Help Center> Data Warehouse Service> Best Practices> Tutorial: Using GDS to Import Data from a Remote Server> Step 1: Preparing an ECS as the GDS Server
View PDF

Step 1: Preparing an ECS as the GDS Server

The following parts are included in this section:

  • Preparing an ECS as the GDS Server

    Before using GDS to import and export data, prepare one or more Linux ECSs in the same VPC as the data warehouse cluster to install the GDS tool package. These ECSs serve as the GDS servers.

    • Importing data

      Before data import, upload source data files to GDS servers. Then, the GDS servers can also be called data servers.

      If source data files are uploaded to multiple servers for storage due to large data volume, you need to install, configure, and start GDS on each data server.

    • Exporting data

      Data is exported from GaussDB(DWS) as data files to GDS servers.

  • Downloading the GDS Tool Package and SSL Certificate

    Before installing GDS, log in to the GaussDB(DWS) management console and download the GDS tool package and SSL certificate that match the cluster version.

Preparing an ECS as the GDS Server

For details about how to purchase a Linux ECS, see Purchasing an ECS with Customized Configurations in the Elastic Cloud Server Getting Started. For details about how to log in to the ECS, see Logging In to a Linux ECS.

The ECS that functions as the GDS server must meet the following requirements:

  • The GDS server must use an operating system supported by the GDS tool package. For details, see Downloading the GDS Tool Package and SSL Certificate.
  • The GDS server can communicate with the data warehouse cluster.
    • The data warehouse cluster connects to the GDS server using the private network address, and they must belong to the same region, VPC, and subnet.

      The data import rate is affected by the network bandwidth. Therefore, the private network address is recommended.

    • The GDS server can properly receive network access requests from the data warehouse cluster.

      Port: Plan a listening port for GDS. This port is used by the data warehouse cluster to connect to the GDS server. You need to specify the listening port when enabling GDS. If it is not specified, the default port 8098 is used.

      Firewall: If a firewall is enabled on the GDS server, add an inbound rule for the GDS listening port to allow the data warehouse cluster to connect to the GDS server. Otherwise, the connection cannot be set up.

      Note that the listening port must be specified to the port that has been opened on the firewall when enabling GDS.

      The following example describes how to create an ECS as the GDS server.

      First of all, open the GDS listening port in the inbound rule of the security group where the ECS resides.

      Table 1 Inbound rule example

      Parameter

      Example Value

      Protocol

      TCP

      Port Range

      5000

      NOTE:

      Enter the listening port of the GDS server.

      Source

      Select IP Address and enter the IP address of the data warehouse cluster. Example value: 192.168.0.10/32.

      If the firewall is enabled on the ECS, ensure that the listening port of the GDS server is opened on the firewall.

      iptables  -I INPUT -p tcp -m tcp --dport <gds_port> -j ACCEPT

Downloading the GDS Tool Package and SSL Certificate

  1. Log in to the GaussDB(DWS) management console.
  2. In the navigation tree on the left, click Connection Management.
  3. Select the GaussDB(DWS) client of the corresponding version from the drop-down list of gsql CLI Client.

    Choose a corresponding client version according to the cluster version and operating system to which the client is to be installed.

    • The Redhat x86_64 client can be used on the following operating systems:
      • RHEL 6.4 to RHEL 7.6.
      • CentOS 6.4 to CentOS 7.4
      • EulerOS 2.3
    • The SUSE x86_64 client can be used on the following operating systems:
      • SLES 11.1 to SLES 11.4
      • SLES 12.0 to SLES 12.3
    • The Euler Kunpeng_64 client can be used on the following operating systems:
      • EulerOS 2.8
    • The Stream Euler X86_64 client can be used on the following operating system:
      • EulerOS 2.2
    • The Stream Euler Kunpeng_64 client can be used on the following operating system:
      • EulerOS 2.8
    Figure 1 Downloading a gsql client

    The CPU architecture of the client must be the same as that of the cluster. If the cluster uses x86 servers, select a x86 client.

  4. Click Download.
  5. (Optional) To start the GDS service in SSL encryption mode, click here to download the SSL certificate in the Download Client and Driver area.

    If GDS is started in SSL encryption mode, data transmitted between the GDS server and the data warehouse cluster is encrypted to ensure data security.

    The SSL mode is more secure than other modes. You are advised to start GDS in SSL mode.