Installing and Configuring GPSS in Greenplum Database

The Greenplum Streaming Server (GPSS) manages communication and data transfer between the VMware Greenplum Connector for Informatica and Greenplum Database. In addition to installing the Connector components in your Informatica cluster, you must install, configure and start GPSS in the Greenplum Database cluster.

Prerequisites

The Greenplum Streaming Server is automatically installed with Greenplum Database 5.6 or later. Install and start a compatible Greenplum Database version before you continue with the procedure.

Process

Follow these steps to install GPSS support in Greenplum Database:

Register the GPSS Extension as described in the Greenplum Streaming Server documentation.
Starting in Greenplum Database version 5.15.x, GPSS uses a new JSON configuration file format. Depending upon the version of your Greenplum Database installation, configure and start a GPSS instance as follows:
1. If you are using the Connector with Greenplum Database version 5.15.x or later, configure and start a GPSS instance as described in the Greenplum Streaming Server documentation.
2. If you are using the Connector with Greenplum Database version 5.14.x or earlier, Configure and Start GPSS as described below.

Configure and Start GPSS

If you are using the Connector to write data to Greenplum Database version 5.14.x or earlier, follow these steps to create a GPSS configuration file and start the GPSS process:

Create a JSON configuration file to configure connection information for the GPSS service. The following output shows the contents of a sample configuration file named config.json:

{
    "listenaddress": {
        "Host": "",
        "Port": 5000,
        "Encryption": {
            "CertFile": "/home/gpadmin/gpdb_bin/ext/server.crt",
            "KeyFile": "/home/gpadmin/gpdb_bin/ext/server.key",
            "CAFile": "/home/gpadmin/gpdb_bin/ext/rootCA.pem"
        }

    },
    "gpfdist": {
        "Host": "",
        "Port": 5001,
        "Encryption": {
            "CertFile": "/home/gpadmin/gpdb_bin/gpdb/server.crt",
            "KeyFile": "/home/gpadmin/gpdb_bin/gpdb/server.key",
            "CAFile": "/home/gpadmin/gpdb_bin/gpdb/rootCA.pem"
        }
    }
}

Each property of the GPSS configuration file is described in the following table.

Property	Description
`listenaddress`	Specifies the `Host`, `Port`, and optional `Encryption` settings that the GPSS service uses to listen for connection requests from the Connector. The default listen address is 0.0.0.0:5000.
`gpfdist`	Specifies the `Host`, `Port`, and optional `Encryption` settings that GPSS uses to provide the `gpfdist` service for loading data into Greenplum Database. The default value is ‘HOSTNAME:5001’, where “HOSTNAME” is the output of the `hostname` command.
`Encryption`	The `Encryption` property is optional. If you include `Encryption` as part of the `listenaddress` object, GPSS uses an HTTPS connection between itself and the Connector, and you must specify the `CertFile`, `KeyFile`, and `CAFile` properties described below. If you include `Encryption` as part of the `gpfdist` object, the GPSS service uses the encrypted `gpfdists` protocol to transfer data between itself and Greenplum Database, and `CertFile`, `KeyFile`, and `CAFile` are also needed to secure the connection. If you do not specify the `Encryption` property in one or both contexts, GPSS uses unencrypted communication (HTTP for communicating with the Connector, or unencrypted `gpfdist` protocol for communicating with Greenplum Database).
`CertFile`	Specifies the absolute path to the server certificate used for authenticating the HTTPS or `gpfdists` connection to GPSS. The CN in `CertFile` must exactly match the hostname on which GPSS runs. For HTTPS communication between the Connector and GPSS, the CN must also match the Loader host address property that the Connector specifies in its connection configuration (see Creating a New Connection to Greenplum Streaming Server).
`KeyFile`	Specifies the absolute path to the server key file used for authenticating the HTTPS or `gpfdists` connection to GPSS.
`CAFile`	Specifies the absolute path to the Certificate Authority file used for authenticating the HTTPS or `gpfdists` connection. The `CAFile` must contain the entire Certificate Authority chain. For HTTPS communication between the Connector and GPSS, the same Certificate Authority must be specified in the Certificate Authority file specified in the Connector configuration (see Creating a New Connection to Greenplum Streaming Server).

Use the gpss command to start the GPSS service, specifying the JSON configuration file to use. For example:
```
$ gpss config.json
```