Port tracking allows you to use information about the operational state of the link so that you can initiate a failure in the link that connects the edge device. Converting the indirect failure to a direct failure triggers a faster recovery process towards redundant links. When enabled, port tracking brings down the configured links based on the failed link and forces the traffic to be redirected to another redundant link.
Generally, hosts can instantly recover from a link failure on a link that is immediately (direct link) connected to a switch. However, recovering from an indirect link failure between switches in a WAN or MAN fabric with a keepalive mechanism is dependent on several factors such as the timeout values (TOVs) and on registered state change notification (RSCN) information.
the following figure, when the direct link 1 to the host fails, recovery can be immediate. However, when the ISL 2 fails between the two switches, recovery depends on TOVs, RSCNs, and other factors.
Traffic Recovery Using Port Tracking
Port tracking monitors and detects failures that cause topology changes and brings down the links that connect the attached devices. When you enable this feature and explicitly configure the linked and tracked ports, the switch software monitors the tracked ports and alters the operational state of the linked ports on detecting a link state change.
The following terms are used in this chapter:
Tracked ports—A port whose operational state is continuously monitored. The operational state of the tracked port is used to alter the operational state of one or more ports. VSAN, SAN port channel, or a Gigabit Ethernet port can be tracked. Generally, ports in E and TE port modes can also be F ports.
Linked ports—A port whose operational state is altered based on the operational state of the tracked ports. Only ports can be linked ports.
Port tracking has the following features:
The application brings the linked port down when the tracked port goes down. When the tracked port recovers from the failure and comes back up again, the linked port is also brought up automatically (unless otherwise configured).
You can forcefully continue to keep the linked port down, even though the tracked port comes back up. In this case, you must explicitly bring up the linked port when required.