![Riverstone Networks RS 8000 Getting Started Manual Download Page 131](http://html1.mh-extra.com/html/riverstone-networks/rs-8000/rs-8000_getting-started-manual_1466193131.webp)
Riverstone Networks RS 8000/8600 Switch Router Getting Started Guide 4-23
Initial Configuration
Using Redundant Control Modules
Each message confirms that the Active and Startup Configurations were saved to the backup CM.
4.9.1
Fail Over
There are two ways that the primary CM can fail – either by a software failure (which causes a
soft fail over
) or by a
hardware failure (which causes a
hard fail over
). Each of these failures cause the backup CM to assume the role of
primary CM in a different way.
Soft Fail Over
If the primary CM experiences a crash because of a software failure, it sends a command to the backup CM to take
over the role of primary CM. Because the backup CM’s configuration has been kept in sync with the primary CM’s
configuration, the switch over to becoming the primary CM occurs immediately. The new primary CM (formerly the
backup CM) maintains existing layer-2 and layer-3 flows and begins learning new flows within 5 to 20 seconds. After
the routing protocols have had a chance to converge (60 seconds), layer-3 flows are aged out gracefully and replaced
with new flows. When the original primary CM reboots or is reset, it assumes the role of the backup CM.
Notice that the default positions for the CMs within the RS 8000/8600 chassis have now changed – the primary CM
now resides in slot CM/1 and the backup CM resides in slot CM. As configuration changes are saved on the new
primary CM (in slot CM/1) they are copied over to the new backup CM (in slot CM). Again, the two Control Modules
always keep their configurations in sync.
Hard Fail Over
During normal operation, the status of the primary CM is communicated to the backup CM through a heartbeat signal.
As long as the backup CM receives heartbeats from the primary CM, the backup CM retains its standby status. If,
however, the primary CM experiences a hardware failure, the heartbeat signals cease. The backup CM waits 20
seconds for the heartbeats to resume, if they do not, the backup CM assumes the role of primary CM. As with the soft
fail over, the new primary CM maintains existing layer-2 and layer-3 flows and begins learning new flows within 5 to
20 seconds. Once the routing protocols have converged (60 seconds), layer-3 flows are aged out gracefully and
replaced with new flows.
Note
The amount of time that the backup CM waits before taking over as the primary
CM because of a hard fail over is configured using the
system set
backup-cm-timeout
command from within Configure mode. The timeout can
be set between 20 and 1000 seconds.
If the failed primary CM (in slot CM) is replaced by a new Control Module, the new CM assumes the role of backup
CM.