What and Why HbaseExplorer?
As there are projects to access HBase using common methods (check DataNucleus or Hibernate), you could use traditional RDBMS utilities like the great Squirrel to find and manipulate data in a HBase datastore. This is certainly possibly, but when the timestamp-aspect of HBase’s (or BigTables’s) data model comes into play, you have to find good workarounds. The mass of data also requires utilities that, in the RDBMS world, are vendor provided or model specific (e.g. pgadmin or todd). So many things (like to copy a database) will trigger MapReduce processes running on a connected cluster.
Unlike other database management and query tools, no query language is offered in HbaseExplorer. This might change with future versions of Hbase.
HBaseExplorer is implemented as Web Application.
So right now, HbaseExplorer shall help in doing simple data administration and monitoring. The following things can be done with the current version:
- Data Source Management: Create, Edit, Modify, Drop Hbase Cluster definition’s.
- Table Management: Table Detail Display (Families settings, Regions), Create Tables, Drop Tables, Clone a Table within the same Hbase instance, gather Table Statistics
- Querying Data: HBase Scan
- Data Display: Top-Version Display, Timestamp-oriented display
Many things could be added, the feature list contains some ideas.
The code is used regularly against fairly big datastores with different data models (e.g. one with many timestamps, one with less but big data rows). Since also Hbase is still young, this version run against the HBase 0.20.2 API, I guess new versions will require small adaptions of hbaseexplorer as well.
See Installation on how to install it.
Please use the Tracker from Sourceforge to file issues.
The HBase Pages have pointers to the standard reading. Many insights to Hbase can be found in Lars’ Blog.