Good day @KLC_Engineer 

I am using 510 with version 5.0.1.4 and we are not experiencing this issue.

Are your edges on 5.2.0?

No, Our 510 edges are on 4.5.1, but we are currently upgrading them to 4.5.2 since I confirmed the problem is resolved there.
We were able to reproduce this issue in 5.0.1.4 by making a business policy rule for some destinations set to "Available Wired" and manually setting the active link to "Public Wireless".

We found after doing this, we could not reach any destinations from a host connected through the edge in the Segment / VLAN applicable to that business policy.

Keep in mind that testing from the edge directly via the remote diagnostics page (Ping Test) will not reproduce this issue, since the remote diag tools do not flow through the business policy logic of the control plane.

Good day @KLC_Engineer ,

Thanks for the great information.

I did not test this with the diagnostic tools, but we literally used this to block the backup traffic for a customer who is not paying for their backup link. So, I know it works because they started complaining.

That's somewhat amusing

I think the legitimate way of doing that would be to make the BP rules for that traffic "Mandatory Wired" instead, since that is how it's supposed to function.

VMware has confirmed that this issue is not fixed in 5.0.1.4, and advised it will be resolved in an upcoming release.

Dear @KLC_Engineer , Was there a LINK DEAD event when the broadband was down?

Thank you.

Hi Benjaman,

On version 4.5.1, I observed the problem both with and without a LINK DEAD condition:

A couple of our branches had a "Public Wired" LINK DEAD event and failed over to wireless. The traffic governed by the Business policy did not restore until the Wired Link restored the next day.

Also, when testing in my lab while on 4.5.1, I relabeled my wired link to "Public wireless", saved the config and all applicable traffic immediately stopped - my Voip phone de-registered and I could not reach any of the destinations in that business policy rule. 

I also restarted the edge in this condition and the problem remained.

I then relabeled the link as "Public wired" and saved the configuration, then the traffic immediately restored.
So in this scenario there was no LINK DEAD event.

This leads me to believe that an edge with only a wireless link available would have this problem permanently.

For version 5.0.1.4, I have only validated the condition by relabeling the link type between wired/wireless, not with a link down event.
I presume it would be the same scenario as on 4.5.1.

Coincidentally, I just acquired a Cradlepoint for my lab, so I could reproduce/test the link DEAD condition while on 5.0.1.4, if that is helpful.

@KLC_Engineer 
From the symptoms, I think the set up is hitting the defect which is not fixed in 5.0.1.4 or 4.5.1.

The defect ID is #97152

This is fixed in 5.1.0.x, 4.5.2 and 4.3.1

Please feel free to check the release notes of those versions.

https://docs.vmware.com/en/VMware-SASE/5.1.0/rn/vmware-sase-510-release-notes/index.html

Thank you for the info Benjaman!