Kylin 安装启动时报：java.net.UnknownHostException

1.异常

    Kylin在启动之后报一下错误：java.net.UnknownHostException: hadoop102:2181: 域名解析暂时失败。    具体错误信息如下：

2019-01-19 22:00:43,140 INFO  [main] zookeeper.ZooKeeper:438 : Initiating client connection, connectString=hadoop102:2181:2181,hadoop103:2181:2181,hadoop104:2181:2181 sessionTimeout=120000 watcher=org.apache.curator.ConnectionState@493dfb8e
2019-01-19 22:00:53,154 ERROR [main] imps.CuratorFrameworkImpl:546 : Background exception was not retry-able or retry gave up
java.net.UnknownHostException: hadoop102:2181: 域名解析暂时失败
    at java.net.Inet6AddressImpl.lookupAllHostAddr(Native Method)
    at java.net.InetAddress$2.lookupAllHostAddr(InetAddress.java:928)
    at java.net.InetAddress.getAddressesFromNameService(InetAddress.java:1323)
    at java.net.InetAddress.getAllByName0(InetAddress.java:1276)
    at java.net.InetAddress.getAllByName(InetAddress.java:1192)
    at java.net.InetAddress.getAllByName(InetAddress.java:1126)
    at org.apache.zookeeper.client.StaticHostProvider.<init>(StaticHostProvider.java:61)
    at org.apache.zookeeper.ZooKeeper.<init>(ZooKeeper.java:445)
    at org.apache.curator.utils.DefaultZookeeperFactory.newZooKeeper(DefaultZookeeperFactory.java:29)
    at org.apache.curator.framework.imps.CuratorFrameworkImpl$2.newZooKeeper(CuratorFrameworkImpl.java:146)
    at org.apache.curator.HandleHolder$1.getZooKeeper(HandleHolder.java:94)
    at org.apache.curator.HandleHolder.getZooKeeper(HandleHolder.java:55)
    at org.apache.curator.ConnectionState.reset(ConnectionState.java:218)
    at org.apache.curator.ConnectionState.start(ConnectionState.java:102)
    at org.apache.curator.CuratorZookeeperClient.start(CuratorZookeeperClient.java:189)
    at org.apache.curator.framework.imps.CuratorFrameworkImpl.start(CuratorFrameworkImpl.java:248)
    at org.apache.kylin.storage.hbase.util.ZookeeperDistributedLock$Factory.getZKClient(ZookeeperDistributedLock.java:81)
    at org.apache.kylin.storage.hbase.util.ZookeeperDistributedLock$Factory.<init>(ZookeeperDistributedLock.java:105)
    at org.apache.kylin.storage.hbase.util.ZookeeperDistributedLock$Factory.<init>(ZookeeperDistributedLock.java:101)
    at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
    at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)
    at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
    at java.lang.reflect.Constructor.newInstance(Constructor.java:423)
    at java.lang.Class.newInstance(Class.java:442)
    at org.apache.kylin.common.util.ClassUtil.newInstance(ClassUtil.java:84)
    at org.apache.kylin.common.KylinConfigBase.getDistributedLockFactory(KylinConfigBase.java:375)
    at org.apache.kylin.storage.hbase.HBaseConnection.createHTableIfNeeded(HBaseConnection.java:327)
    at org.apache.kylin.storage.hbase.HBaseResourceStore.createHTableIfNeeded(HBaseResourceStore.java:112)
    at org.apache.kylin.storage.hbase.HBaseResourceStore.<init>(HBaseResourceStore.java:93)
    at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
    at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)
    at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
    at java.lang.reflect.Constructor.newInstance(Constructor.java:423)
    at org.apache.kylin.common.persistence.ResourceStore.createResourceStore(ResourceStore.java:92)
    at org.apache.kylin.common.persistence.ResourceStore.getStore(ResourceStore.java:110)
    at org.apache.kylin.rest.service.AclTableMigrationTool.checkIfNeedMigrate(AclTableMigrationTool.java:98)
    at org.apache.kylin.tool.AclTableMigrationCLI.main(AclTableMigrationCLI.java:41)

    这是由于端口号2181在zk connectString里写了两遍，从而造成了master:2181:2181的情况。

    cd /opt/module/hbase/conf  vi hbase-site.xml,可以看到我们之前的配置时同时配置了主机名：端口号，在kylin连接zookeeper时会调用connectString，从再添加一边端口号，从而造成上述问题。

<configuration>

    <property>
        <name>hbase.rootdir</name>
        <value>hdfs://hadoop103:9000/hbase</value>
    </property>

    <property>
        <name>hbase.cluster.distributed</name>
        <value>true</value>
    </property>

   <!-- 0.98后的新变动，之前版本没有.port,默认端口为60000 -->
    <property>
        <name>hbase.master.port</name>
        <value>16000</value>
    </property>

    <property>
        <name>hbase.zookeeper.quorum</name>
         <value>hadoop102:2181,hadoop103:2181,hadoop104:2181</value>
    </property>

    <property>
         <name>hbase.zookeeper.property.dataDir</name>
         <value>/opt/module/zookeeper-3.4.10/zkData</value>
    </property>

    <property>
            <name>hbase.coprocessor.region.classes</name>
            <value>com.atguigu.coprocessor.CalleeWriteObserver</value>
    </property>

</configuration>

2.解决方法：    

    在$HBASE_HOME/conf/的文件夹下，修改hbase-site.xml的hbase.zookeeper.quorum，该项只需配置Host不需要配置端口号Port。添加如下内容

<configuration>

    <property>
        <name>hbase.rootdir</name>
        <value>hdfs://hadoop103:9000/hbase</value>
    </property>

    <property>
        <name>hbase.cluster.distributed</name>
        <value>true</value>
    </property>

   <!-- 0.98后的新变动，之前版本没有.port,默认端口为60000 -->
    <property>
        <name>hbase.master.port</name>
        <value>16000</value>
    </property>

    <property>
         <name>hbase.zookeeper.quorum</name>
         <value>hadoop102,hadoop103hadoop104</value>
    </property>

    <property>
        <name>hbase.zookeeper.property.dataDir</name>
        <value>/opt/module/zookeeper-3.4.10/zkData</value>
    </property>

    <property>
            <name>hbase.coprocessor.region.classes</name>
            <value>com.atguigu.coprocessor.CalleeWriteObserver</value>
    </property>

</configuration>

    添加完毕，重启Hbase服务即可。