adept technology GadgetGateway, Руководство пользователя

Страница: 23 / 124

2-13
GadgetGateway Overview 2.5Redundant Twin Mode
2.5.4 Definitions
For the purpose of clarifying the descriptions the following definitions are used:
Failure: A failure is detected whenever a HeartBeat times out without receiving a moni-
toring packet from both interfaces. Only the active node sends monitoring packets. The
inactive node passively listens for the monitoring packets. The inactive twin always for-
wards monitoring packets. In order for an active node to receive a monitoring packet it has
to complete a round trip, such as, out IP side to twin, in IP side of twin, out 709.x side of
twin, in 709.x side, or going the other way, out 709.x side to twin, in 709.x side of twin,
out IP side of twin, in IP side. A failure may be detected on one or both interfaces.
Fault: Once a failure is detected both twins perform a diagnostic by actively interrogating
each other on both interfaces. If the interrogation on a particular interface fails then a fault
has occurred on that interface. An alarm is generated when a fault has been determined. A
fault on a particular interface is cleared whenever a monitoring packet is received or if a
diagnostic interrogation succeeds. A cleared fault generates an alarm cleared.
Both nodes independently report failures and faults. It is possible to have a failure but not
a fault. The converse is not true. It is possible for only one twin to report a failure. For
example if either interface has failed the active node will not receive any round trip moni-
toring packets so it will report a failure on both interfaces. However it will only report a
fault on one. In the same event the inactive twin will report a failure on only one interface
not both. The inactive will report a fault on one interface.
Alternatively if one interface fails and then some time later the other interface fails, the
initially active twin will not diagnose the second fault. The initially inactive twin, how-
ever, will diagnose the second fault. Therefore in order to fully characterize the failure and
fault state of a redundant pair the state of both devices must be examined. Moreover, the
monitoring application is on the LON side. In the event of an IP failure the alarm SNVT
sent by the active node may not be received by a monitor HMI on the IP side. Although the
alarm is sent out both sides, the IP side has failed so the alarm can’t propagate on the IP
side and the Inactive twin may not have switched to forwarding mode in time to forward
the alarm packet. Nevertheless, the inactive device will also detect the fault and its alarm
will propagate.
2.5.5 Status SNVT
The twin monitoring application has a status SNVT type 93. If bound the status SNVT is
propagated either on a timer or when it is updated by the monitoring application or both or
neither. If propagate of update is off and the update time is zero then the status SNVT will
never be scheduled for propagation. In this case the only way to read the status SNVT is to
poll it. If propagate on update is off and update time is non zero then the status SNVT will
propagate at an interval specified by the update time. If propagate on update is on and
update time is non zero then the status SNVT will propagate both on the update time inter-
val and anytime the status is changed. If update time is zero and propagate on update is on