VSF FAQs

What is VSF?

Virtual Switching Framework, or VSF, defines a single virtual switch comprised of multiple individual physical switches that are interconnected through standard Ethernet links. These links are referred to as VSF links.

These physical switches will function as one device with a unified control and management plane.

Multiport VSF links are supported.

What are the supported platforms for VSF?

The Aruba 6300F/M Switch Series supports VSF.

VSF can be formed with a combination of any of the Aruba 6300F/M Switch Series (JL658A, JL659A, JL660A, JL661A, JL662A, JL663A, JL664A, JL665A, JL666A, JL667A, JL668A) or a combination of any of the Aruba 6200F Switch Series (JL724A, JL725A, JL726A, JL727A, JL728A).

NOTE:

Aruba 6200F Switch Series only supports fixed SKUs.

What port speeds do VSF links support?

For Aruba 6300F/M Switch Series: All uplink ports with 10G, 25G, and 50G speeds can be configured as VSF links.

For Aruba 6200F Switch Series: All uplink ports with 10G speed can be configured as VSF links.

Aruba recommends that all VSF links be configured to run at the same speed.

Can VSF be disabled?

Users cannot disable VSF. A factory default switch boots up as a VSF-enabled device with its Member ID set to 1.

What is a primary switch in VSF stack? Is it configurable?

Only the switch with a Member ID of 1 will be the primary switch in a VSF stack. This switch will function as the stack master and will drive the control and management plane for the stack.

What is a secondary switch in a VSF stack? Is it configurable?

The secondary switch will function as the standby in a stack. There is no secondary switch by default. Any member other than Member 1 can be configured as the secondary switch using the vsf secondary-member <MEMBER-ID> command.

Aruba strongly recommends that you configure a secondary member (standby) for stack high-availability.

How many secondary member switches are configurable in a VSF stack?

A VSF stack can be configured with one secondary member only.

Once it is configured, is it possible to change the secondary member?

Yes. Remove the current secondary member using the no vsf secondary-member command. This action will trigger the member to reboot and join the stack (not a standby anymore).

A new secondary member can be configured using the vsf secondary-member <MEMBER-ID> command. The device will reboot and rejoin the stack back as standby.

The secondary member configuration can only be changed when Member 1 is master of the stack.

How are master and standby for a stack determined?

By default, the primary member (Member 1) becomes the master of the stack and the user-configured secondary member becomes the standby.

The secondary member synchronizes all its states with the master. If the current master (Member 1) fails, the standby (secondary member) will seamlessly transition to the master role. In this state, if Member 1 comes back up, it will take the standby role.

Only primary and secondary members can take up master and standby roles in a stack.

What is the role of other members in a stack?

All devices other than the master and standby are called members. These devices do not have any network, control, or management plane functions. Their interfaces are directly controlled and managed by the master switch.

Is there any restriction in the order of VSF member numbering?

There is no restriction on the order in which VSF members can be numbered. Each member, however, must have a unique number in the range of 1-10. Each member, however, must have a unique number in the range of 1-8.

What is the supported stack height and topology?

6200F platforms can stack up to 8 members with no modular SKU (only fixed SKU).

6300 F/M platforms can stack up to 10 members in a chain or ring topology.

Ring is the recommended topology. This topology requires that each member is configured with two VSF links, interconnecting each member with two other members in the stack.

On a two-member stack, configure only one VSF link that connects to its peer.

There is no concept of a ring topology in a two-member stack.

Can features be configured on a VSF link?

Once an interface becomes part of a VSF link, no standard network layer protocol or feature can run on that interface because it is part of the VSF stack fabric.

Will configurations in an individual member switch be retained after joining a stack?

Individual member device configurations are not retained after the switch is renumbered and becomes part of a stack.

How do the consoles of each member in a stack work?

The console of the master switch provides a full CLI that can be used to manage the stack. Consoles of other stack members, including the standby, have a limited set of CLI commands that are useful for troubleshooting the device from a stacking functionality standpoint.

How does an image upgrade for a stack work?

To upgrade a stack to a new firmware image, use the copy <TFTP/SFTP> image command to download the image to the device. The image will be downloaded to the stack master first and then be distributed to the other members of the stack automatically.

After downloading the firmware, reboot the stack using the boot system <PRIMARY/SECONDARY> command. This action completes the upgrade process.

Adding or rebooting individual members before the upgrade process is complete can cause the individual member to fail while joining the stack.

Can I add a member to the VSF stack when the member is running an image with a different version than the stack?

When a device joins a stack and its firmware version is different from the version on the master, the master will push its firmware copy to the device. Once the device receives a copy of the firmware, it will reboot and rejoin the stack, now running the same version as the master.

What happens when the VSF master switch goes down?

The standby switch, if present, will take the role of the master. The old master switch will boot and join the stack as the standby switch. This transition will be seamless with limited network impact.

In the absence of a standby (no secondary member configuration), master device failure causes the remaining VSF members to reboot and come back up. At this point, members will enter a state in which they are waiting for the master to come back up. During this time, front plane ports of the members will be down.

How do I recover a device that has not joined a stack due to misconfiguration?

The vsf renumber-to command is used to trigger a device to take up its new member number and light up its VSF links. This command causes the device to reboot, come back up and wait for messages from the stack master. If the VSF link is configured incorrectly or the member number is wrong, the device could be waiting in this state indefinitely.

To recover a device in this state, execute the following commands:

  1. Execute the ctrl+c command on the device console. This action launches the recovery console.

  2. Execute the vsf-factory-reset command on the recovery console.

This action resets the device to factory-default.

  • The device will come back up as member ID 1 with no VSF link configuration.

  • The device can be configured with the correct member number and VSF links.

  • The vsf renumber-to command will trigger this configuration to take effect.

The recovery console also has commands that allow the user to copy support files to an external server. This functionality is useful for troubleshooting stacking-related issues.

How do the management ports of each member in a stack work?

In a stack, only the master management interface is active. The management interface can be assigned an IP address for device management purposes. When a master device fails, the standby becomes master and activates its management interface.

How does replacing the master switch in a stack work?

The replacement device must be of the same part number as the switch being replaced. You must also have a standby switch configured for replacing the master of a stack without losing configuration.

Complete the following steps:

  1. Execute the vsf switchover command to trigger the standby switch to take over the master role.

  2. Once the stack is up with the new master, remove all physical connections from the old master switch that is being replaced.

  3. Configure VSF interfaces/links on the new device. It is of critical importance to match the interfaces configured on the switch being replaced.

  4. Physically connect the new device to the stack through configured VSF links.

  5. The new switch will join the stack and take up the role of standby.

What is the workflow for replacing a standby or member switch?

The replacement device must be of the same part number as the switch being replaced.

Complete the following steps:

  1. Configure VSF interfaces/links on the new device. It is of critical importance to match the interfaces configured on the switch being replaced.

  2. Renumber the new device to match the switch being replaced.

  3. Physically connect the new device to the stack through configured VSF links.

  4. The new switch will join the stack and take up the standby or a member role based on the secondary configuration for the stack.

What happens if a VSF link fails?

  • If the stack topology is a ring, it will degenerate to a chain when a VSF link in the stack fails.

  • If the topology is a chain, a VSF link failure will result in a stack being split into two independent stack fragments.

  • When a stack splits and the master and standby of the stack become part of two different fragments, the standby takes up the master role for its fragment. Network disruption can result because the two fragments are simultaneously active. Aruba highly recommends enabling VSF split-detection to gracefully handle split brain scenarios.

  • If a stack splits and the master and standby are in the same fragment with the other members on a different fragment, the members-only fragment will:

    • Reboot.

    • Come back up.

    • Wait for communication from the stack master.

What is VSF split-detect?

When a stack splits, the split-detect feature provides a mechanism for the fragments to discover each other.

Once the two stack fragments are discovered, the fragment that has the primary member becomes the active fragment and keeps its front plane (non-VSF) interfaces up and running. The other fragment becomes inactive and all non-VSF interfaces on the inactive fragment are brought down to avoid network disruption.

How do I configure split-detect?

VSF supports split-detection through the management interface.

Connect the management interfaces of the primary and secondary members to the same management VLAN/network or connect them directly to one another. The CLI command to enable split detection is vsf split-detect mgmt.

How do I remove the non-VSF configurations in a stack?

Use the erase startup-config command on the VSF stack. This action will remove all non-VSF related configurations from the startup-config. Then reboot the stack.

Can a VSF member be removed from a stack?

Yes, remove a member from the stack using the no vsf member <MEMBER-ID> command. All configurations associated with the member will also be removed. The member will boot and come back up with the factory default configuration.

How do I remove the master switch from the stack?

Aruba does not recommend removing a member that is master of a stack.

If the master switch has to be removed, complete a switchover and wait for:

  • the standby to take up the master role, and

  • the old master to reboot and join the stack as standby.

Then use a member remove command to remove the device from the stack.

How can I boot the whole VSF stack and individual members using CLI?

The boot system command can be used to boot the whole stack.

To boot an individual member, use the vsf member <MEMBER-ID> reboot command.

Is modifying the VSF-specific configuration using Checkpoint restore or TFTP/SFTP/USB download supported?

This functionality is not supported. Before applying a configuration on a stack through Checkpoint restore or TFTP/SFTP/USB download, you must ensure that the following configurations match exactly:

  • The current stack VSF configurations.

  • The VSF configurations that are part of the configuration file that is being restored or downloaded from the server.

Specifically, the current VSF stack and the Checkpoint/downloaded configuration that will be applied on the stack must have the same:

  • Number of members

  • Member part number (J#)

  • Member number

  • VSF link configurations

  • Secondary member configuration

  • Split-detect configuration

How can I dismantle a stack?

A VSF stack can be dismantled by using the erase all zeroize command.

This action will cause each member to reboot, come back up with factory defaults, and function as individual/standalone devices.

How do I collect support files for a stacked device?

The copy support-files all command executed on the master console will collect support and troubleshooting information from all members that are part of the stack.

If a member is not part of the stack, you must run the same command from the recovery console of the respective member.

If a stack has split into two fragments, both fragments will have a master. Execute the same command on the master console of both fragments.