Hadoop集群(三) Hbase搭建

前面已经完成Zookeeper和HDFS的安装，本文会详细介绍Hbase的安装步骤。以及安装过程中遇到问题的汇总。

系列文章：

 

Hadoop集群(一) Zookeeper搭建

Hadoop集群(二) HDFS搭建

Hadoop集群(四) Hadoop升级

 

下面开始Hbase的安装。

Hbase的服务器规划

192.168.67.101 c6701 --Master + regionserver192.168.67.102 c6702 --Master(standby)+regionserver192.168.67.103 c6703 --regionserver

---在c6701上安装Hbase

1. 创建hbase用户，及创建相关目录

su - root

useradd hbase

echo "hbase:hbase" | chpasswd

mkdir -p /data/zookeeper

mkdir -p /data/hbase/tmp

mkdir -p /data/hbase/logs

chown -R hbase:hbase /data/hbase

chown -R a+rx /home/hdfs      <<<<<<<<<<<<<让hadoop目录，其他人可以用读和执行--很重要hbase需要访问

2. 解压软件

su - hbase

cd /tmp/software

tar -zxvf hbase-1.1.3.tar.gz -C /home/hbase/

3. 设置hbase-site.xml的参数

[hbase@c6701 conf]$ more hbase-site.xml <configuration>

        <property>

                <name>hbase.tmp.dir</name>

                <value>/data/hbase/tmp</value>

        </property>

        <property>

                <name>hbase.rootdir</name>

                <value>hdfs://ns/hbase</value>  <<<<<<<<<<<<<<<<<这里要注意，ns是namenode的名字，Hbase可以访问很多HDFS，在这里标注namenode，才是指定访问这个namenode，实际在hdfs看到是/hbase的目录，不会看到ns的。

        </property>

        <property>

                <name>hbase.cluster.distributed</name>

                <value>true</value>

        </property>

        <property>

                <name>hbase.master</name>

                <value>60000</value>

        </property>

        <property>

                <name>hbase.zookeeper.quorum</name>

                <value>c6701,c6702,c6703</value>

        </property>

        <property>

                <name>hbase.zookeeper.property.clientPort</name>

                <value>2181</value>

        </property>

        <property>

                <name>hbase.zookeeper.property.dataDir</name>

                <value>/data/zookeeper</value>

        </property></configuration>

4. 设置hbase-env.sh的参数

[hbase@c6701 conf]$ cat hbase-env.sh |grep -v "^#"

export JAVA_HOME=/usr/local/jdk1.8.0_144

export HBASE_CLASSPATH=$HADOOP_HOME/etc/hadoop/

export HBASE_HEAPSIZE=500M

export HBASE_OPTS="-XX:+UseConcMarkSweepGC"

export HBASE_REGIONSERVER_OPTS="-Xmx1g -Xms400m -Xmn128m -XX:+UseParNewGC -XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFraction=70"

export HBASE_MASTER_OPTS="$HBASE_MASTER_OPTS  -Xmx1g -Xms400m -XX:PermSize=128m -XX:MaxPermSize=128m"

export HBASE_REGIONSERVER_OPTS="$HBASE_REGIONSERVER_OPTS -XX:PermSize=128m -XX:MaxPermSize=128m"

export HBASE_LOG_DIR=/data/hbase/logs

export HBASE_PID_DIR=/data/hbase/hadoopPidDir

export HBASE_MANAGES_ZK=true

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HBASE_HOME/lib/:/usr/lib64/

export HBASE_LIBRARY_PATH=$HBASE_LIBRARY_PATH:$HBASE_HOME/lib/:/usr/lib64/

5. 注意一下内存设置，由于是测试环境，设置过大，导致内存不足，无法启动问题

export HBASE_REGIONSERVER_OPTS="-Xmx1g -Xms400m -Xmn128m -XX:+UseParNewGC -XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFraction=70"

export HBASE_MASTER_OPTS="$HBASE_MASTER_OPTS  -Xmx1g -Xms400m -XX:PermSize=128m -XX:MaxPermSize=128m"

内存不足的错误

[hbase@c6701 bin]$ ./hbase-daemon.sh start masterstarting master, logging to /data/hbase/logs/hbase-hbase-master-c6701.python279.org.out

Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0

Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0

Java HotSpot(TM) 64-Bit Server VM warning: INFO: os::commit_memory(0x00000006c5330000, 2060255232, 0) failed; error='Cannot allocate memory' (errno=12)## There is insufficient memory for the Java Runtime Environment to continue.# Native memory allocation (mmap) failed to map 2060255232 bytes for committing reserved memory.# An error report file with more information is saved as:# /home/hbase/hbase-1.1.3/bin/hs_err_pid7507.log

6. 增加hadoop的参数到/etc/profile中，后面hbase运行，需要知道hadoop_home

export JAVA_HOME=/usr/local/jdk1.8.0_144

export JRE_HOME=/usr/local/jdk1.8.0_144/jre

export PATH=$JAVA_HOME/bin:$PATH:/home/hbase/hbase-1.1.3/bin

export HADOOP_HOME=/home/hdfs/hadoop-2.6.0-EDH-0u2

export HADOOP_INSTALL=$HADOOP_HOME

export HADOOP_MAPRED_HOME=$HADOOP_HOME

export HADOOP_COMMON_HOME=$HADOOP_HOME

export HADOOP_HDFS_HOME=$HADOOP_HOME

export YARN_HOME=$HADOOP_HOME

export HADOOP_COMMON_LIB_NATIVE_DIR=$HADOOP_HOME/lib/native

export PATH=$PATH:$HADOOP_HOME/sbin:$HADOOP_HOME/bin

export  HADOOP_OPTS="-Djava.library.path=$HADOOP_HOME/lib:$HADOOP_COMMON_LIB_NATIVE_DIR"

---安装c6702的hbase

7. 创建hbase用户 

ssh c6702 "useradd hbase;echo "hbase:hbase" | chpasswd"

8. 为hbase用户ssh免密

ssh-copy-id  hbase@c6702

9. 拷贝软件,创建目录，解压软件

scp -r /tmp/software/hbase-1.1.3.tar.gz root@c6702:/tmp/software/.

ssh c6702 "chmod 777 /tmp/software/*;mkdir -p /data/zookeeper;mkdir -p /data/hbase/tmp;mkdir -p /data/hbase/logs;chown -R hbase:hbase /data/hbase"

ssh c6702 "chmod -R a+rx /home/hdfs"

ssh hbase@c6702 "tar -zxvf /tmp/software/hbase-1.1.3.tar.gz -C /home/hbase"

10.复制配置文件

scp -r /etc/profile root@c6702:/etc/profile

scp -r /home/hbase/hbase-1.1.3/conf/hbase-site.xml hbase@c6702:/home/hbase/hbase-1.1.3/conf/.

scp -r /home/hbase/hbase-1.1.3/conf/hbase-env.sh hbase@c6702:/home/hbase/hbase-1.1.3/conf/.

---安装c6703的hbase

8. 创建hbase用户 

ssh c6703 "useradd hbase;echo "hbase:hbase" | chpasswd"

9. 为hbase用户ssh免密

ssh-copy-id  hbase@c6703

10. 拷贝软件,创建目录，解压软件

scp -r /tmp/software/hbase-1.1.3.tar.gz root@c6703:/tmp/software/.

ssh c6703 "chmod 777 /tmp/software/*;mkdir -p /data/zookeeper;mkdir -p /data/hbase/tmp;mkdir -p /data/hbase/logs;chown -R hbase:hbase /data/hbase"

ssh c6703 "chmod -R a+rx /home/hdfs"

ssh hbase@c6703 "tar -zxvf /tmp/software/hbase-1.1.3.tar.gz -C /home/hbase"

11. 复制配置文件

scp -r /etc/profile root@c6703:/etc/profile

scp -r /home/hbase/hbase-1.1.3/conf/hbase-site.xml hbase@c6703:/home/hbase/hbase-1.1.3/conf/.

scp -r /home/hbase/hbase-1.1.3/conf/hbase-env.sh hbase@c6703:/home/hbase/hbase-1.1.3/conf/.

12. 启动Hbase master(在c6701和c6702)

hbase-daemon.sh start master

ssh -t -q c6702  sudo su -l hbase -c "/usr/local/hadoop/hbase-release/bin/hbase-daemon.sh start master"

13. 启动Hbase regionserver(在c6701和c6702和c6703)

hbase-daemon.sh start regionserver

ssh -t -q c6702  sudo su -l hbase -c "/usr/local/hadoop/hbase-release/bin/hbase-daemon.sh start regionserver"

ssh -t -q c6703  sudo su -l hbase -c "/usr/local/hadoop/hbase-release/bin/hbase-daemon.sh start regionserver"

HBASE安装过程，问题汇总

1. 首先Hbase-site.xml设置错误，ns是namenode的名字，Hbase可以访问很多HDFS，在这里标注namenode，才是指定访问这个namenode，实际在hdfs看到是/hbase的目录，不会看到ns的。

 <name>hbase.rootdir</name><value>hdfs://ns/hbase</value>  <<<<<<<<<<<<<<<<<这里要注意

2.通过hbase访问hdfs遇到问题，虽然也可以对hdfs操作，但是有warning

[root@c6701 home]# su - hbase

$ hdfs dfs -mkdir /hbase

$ hdfs dfs -ls /

Found 1 items

drwxrwx---   - hdfs hadoop          0 2017-10-25 10:18 /hbase

权限不对，需要hbase是owner才可以让hbase正常访问hdfs路径

$ hadoop fs -chown hbase:hbase /hbase

$ hdfs dfs -ls /

Found 1 items

drwxrwx---   - hbase hbase          0 2017-10-25 10:18 /hbase

已经修改成功[hbase@c6701 ~]$ hdfs dfs -ls /hbase/test17/09/27 07:45:59 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable

最后注释掉/etc/profile中的CLASSPATH=解决的这个问题

export JAVA_HOME=/usr/local/jdk1.8.0_144

export JRE_HOME=/usr/local/jdk1.8.0_144/jre#export CLASSPATH=.:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar:$JRE_HOME/lib:$CLASSPATH

export PATH=$JAVA_HOME/bin:$PATH:/home/hbase/hbase-1.1.3/bin

3. 启动Hbase的时候，还遇到权限问题，无法找到hadoop。首先通过hbase用户检查hadoop访问权限

su - hbase/home/hdfs/hadoop-2.6.0-EDH-0u2/bin/hadoop version

如果权限有问题，需要增加hbase执行hadoop文件的权限

 

4. 启动过程中，遇到错误，是由于/home/hbase/hbase-1.1.3/lib/原有的一些jar包版本与hadoop不同，进而影响，删除即可。hbase会从hadoop中读取jar包

[hbase@c6702 ~]$ hbase-daemon.sh start regionserverstarting regionserver, logging to /data/hbase/logs/hbase-hbase-regionserver-c6702.python279.org.out

Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0

Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0

SLF4J: Class path contains multiple SLF4J bindings.

SLF4J: Found binding in [jar:file:/home/hbase/hbase-1.1.3/lib/kylin-jdbc-1.5.2.jar!/org/slf4j/impl/StaticLoggerBinder.class]

SLF4J: Found binding in [jar:file:/home/hbase/hbase-1.1.3/lib/kylin-job-1.5.2.jar!/org/slf4j/impl/StaticLoggerBinder.class]

SLF4J: Found binding in [jar:file:/home/hbase/hbase-1.1.3/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]

SLF4J: Found binding in [jar:file:/home/hdfs/hadoop-2.6.0-EDH-0u2/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]

SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.

SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]

至此Zookeeper+HDFS+Hbase安装全部完成，前面还是比较顺利的，但是在Hbase安装的过程中，由于与Hadoop的衔接上，出现一些问题，耗费一些时间分析解决。

后续会继续测试，hadoop升级的过程。