Skip to main content
Start a Conversation

Solved!

Go to Solution

The_LostIT_Guy

1 Rookie

 • 

83 Posts

137

February 13th, 2025 14:59

Best practice for performing maintenance on management switch that VLT domain uses for backup-links

My understanding is that if the VLT peers lose the ability to communicate via their backup links, a split-brain scenario can occur. That's obviously not good, so I am wondering what the recommended method of performing maintenance on the management switch attached to a VLT domain would be. Here are my current thoughts:

  •  Establish an alternative backup-link between the peers by configuring in-band management ports (front facing) and directly connect the two peers.
    • Can more than one backup-link be configured at a time and will the peers automatically attempt to send heart beats over the other if the first fails?
    • If no to the above, is it feasible to temporarily remove the first back up link and then configure an alternative one, or does this have the potential to cause a split-brain scenario since the peers won't be able to send heart beats for a very brief moment?
  • In lieu of alternative backup-links, identify the secondary VLT peer and shut it down while performing maintenance on the management switch. 

DELL-Charles R

Moderator

 • 

4.5K Posts

February 13th, 2025 19:41

Hello,

 

As we understand, under each switch vlt-domain, you currently have “backup destination x.x.x.x” pointing to the vlt peer management IP address. You have a connection from the management interface OOB of each switch going to a management switch, on which you want to perform maintenance. This “heartbeat” connection is used for peer liveliness check.

In addition, you have the “VLTi links” that connect the 2 VLT switches for data traffic to sync up mac and arp tables.

 

Here are some rules (not all) of how “heartbeat” and “VLTi links” are used:

  • If you only unplug the “heartbeat” connection that goes through management, but you leave the VLTi functioning, there will not be service impact.
  • If only the VLTi link fails, but the peer was found alive via the “heartbeat” check, the secondary VLT peer shuts down its VLT ports to prevent split brain scenario. 
  • If the primary VLT node fails (the whole node fails), then both the VLTi and “heartbeat” fail, then the current secondary peer takes over the primary role.

 

So in your case, you can simply connect directly the OOB management ports of the 2 VLT peer switches (unplugging the links and connecting direct will not have service impact). In fact, we have a lot of users who have VLT heartbeat setup this way permanently.

 

You can refer to this VLT Guide and search for the word “heartbeat” to read more.

https://dell.to/4k1M4Tx

The_LostIT_Guy

1 Rookie

 • 

83 Posts

February 14th, 2025 13:25

@DELL-Charles R​ Thank you for such a thorough answer. It really helped fill the gaps in my knowledge/understanding! 

Was this post helpful?

View All

No Events found!

Top