Comment # 9 on bug 1075878 from Fabian Vogt

(In reply to Johannes P�hlmann from comment #8)
> I had "Apocalyptic loss of proc/tmpfs/sysfs mounts" also on openSUSE 15.1.
> By stopping open VPN or by pulling a usb network adapter. in fact every
> taking down of a network adapter.
> 
> I found a possible trigger: I had managed somehow to have network 
> manager AND wickedd active. After disabling the network manager packages and 
> switching to wicked in Yast, the Problem has vanished. 
> 
> I am fine now. The remaining root questions  are: 
> 
> * Why was i able to activate network manager AND wicked (Reasons on Level
> 1-7 ;-)

Probably by not using network.service, but also wicked.service or
NetworkManager.service. Or maybe dbus activation.

> * Why can the combination of network manager and wickedd do such damage ? 

Does /etc/NetworkManager/dispatcher.d/nfs tun0 down trigger the issue?
You might be able to run it as non-root, in which case the errors should
indicate whether it would've broken the mounts.

If so, please provide the output of:
bash -x /etc/NetworkManager/dispatcher.d/nfs tun0 down

> * Should the kernel not defend itself better against such misconfiguration ?

The kernel is most likely innocent here and doesn't care about /proc and such
at all. If userspace decides to break itself, it won't mind.

> uname -a:
> Linux xxx 4.12.14-lp151.28.48-default #1 SMP Fri Apr 17 05:38:36 UTC 2020
> (18849d1) x86_64 x86_64 x86_64 GNU/Linux