Virtual Link Trunking (VLT) is a mechanism to aggregate the two connections by spreading across two upstream VLT devices. It allows connection between two different VLT switches, such as Edge, Access, or top-of-rack (ToR). As a result, the two physical switches appear as a single switch to the connected devices. VLT technology connects the devices and presents to single logical LACP end-point to connecting device/server, the device/server supports LACP sees the two VLT switch as single LACP end point. It provides multiple benefits like loop-free topology, no spanning-tree block port, load-balance traffic, port switch in link redundancy. The benefit also allows a single server to spread across two upstream switches. This allows for a nonblocking, multipathing scenario. The VLT works in Layer 2 traffic in VLT domains, the media access control (MAC) address of any host connected to the VLT peers is synchronized between the VLT nodes. 

Figure 6.   VLT physical vs logical comparison

The delay-restore feature for non-VLT ports is used to delay non-VLT ports presentation. This day provides the time for the protocols and features to converge. If there is no such mechanism available, traffic may get blocked.

The delay-restore feature is similar to the VLT delay-restore mechanism for VLT port channels. The delay-restore mechanism is applicable to the following two scenarios:

• VLT peer reload: When the reloaded node joins as a secondary node in the VLT domain or fabric, VLT port channels are brought up after delay-restore timer expires.
• VLTi link failure: If a VLTi link goes down while the VLT heartbeat is configured and functional, the secondary VLT peer brings down its local VLT port channels. All devices connected to the VLT port channel interfaces are send and receive traffic through the VLT primary device. When the VLTi link comes back up, the secondary VLT peer does not bring its VLT port channel up immediately. These channels are only brought up after VLT delay-restore timer expires, assuming that heartbeat is up.

There are two sets of configurations, one for non-VLT deployments and the other for VLT deployments.

Delay-restore port is used for non-VLT deployments. The configured ports are kept down after a reload and are brought up only after the configured delay-restore port timer expires. The VLT delay-restore timer that controls the VLT port channels also controls the orphan ports.

The delay-restore for orphan ports feature extends the support of the delay-restore timer for the VLT port channels to orphan ports. The timer corresponding to the VLT port channel also control the orphan-ports. When delay-restore orphan port is configured, the orphan port responds similarly to VLT port channels during VLT peer reload and VLTi link failure scenarios.

During reload, ethernet1/1/1 is kept down. The delay-restore timer is started and is run for 120 seconds. After the timer expires, ethernet1/1/1 is brought up.

Use the following script to view the current state of configurations and the timer:

Table 1 explains how to disable delay-restore port on an interface and revert the timer value to Default in 90 seconds.

Table 1.      Disable delay-restore port

Note: LAG does not provide the redundancy between switches, can only terminate to single switch where VLT can spread the peering to 2 different physical switches. Stacking incurs the faith-sharing, when you reload a switch, the entire switch will be affected.