Overview

The Client Failover service is an application that minimizes data loss by enabling adapter-level failover. This service manages multiple failover groups by assigning primary and secondary roles to adapters in each individual failover group.

Client Failover Service installation

The Client Failover service is installed with a downloadable install kit. See the Install the Client Failover Service for additional information.

This version only supports Windows Operating Systems. See the System requirements section for more details.

Client Failover Service configuration

The Client Failover service supports minimal configuration options, all of which can be configured using REST endpoints.

To add adapters to a failover group, you must update the Failover Configuration Facet section in the adapter configuration.This service acts as an endpoint that receives and sends failover messages to and from adapters. These messages report on the health of the adapter, help inform the adapter of its role within the failover group, and provides users with the ability to monitor the adapter's health.

If a primary adapter stops reporting its own health and status to the Client Failover Service, this service promotes an adapter in the secondary role to the primary role to ensure continuous data flow.

The Client Failover service supports cold, warm, and hot failover modes.

Note: Only the OPC UA (version 1.4 and newer), BACnet (version 1.2), Modbus TCP (version 1.6), DNP3 (version 1.2), and MQTT (version 1.3) adapters currently support Client Failover Service.

Options for hosting

The Client Failover service can be hosted on-premises or within CONNECT data services. There are no functional differences between the cloud and on-premises failover endpoints. Both options enable customers to choose a failover deployment that best suits their needs, environments, and use cases. In general, customers that are writing data to an on-premises PI Server should leverage the on-premises version of the Client Failover Service while customers writing to CONNECT data services should utilize the cloud version of the Client Failover Service.

Additional information on utilizing Client Failover service in an on-premises configuration can be found in this document.

For additional information on utilizing the Client Failover service within CONNECT data services, see the Client failover section of the CONNECT data services guide.

Recommended Architecture for the on-premises Client Failover Service and Adapter Pair

For a redundant adapter architecture, there are three separate components that must be considered:

• Client Failover service

• Adapter 1

• Adapter 2

Customers collecting data on-premises have a wealth of options when it comes to the installation location of the Client Failover Service and subsequent adapters. However, we recommend installing each of the components listed above on their own separate node within the same network. Separating the Client Failover Service from an adapter node increases the safety, reliability, and tolerance of the failover pair.