bhyve is eating my memory and running far beyond the configured limits

From: <elk-freebsd_at_wetznet.de>
Date: Wed, 18 Sep 2024 12:25:36 UTC
Hi,

I running Truenas 13.3 which is on Freebsd 13.3-RELEASE-p4 on AMD Ryzen 5 PRO 4650G

bhyve is configured for two VMs one with 4GB (Windows) and one with 8GB (linux)

But If I start memory intensive actions inside the linux VM inside a docker this raised the memory usage far beyond the 
8GB..
14,6GB was the last I saw yesterday and this morning the VMs were killed by Truenas cos of OOM.

The

vm is configured with this xml ...

<domain type='bhyve' xmlns:bhyve='http://libvirt.org/schemas/domain/bhyve/1.0'>
   <name>1_Jellyfin</name>
   <uuid>615af2a2-84e0-434e-b544-b4c9e7c0872d</uuid>
   <title>Jellyfin</title>
   <description>Jellyfin</description>
   <memory unit='KiB'>8388608</memory>
   <currentMemory unit='KiB'>8388608</currentMemory>
   <vcpu placement='static'>8</vcpu>
   <os>
     <type arch='x86_64'>hvm</type>
     <loader readonly='yes' type='pflash'>/usr/local/share/uefi-firmware/BHYVE_UEFI.fd</loader>
     <boot dev='hd'/>
   </os>
   <features>
     <acpi/>
     <msrs unknown='ignore'/>
   </features>
   <cpu>
     <topology sockets='1' dies='1' cores='4' threads='2'/>
   </cpu>
   <clock offset='localtime'/>
.....

But the bhyve VM Jellyfin is using 14G at this moment
Wed Sep 18 14:24:00 CEST 2024
   UID  PID PPID  C PRI NI      VSZ      RSS MWCHAN   STAT TT        TIME COMMAND
     0 2106    1 10  20  0 14774616 14149064 kqread   SC    - 1689:51.13  bhyve: 1_Jellyfin (bhyve)

What can I do to support the issue solving process ?

elk