Guides → Set up Data Agent with Proxy

Starting 2024.7, the data agent uses two separate network protocols to communicate with the Incorta cluster: one for the controller and one for the agent. When the data agent host must reach the cluster through a proxy server, you must configure each component with the appropriate proxy settings before starting the data agent.

Setup steps

Here are the high-level steps to configure a data agent through a proxy:

1. Confirm proxy protocol support.
2. Install the data agent.
3. Add HTTP CONNECT proxy settings to the controller component.
4. Add SOCKS5 proxy settings to the agent component.
5. Start the data agent and verify the connection.

Confirm proxy protocol support

Before you begin, confirm the following:

• Your proxy supports both of the following protocols, with access to the Incorta cluster's data agent service ports:

• Ensure that the proxy server and any device on the network path forward the Transport Layer Security (TLS) traffic without modification. TLS interception (also known as SSL/TLS inspection or SSL bumping) breaks the data agent connection.

Install the data agent

Follow the standard data agent installation steps. Do not start the data agent yet.

Configure the controller proxy

The controller uses HTTP CONNECT to communicate with the cluster on port 6260. Configure the JVM proxy settings in the controller startup script.

1. On Linux or macOS, open incorta.dataagent.controller/bin/controller.sh. On Windows, open controller.bat.

2. Locate the section that sets the JAVA_OPTS based on the JDK version, and add the following immediately below it:

   • Linux / macOS:

     Copy
     JAVA_OPTS="$JAVA_OPTS -Dhttps.proxyHost=<PROXY_HOST> -Dhttps.proxyPort=<PROXY_PORT>"
     export JAVA_OPTS
     
     

   • Windows:

     Copy
     set JAVA_OPTS=%JAVA_OPTS% -Dhttps.proxyHost=<PROXY_HOST> -Dhttps.proxyPort=<PROXY_PORT>
     
     

   Note

   Replace <PROXY_HOST> and <PROXY_PORT> with the HTTP CONNECT proxy address and port.

3. If the proxy requires authentication, append the proxy credentials to the previous Java option. For example:

   Copy
   JAVA_OPTS="$JAVA_OPTS -Dhttps.proxyHost=<PROXY_HOST> -Dhttps.proxyPort=<PROXY_PORT> -Dhttps.proxyUser=<PROXY_USER> -Dhttps.proxyPassword=<PROXY_PASSWORD>"
   
   

4. Save the file.

Configure the agent proxy

The agent component uses SOCKS5 to communicate with the cluster on ports 6767 and 6768. Configure the JVM proxy settings in the agent startup script.

1. On Linux or macOS, open incorta.dataagent/bin/agent.sh. On Windows, open agent.bat.

2. Locate the section that sets the JAVA_OPTS based on the JDK version, and add the following immediately below it:

   • Linux / macOS:

     Copy
     JAVA_OPTS="$JAVA_OPTS -DsocksProxyHost=<PROXY_HOST> -DsocksProxyPort=<PROXY_PORT>"
     export JAVA_OPTS
     
     

   • Windows:

     Copy
     set JAVA_OPTS=%JAVA_OPTS% -DsocksProxyHost=<PROXY_HOST> -DsocksProxyPort=<PROXY_PORT>
     
     

   Note

   Replace <PROXY_HOST> and <PROXY_PORT> with the SOCKS5 proxy address and port.

3. Save the file.

Start the data agent

In releases starting 2024.7, start the controller first, then start the agent.

• To start the controller, run the following. Once started, use the Incorta platform to start the agent:

  Copy
  cd <DATA_AGENT_INSTALL_DIR>/incorta.dataagent.controller
  ./bin/controller.sh start
  
  

• To start the data agent for releases before 2024.7, run the following:

  Copy
  cd ../incorta.dataagent
  ./bin/agent.sh start
  
  

For more details, refer to Tools → Data Agent → Start the data agent service.

Verify the connection

In the Cluster Management Console (CMC), confirm that the data agent displays Connected status.

Alternative setup: direct outbound access

If the available proxy does not support the required protocols, you can configure a direct egress allow-list for the data agent host to reach the cluster on the following destinations. The host values for each destination are visible in the CMC.

Confirm that the allow-list is active, and verify connectivity from the data agent host before starting it:

• Linux/macOS:

  Copy
  nc -vz <CLUSTER_HOST> <PORT>
  
  

• Windows:

  Copy
  Test-NetConnection -ComputerName <CLUSTER_HOST> -Port <PORT>
  
  

A successful TCP connection to each destination confirms that the allow-list is active. Start the data agent.