INTRODUCTION
WBPEEUI250016A1
C
-
1
APPENDIX C - WARM FAILOVER FEATURE
INTRODUCTION
The warm failover feature was developed for the process con-
trol unit interface to minimize the interruption in the flow of
exception reports that occurs when the backup process control
unit interface re-establishes exception report routes once it
becomes the primary process control unit interface. The
elapsed time of this interruption is dependent on the number
of exception report routes that need to be re-established prior
to warm failover or for a cold failover. This interruption may
exceed one minute for a process control unit interface with a
large configuration.
During a warm failover, points configured in a process control
unit interface that has failed do not reflect bad quality. In con-
trast, cold failover protocol dictates that all other nodes mark
all exception report data from a process control unit interface
that has failed as bad quality until all exception report routes
are re-established and new good quality exception reports are
received. Warm failover allows other nodes to maintain good
quality briefly during a warm failover.
OPERATION
The warm failover feature requires an INNPM11 module with
firmware revision level C_0 or later and an NKMP01 or
NKMP11 cable. While the cabinet is powering up or whenever a
backup NPM module is inserted into the module mounting
unit, the two redundant NPM modules arbitrate for the pri-
mary and backup roles. This arbitration takes place over the
Controlway. When the primary and backup roles are estab-
lished, the backup module requests an image of the dipswitch
settings of the primary NPM module. This image is transferred
over the Controlway. All further redundancy communication
takes place over the NKMP01 and NKMP11 cable.
Warm failover requires the primary NPM module to transfer a
copy of its database of node, module, block, and exception
report route records to the backup NPM module. On warm
failover, the new primary NPM module broadcasts a warm
failover node restart broadcast message on the loop and imme-
diately obtains new exception reports from control modules in
its node. These exception reports are sent to all nodes that
have established exception report routes to the former primary
NPM module. Other nodes that recognize the warm failover
restart broadcast message will send the new primary NPM
module updated exception reports for all points that the former
primary NPM module requested. In this way the new primary
NPM module transfers updated exception reports out of and
