zoukankan      html  css  js  c++  java
  • 修改NameNode端口后,hive表查询报错

    在进行使用hive查询表数据的时候,抛出异常


    hive> select*from blackList;
    FAILED: SemanticException Unable to determine if hdfs://node1:8020/opt/hive/warehouse is encrypted: java.lang.IllegalArgumentException: Wrong FS: hdfs://node1:8020/opt/hive/warehouse, expected: hdfs://hadoop-node1.com:8020

    这个是因为我改变了hdfs nameNode的端口导致的,是hive元数据的问题,我的hive元数据是保存在mysql上,mysql中保存的还是以前的nameNode的端口。


    我hive的在mysql中的数据库是sparkStreaming_db,
    use sparkStreaming_db;(进入到数据库)
    show tables;     (查看该数据库中所有表)
    其中有两张表
    DBS : Hive数据仓库的路径
    SDS : Hive每张表对应的路径

    mysql> select*from DBS;
    +-------+-----------------------+-------------------------------------------------+---------+------------+------------+
    | DB_ID | DESC | DB_LOCATION_URI | NAME | OWNER_NAME | OWNER_TYPE |
    +-------+-----------------------+-------------------------------------------------+---------+------------+------------+
    | 1 | Default Hive database | hdfs://hadoop-node1.com:9000/opt/hive/warehouse | default | public | ROLE |
    +-------+-----------------------+-------------------------------------------------+---------+------------+------------+
    hdfs://hadoop-node1.com:9000 是修改前的端口

    mysql> select*from SDS;
    +-------+-------+------------------------------------------+---------------+---------------------------+----------------------------------------+-------------
    | SD_ID | CD_ID | INPUT_FORMAT | IS_COMPRESSED | IS_STOREDASSUBDIRECTORIES | LOCATION | NUM_BUCKETS | OUTPUT_FORMAT | SERDE_ID |
    +-------+-------+------------------------------------------+---------------+---------------------------+----------------------------------------+------------- +------------------------------------------------------------+----------+
    | 3 | 3 | org.apache.hadoop.mapred.TextInputFormat | | | hdfs://node1:8020/opt/hive/hive_tables | -1 | org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat | 3 |
    | 4 | 4 | org.apache.hadoop.mapred.TextInputFormat | | | hdfs://node1:8020/opt/hive/hive_tables | -1 | org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat | 4 |
    +-------+-------+------------------------------------------+---------------+---------------------------+----------------------------------------+-------------

    从这里可以发现地址(hdfs://hadoop-node1.com:9000, hdfs://node1:8020)和端口是不对的,即便是hadoop-node.com和node1指向的都是同一个ip也是不行的,名称必须要一样
    端口也要一样。后面存储的地址可以不同的

    修改命令:

    update DBS t set t.DB_LOCATION_URI='hdfs://hadoop-node1.com:8020/opt/hive/warehouse' where t.DB_ID='1'
    update SDS t set t.LOCATION='hdfs://hadoop-node1.com:8020/opt/hive/tables' where t.DB_ID='3'

  • 相关阅读:
    jboss hello world
    jboss 7 启动问题
    jboss [how to access the admin console]
    tomee 消息持久化
    Python、C和Java对比
    编程语言产生时间表
    从关系型数据库到非关系型数据库
    约束和异常处理
    类与类之间的关系
    类的成员
  • 原文地址:https://www.cnblogs.com/zhangXingSheng/p/7073584.html
Copyright © 2011-2022 走看看