Rockwell Automation FactoryTalk Historian SE 3.0 H2H Interface User Guide - Manual (Page 81)

FactoryTalk Historian To Historian Interface User Guide 75
Chapter 9.
UniInt Failover Configuration
Introduction
To minimize data loss during a single point of failure within a system, UniInt provides two
failover schemas: (1) synchronization through the data source and (2) synchronization
through a shared file. Synchronization through the data source is Phase 1, and
synchronization through a shared file is Phase 2.
Phase 1 UniInt Failover uses the data source itself to synchronize failover operations and
provides a hot failover, no data loss solution when a single point of failure occurs. For this
option, the data source must be able to communicate with and provide data for two interfaces
simultaneously. Additionally, the failover configuration requires the interface to support
outputs.
Phase 2 UniInt Failover uses a shared file to synchronize failover operations and provides for
hot, warm, or cold failover. The Phase 2 hot failover configuration provides a no data loss
solution for a single point of failure similar to Phase 1. However, in warm and cold failover
configurations, you can expect a small period of data loss during a single point of failure
transition.
Note: This interface supports only Phase 2 failover.
You can also configure the UniInt interface level failover to send data to a High Availability
(HA) Historian Server collective. The collective provides redundant Historian Servers to
allow for the uninterrupted collection and presentation of Historian Time series data. In an
HA configuration, Historian Servers can be taken down for maintenance or repair. The HA
Historian Server collective is described in the Historian Server Reference Guide.
When configured for UniInt failover, the interface routes all FactoryTalk Historian data
through a state machine. The state machine determines whether to queue data or send it
directly to Historian depending on the current state of the interface. When the interface is in
the active state, data sent through the interface gets routed directly to Historian. In the backup
state, data from the interface gets queued for a short period. Queued data in the backup
interface ensures a no-data loss failover under normal circumstances for Phase 1 and for the
hot failover configuration of Phase 2. The same algorithm of queuing events while in backup
is used for output data.