Recommendations and restrictions

  • Before applying a configuration on a stack through checkpoint restore or TFTP/SFTP/USB download, make sure that current VSF-specific configurations and the intended configurations match exactly. In other words, the VSF stack and the intended configuration must have the same:

    • Total number of members

    • Member types

    • Member number/ID

    • VSF link configurations

    • Secondary member configuration

    • Split-detect configuration

  • A functional stack must be configured with a standby for redundancy purposes. If the master fails and there is no standby, the stack will fail.

  • If the master fails and there is a standby device, the standby becomes the new master and will take over stack management. When the old master device is replaced, it seamlessly becomes the standby device for the stack and there no disruption.

  • After downloading firmware to a stack, the stack must be rebooted to complete the upgrade process. Adding or rebooting individual members before the upgrade process is completed can cause the individual member to fail while joining the stack.

  • If there is a discrepancy between a VSF member link configuration on the master and the VSF member link configuration on the member, the link configuration on the member is used.

  • Configure only one link between a pair of stack members. This recommendation is also applicable for a two-member VSF stack.

  • If there is a split, failure in the connectivity between management interfaces of the master and standby might result in two active fragments. This can happen even if management split-detect is enabled.

  • Ring topology is not supported for two-member VSF.

  • Replacing member 1 in a stack without a standby with a new switch booted as member 1 will reset all configurations on the stack.

  • Do not connect a renumbered member to multiple primary devices through VSF links.

  • Before removing an individual interface from VSF link using the command no vsf link <x> <interface> , ensure that the interface is admin shutdown at both local and peer ends.

    For example: Interface 1/1/25 on member 1 link 1 is connected to 2/1/25 on member 2 link 2. The user intends to remove 1/1/25 from link 1 of member 1. Both the interfaces 1/1/25 and 2/1/25 have to be admin shutdown before actually removing them from the link configuration.

    To delete the link completely using the no link <x> command, all individual interfaces in the VSF link have to be admin shutdown both at local and peer ends.

  • Before removing an individual interface from the VSF link using the no vsf link <x> <interface> command, ensure that the interface is admin shutdown at both local and peer ends.

    Example: Interface 1/1/25 on member 1 link 1 is connected to 2/1/25 on member 2 link 2. The user intends to remove 1/1/25 from link 1 of member 1. Interfaces 1/1/25 and 2/1/25 must be admin shutdown before removing them from the link configuration.

    Before deleting the link using the no link <x> command, all individual interfaces in the VSF link must be admin shutdown both at local and peer ends.

  • There may be instances in which a master switch with vsf secondary <id> configuration is unable to discover the standby switch. In such cases, the master switch will wait for up to ten minutes to detect the standby switch.