【技術の悟り】Hive HBAse区別2020_1_11

4658 ワード

    
	     
		Hive Hadoop(HDFS+YARN)        ·    (YARN.MR       /  )(    )
		HBase        (HMaster+HRegionServer+ZooKeeper)      NoSQL       HDFS  
	     
		Hive    HDFS               (  Hadoop           )
		HBase                   Cell{rowKey,column,version}/KV    (k=RowKey+CF+CQ+TS+T)
	     
		Hive    SQL  
		HBase   SQL,        
	     
		Hive  YARN   MR    (            ),m   ,  
		HBase  Key-Value   ,s  ,      
	    
		Hive     RDBMS.database+table+columnqualify     
		HBase  NoSQL.Namespace+table+ColumnFamily+ColumnQualify,      [         ]
	    
		Hive     =>         SQL     HDFS       [        ,   MR  20-30s  ]
		HBase     =>    ,   ,     ,    (    X    )/   

   
	Hive       HDFS       SQL  /  MR(Driver/Map/Reduce)          
	HBase          (    X    )          

    
	       -Flume/Kafka-> Flume/    -write-> HDFS -Hive.load-> Hive -ETL/Analyze-> HBase
	       -Kafka-> SparkStreaming/    -write-> HDFS [-SparkSQL-> Hive] && [-SparkSQL-> HBase]

    
	Hive(MR    ) => Hive(Tez  ) => Spark    (       Hive,   YARN)
	Hive                 => HBase       CRUD