> ## Documentation Index
> Fetch the complete documentation index at: https://docs.planasonix.com/llms.txt
> Use this file to discover all available pages before exploring further.

# JDBC gateway

> Connect to databases via JDBC drivers including Teradata, Hive, Druid, and more.

The **JDBC gateway** runs queries through vendor **JDBC drivers** when Planasonix does not ship a first-class connector for your engine. You upload approved driver JARs, define connection URLs, and use SQL nodes against the gateway-managed pools.

## JDBC gateway concept

Instead of embedding drivers in every worker, the gateway:

* Hosts **driver isolation** per database type and version
* Pools connections with configurable limits
* Translates Planasonix SQL requests into JDBC calls your database understands
* Surfaces errors back to pipeline logs with JDBC SQLState context

Admins operate the gateway cluster; analysts reference **connections** the admin published.

<Tip>
  Prefer first-class connectors when available—they include tuned pagination, metadata discovery, and CDC hooks the generic JDBC path may lack.
</Tip>

## Supported databases

The JDBC path supports many engines when you supply a compatible driver. Common enterprise examples include:

| Engine             | Notes                                                            |
| ------------------ | ---------------------------------------------------------------- |
| Teradata           | Use official JDBC; respect session limits on shared appliances   |
| Apache Hive        | Connect to HiveServer2; watch metastore latency on wide scans    |
| Apache Druid       | JDBC exposes SQL; complex ingest still uses native batch tooling |
| Exasol             | High-performance analytics; tune fetch size for wide rows        |
| Netezza (PureData) | Verify supported driver generation for your appliance firmware   |
| Firebolt           | Use Firebolt JDBC; align warehouse scaling with burst workloads  |

Exact support depends on driver licensing—your legal team must approve redistribution if drivers are not publicly downloadable.

## Driver upload

<Steps>
  <Step title="Obtain the JAR">
    Download from the vendor or your internal artifact repository; verify checksums.
  </Step>

  <Step title="Upload in admin console">
    Navigate to **Connectors** → **JDBC gateway** → **Drivers** and upload with a descriptive label (`teradata-17.20-jdbc4`).
  </Step>

  <Step title="Create a connection template">
    Supply JDBC URL pattern, default properties, and credential mapping.
  </Step>

  <Step title="Test connectivity">
    Run `SELECT 1` from a privileged admin account, then validate least-privilege service accounts for pipelines.
  </Step>
</Steps>

<Warning>
  Mismatched driver and database versions cause opaque protocol errors—pin versions in change management.
</Warning>

## Operational guidance

<AccordionGroup>
  <Accordion title="Network paths">
    Open firewall rules from gateway hosts to database listeners; avoid NAT ambiguity with multi-subnet clusters.
  </Accordion>

  <Accordion title="Secrets">
    Store passwords and keytabs in your vault; rotate on the same schedule as native DB users.
  </Accordion>

  <Accordion title="Performance">
    Tune fetch size and parallel readers based on DBA guidance; JDBC defaults are rarely optimal for ETL throughput.
  </Accordion>
</AccordionGroup>

## Related topics

<CardGroup cols={2}>
  <Card title="Connections overview" icon="link" href="/connections/overview">
    How credentials and networking fit together.
  </Card>

  <Card title="SQL assistance" icon="database" href="/ai-copilot/sql-assistance">
    Author JDBC-backed SQL with Copilot help.
  </Card>
</CardGroup>