System availability

To help ensure that critical data remains available to clients in a distributed automation system, even during server disruptions,

FactoryTalk Alarms and Events

supports the alarm buffering and server redundancy features of the

FactoryTalk Services Platform

and buffers alarm and event information if the connection to the controller or the database is lost.

Connection to the controller is lost

To receive device-based alarms, the alarm server (

FactoryTalk Linx

) establishes a subscription to the alarms in the

Studio 5000

controller. The controller maintains a connection to each subscriber and monitors the status of that connection.

As alarm state changes occur, the controller caches the necessary information (such as timestamps, alarm state and associated tag values) and transmits the alarm information to all of the subscribers.

If any subscriber fails to confirm receipt of the alarm information, or if the connection to a known subscriber is not good, the controller stores the undelivered alarm information in a 100 KB buffer. This buffer is created when the subscriber establishes its initial connection and is maintained for a configurable length of time (0 to 120 minutes, default is 20 minutes) after a subscriber loses its connection. Settings for the buffer behavior, such as the

Buffer Timeout

setting can be configured on each

FactoryTalk Linx

Device Shortcut

.

Each subscriber has a separate buffer so communication problems with one subscriber do not interfere with alarm delivery to other subscribers. When the buffer is full, newer alarm information is discarded and a

FactoryTalk Diagnostics

message is logged.

Connection to the database is lost

If a connection to an alarm and event history log is lost, the Alarm and Event Historian software buffers data until the connection is restored and then moves the data into the log. The Alarm and Event Historian creates its own log file on the computer where it is installed and buffers data until disk space is filled. You can limit database buffering to a specified number of days or percentage of available disk space.

If the connection to the database is lost, alarm and event information will continue to be cached to files. Once the connection is restored, the cached information will be sent to the database.

The number of days to buffer data and the maximum amount disk space to be used is configured on the

Advanced

tab of the

Alarm and Event Historian Database Properties

dialog box.

Server redundancy

When creating and configuring an alarm server, you can identify an alternate server on another computer to serve as a standby. One server in a redundant server pair is called the

active

server. The partner server waiting to provide service in the case of a failure with the active server is called the

standby

server. The

FactoryTalk

Health Monitoring service manages alarm states and identifies which server is currently active and which is standby.