A load balancer deployment includes the following traits: an external system database, a key location, a load balancer, and the Hybrid Data Pipeline server running on one or more nodes behind the load balancer. Backing up a load balancer deployment involves backing up files or systems that correspond to each of these traits.

System database

Hybrid Data Pipeline requires a system database for storing user and configuration information. When deploying the service behind a load balancer, you must use a supported external database.

To promote the continuous availability of the service, the system database should be replicated or mirrored, and the system database should be routinely backed up. In the event that the system database fails, the backup database can be integrated with the connectivity service by performing an upgrade installation of the server on the node or nodes supporting the service. See Recovering an external system database for more information.

Note: Describing the backup procedure for the system database is beyond the scope of this document. Refer to your database vendor documentation for instructions on backing up the database.

Key location

Hybrid Data Pipeline requires the specification of a key location during installation. The installation program writes shared files used in the operation of the Hybrid Data Pipeline server to this directory. The key location must be backed up for recovery of individual nodes or recovery of the service as a whole.

Note: The KEY_LOCATION property in the <install_dir>/ddcloud/update.properties file (where <install_dir> is the installation directory of a Hybrid Data Pipeline server) contains the path of the key location.
Warning: The key location directory contains encryption keys that are used to protect sensitive data in the system database. If the keys are lost, any previously configured data sources would not be usable. Access to the backup of the key location directory should be secured using the same access controls as the key location directory itself.

The key location contains the following files that are critical for recovery. (For details, refer to Shared files and the key location in the Deployment Guide.)

  • Global properties file
  • Redistributable files
  • System database encryption keys
  • System database encrypted information
  • On-Premises Connector authentication key
  • On-Premises Connector encryption key
  • SSL keystore
  • SSL truststore
  • Third party JDBC drivers that have been integrated with Hybrid Data Pipeline
  • Installer-generated .cfg files created for each cluster node. These files can be used to formulate the haproxy.cfg file and thus the HAProxy load balancer configuration. The information in the .cfg files can also be useful for configuring other types of load balancers.

Load balancer

The Hybrid Data Pipeline product package does not include a load balancer. However, Hybrid Data Pipeline can be deployed on one or more nodes behind a load balancer to provide high availability and scalability. To support system recovery and upgrade, the following load balancer configuration and routing information should be backed up. (For details, refer to Load balancer configuration in the Deployment Guide.)

  • The SSL certificate used to provide SSL termination at the load balancer should be backed up. The backup should be secured with the same access controls as the standing certificate.
  • The settings for session affinity should be documented.
  • The host names passed in the Host header should be documented.
  • The configuration for supplying the X-Forwarded-Proto header should be documented.
  • The configuration for supplying the X-Forwarded-For header if the client IP address is needed for Hybrid Data Pipeline access logs should be documented.
  • Routing and port information for configuration of default routing and for configuration of routing to on-premises data stores.
    Note: If you are using HAProxy as your load balancer, you should backup the haproxy.cfg file as a matter of best practice. Once you spin up HAProxy on a new machine, you can quickly configure HAProxy to handle Hybrid Data Pipeline requests by adding the haproxy.cfg file to the /etc/haproxy folder and executing the command /etc/init.d/haproxy reload.

Server nodes

Individual server nodes do not themselves need to be backed up for recovery or upgrade purposes. However, you may wish to back up the installation directories for reference information such as installation log files. Best practices also recommend making a backup copy of the installation program for the particular version of Hybrid Data Pipeline that is deployed.