My organization ran across the same issue. In our case we went from E85.00.0007 to E86.00.0007 with dynamic upgrade package.
The same workaround worked for us. In some cases, we had to remove anti-ransomware/forensics/anti-bot blades so the folder could be deleted, then add the blades back. Unfortunately, the issue didn't come up during our smaller test groups, only once released org-wide and of course there is no downgrading using deployment rules.
Apparently CheckPoint has a driver fix that they are testing. We opted not to test it since it is more involved to disable self-protection so the driver can be updated than to follow one of the work arounds.
When I expressed our displeasure at external drives not working and systems hanging for a purely re-branding reason, the CP engineer said
"Thanks for the updates; we understand how this can cause inconvenient for the users but this was known issue which sometimes happens."
I didn't find this known issue in any release notes anywhere. I only found this thread once I had the right keyword "SandblastBackup".
So, we're not happy Check Point for the egg-on-face moment so that you can rebrand a hidden folder that no one even knows exists and would otherwise never see. Upper management losing their external disks for no valid reason whatsoever does not reflect well on your product.