Comic learn hbase ---- most straightforward Hbase architecture principles to resolve HTTP: // developer.51cto.com/art/201904/595698.htm
The concept of 1.1 hbase
-
hbase based on Google's BigTable paper is to establish the hdfs on providing high reliability , high performance , column-store , scalable , real-time reading and writing distributed database system. In need real-time read and write random accessing very large scale data sets, you can use hbase.
1.2 hbase features
-
Mass Storage
-
You can store large quantities of data
-
-
Columnar storage
-
Hbase data table is stored based on the column group, the group is divided in the column direction of the column.
-
-
Easy expansion
-
Underlying rely HDFS, when insufficient disk space, only need to dynamically increase datanode service node (machine) on it
-
You can increase the storage capacity by adding server clusters
-
-
High concurrency
-
Support high concurrent read and write requests
-
-
Thin
-
Sparse mainly for flexibility Hbase column, the column family, you can specify any number of columns, column data in the case of empty, will not take up storage space.
-
-
Multiple versions of data
-
Hbase data table can have multiple versions values, default is distinguished according to the version number, the version number is the time stamp data inserted
-
-
A single data type
-
All the data is in hbase byte array stores
-
2, hbase overall architecture
-
1、Client
-
Client
-
Client contains the interface to access Hbase
-
Further Client also maintains a cache to speed access to the corresponding Hbase, such as the cache .META. Metadata information
-
-
-
2、Zookeeper
-
zookeeper cluster
-
effect
-
To achieve a high availability HMaster
-
Hbase saved metadata information is hbase table entry address all
-
HMaster realization of monitoring and HRegionServer
-
-
-
-
3、HMaster
-
Boss hbase cluster
-
effect
-
For the allocation Region HRegionServer
-
Maintenance of the entire cluster load balancing
-
Maintain metadata cluster information
-
It found that the failure of the Region, and the failure of Region assign to a normal HRegionServer
-
-
-
-
4、HRegionServer
-
hbase cluster brother
-
Region responsible for managing
-
Receiving client requests to read and write data
-
Slicing during operation becomes large region
-
-
-
5、Region
-
Minimum unit stored in the distributed cluster hbase
3, hbase table data model
-
rowkey
-
Line key
-
Primary keys of the table, the table records lexicographically sorted by the rowkey
-
-
-
Column Family
-
Column family
-
Hbase each column in the table, are vested with a column family. Column group is part of a table schema (rather than columns), the table must be defined before use.
-
-
-
Timestamp
-
Timestamp
-
Each time stamp data corresponding to the operation, can be seen as a version number version number data
-
-
-
Column
-
Row
-
The following specific column column family
-
It belongs to a ColumnFamily, similar to the concrete column we created among mysql
-
-
-
cell
-
Cell
-
Key Row from the {, column (= <Family> + <label> ), Version} uniquely determined unit
-
Data is not the type of cell, the entire memory array in bytes
-
-