zk  #zk data storage directory to be changed, directory customization

# Add the following three lines at the bottom
server.1=slave2:2888:3888
server.2=slave3:2888:3888
server.3=slave4:2888:3888

Make sure you save and quit

You can't run zookeeper after you have the configuration file. You need to generate an ID file.

mkdir -p /var/opt/zk #02 03 04 nodes need to be created



[root@slave02 ~]# echo 1 > /var/opt/zk/myid
[root@slave03 ~]# echo 2 > /var/opt/zk/myid
[root@slave04 ~]# echo 3 > /var/opt/zk/myid

1, 2, 3 correspond to the above configuration file.

When everything is ready, hit zkS on the command line and TAB will automatically complete it.

zkServer.sh start #Start zookeeper, running at 2.3 and 4 nodes

jps to see if there is a zookeeper process

**~~

4. Start journalnode

~~ **

hdfs --daemon start journalnode  #Run on three configured journalnode nodes

Format namenode on master node

hdfs namenode -format

# Start namenode on master node

 hadoop-daemon.sh start namenode

jps view process

5. Synchronization

**
Run this command at Node 2

hdfs namenode -bootstrapStandby #Synchronization command

cd /var/cjc/hadoop/ha/name/current
cat VERSION

You can see that the cluster ID is the same
Prove synchronization success

There was a time when the namenode of my node 2 could not be started. I looked through the log and found that a file of fsimage was missing.
Solution: https://www.cnblogs.com/zy1234567/p/10643895.html

hdfs zkfc -formatZK    #Format zk


start-dfs.sh    #Start the cluster

See these proven startup success

Browser Running Address

HA High Availability Build Successfully!

Topics: Hadoop Zookeeper xml Apache

Programmer Think