Download HBase JDBC Driver

Download the HBase JDBC driver, copy the JDBC URL into DbSchema, and start designing schemas with ER diagrams — free Community Edition.

About the HBase JDBC driver

A JDBC driver is a Java .jar library that lets applications connect to HBase. DbSchema uses JDBC to reverse engineer schemas, run SQL, and generate documentation. Driver files are usually published by the database vendor or an open-source project.

JDBC URL format

Host, port, database name, and SSL options are passed in a URL starting with jdbc:. Each HBase driver uses its own syntax — see the connection details below.

The HBase JDBC driver

Apache HBase is a distributed, column-family NoSQL database modeled after Google Bigtable, running on top of HDFS in a Hadoop cluster. It provides random read/write access to billions of rows and petabytes of data, making it ideal for sparse, wide-column data like user activity logs and event streams.

DbSchema can connect to HBase using the Phoenix JDBC driver. This driver requires Phoenix to be deployed on HBase cluster. Please check this link for details.

Download HBase JDBC Driver

The driver archive is a zip file. Extract it and load the .jar files using DbSchema's Driver Manager.

Connecting to HBase through Phoenix

HBase has no SQL interface of its own. Apache Phoenix provides it, and Phoenix is not a client-side driver you can simply download — part of it runs inside the HBase cluster.

1. Install Phoenix on the HBase servers. The Phoenix server jar has to be on the region servers' classpath, and the cluster restarted. Without it the driver connects and then fails on the system catalog Phoenix expects to find.
2. Match the client jar to the server version. Phoenix client and server versions must agree, and the pairing is tied to the HBase major version. A mismatched client is the single most common failure here, and the error it produces rarely mentions versions.
3. Point the URL at ZooKeeper, not at HBase. The thick Phoenix client finds the cluster through ZooKeeper, so the host and port are the ZooKeeper quorum — port 2181, not an HBase port. This is why DbSchema prefills 2181 here.
4. Open port 2181 in the server's firewall. The database listens on port 2181 unless it was installed with a different one. A host firewall on the server blocks incoming connections on that port until a rule allows them. The firewall guide has the commands for Windows, ufw, firewalld and iptables.
5. Connect from DbSchema. In the Connection Dialog choose Remote computer or custom port, fill in the server host and the port, then press Test Connection to check that the server answers before you connect.

The client also needs to reach the region servers that ZooKeeper points it at, so opening the ZooKeeper port alone is not always enough on a firewalled cluster.

DbSchema and HBase

DbSchema connects to HBase via the Phoenix JDBC driver (which adds SQL semantics on top of HBase) and renders Phoenix tables and secondary indexes as a visual schema. Use the SQL Editor to run standard SQL queries against HBase data through the Phoenix layer.