Rockwell Automation FactoryTalk Historian SE 3.0 UniInt Interface User Guide - Manual (Page 100)

UniInt Failover Scheme
94
Primary Stops Updating Heartbeat Point for less than 5 Intervals
This situation might occur if the primary interface copy fails to update its heartbeat point
within four failover intervals due to the interface being ―stuck‖ somewhere in the code. In
this scenario, the primary copy does come back to life and updates its heartbeat before the
backup has assumed control. Moreover, it is assumed that the primary has missed scans for
data during the time it failed to update its heartbeat. For this reason, the backup will send the
data it had queued during the time frame where the primary stopped updating its heartbeat
point. This results in two periods of overlapping data. The first period covers the time when
the primary initially hangs and the second period covers the interval when the primary comes
alive again. The first overlap is the same as if the primary never came back to life. The
additional overlap occurs between the times when the primary came back and before the
backup noticed.
 Every failover update interval, the backup interface reads the value of the heartbeat point
for the primary interface and remembers the value internally.
 If on the third update interval the value of the primary‘s heartbeat point has not changed,
the backup transitions into a temporary PrimaryStale state. This state allows the
primary two additional failover update intervals of time to recover before the backup
assumes control. If there was no recovery time for the primary, thrashing might result
due to latency in the system architecture.
 Upon entering the PrimaryStale state, the backup stops discarding queued data.
 The backup interface remains in the PrimaryStale state for two failover update
intervals waiting for the primary interface to come back to life.
 Before two failover update intervals elapse, the primary resumes normal operation and
updates its heartbeat point.
 At the next failover update interval, the backup will see that the primary has resumed
updating its heartbeat point and will transition back to the backup role.
 When the backup exits the PrimaryStale state, it will send its queued data to the
Historian Server.
This scenario results in no loss of data even if the primary interface hangs between two and
five intervals. If the primary hangs for less than two intervals, the backup never transitions to
the
PrimaryStale state and therefore never sends any queued data. This may result in
missing scans for less than two intervals at a time.
A depiction of the timing chart for a scenario where the primary interface stops updating its
heartbeat point for less than five intervals is shown in Figure 6, below.