PVE6 slab cache grows until VMs start to crash

Indeed, this also happens on non-PVE machines. Probably combined with the generally more powerful hardware of the PVE hosts it's really only noticeable there.

On a non-PVE host, I see this:

Code:
● system-check_mk.slice
   Loaded: loaded
   Active: active since Tue 2019-10-08 19:50:50 CEST; 4 days ago
    Tasks: 0
   Memory: 2.7M
   CGroup: /system.slice/system-check_mk.slice

Which is not great, but ... boring :)

FTR: this is with kernel 4.19.67-2+deb10u1, systemd 241-7~deb10u1, absolutely no PVE packages installed.