Windows VM, network traffic crashes host network

[2oldschool]

I am new to using proxmox. I did a clean install of proxmox the other day (before 8.4). I only tweaked, notifications, updates, and added a NAS storage via SMB. Host hardware is a new Dell OptiPlex. I hadn't made a VM on it yet. When 8.4 came out I updated and everything worked fine. Today I created a windows VM on it and started using it. Everything seemed fine. After I got the VM up and running I installed ScreenConnect/ConnectWise Control on it to access it remotely. I was going through some normal customization of windows, etc then I transferred a few small size files VIA ScreenConnect and all of a sudden I losted access to the VM. I checked the host and I couldn't access that, couldn't ping it either. Looked at the physical console and it was still up. I logged into it but I am not smart with Linux so I didn't know how to check stuff but I couldn't ping from the host either. I restarted the host and went on using the VM again. Out of curiosity I made sure to disable all firewalls on DC/Host, etc. I also transfer a large file via a network share into this VM and sure enough soon into the transfer same thing happened. I looked at what I think are the logs and I found this (below). Any ideas on what I am doing wrong? or what I can do?

Apr 10 12:50:50 BIO-TRI-HOST-1 systemd[1]: Starting systemd-update-utmp-runlevel.service - Record Runlevel Change in UTMP...
Apr 10 12:50:50 BIO-TRI-HOST-1 systemd[1]: systemd-update-utmp-runlevel.service: Deactivated successfully.
Apr 10 12:50:50 BIO-TRI-HOST-1 systemd[1]: Finished systemd-update-utmp-runlevel.service - Record Runlevel Change in UTMP.
Apr 10 12:50:50 BIO-TRI-HOST-1 systemd[1]: Startup finished in 17.552s (firmware) + 6.331s (loader) + 3.663s (kernel) + 7.073s (userspace) = 34.620s.
Apr 10 12:50:50 BIO-TRI-HOST-1 systemd[1]: systemd-rfkill.service: Deactivated successfully.
Apr 10 12:50:55 BIO-TRI-HOST-1 chronyd[1102]: Selected source 23.168.136.132 (2.debian.pool.ntp.org)
Apr 10 12:50:55 BIO-TRI-HOST-1 chronyd[1102]: System clock TAI offset set to 37 seconds
Apr 10 12:50:58 BIO-TRI-HOST-1 kernel: netfs: FS-Cache loaded
Apr 10 12:50:58 BIO-TRI-HOST-1 kernel: Key type cifs.spnego registered
Apr 10 12:50:58 BIO-TRI-HOST-1 kernel: Key type cifs.idmap registered
Apr 10 12:50:58 BIO-TRI-HOST-1 kernel: CIFS: Attempting to mount //192.168.2.21/backup-group-01
Apr 10 12:51:09 BIO-TRI-HOST-1 pvedaemon[1307]: <root@pam> successful auth for user 'root@pam'
Apr 10 12:51:13 BIO-TRI-HOST-1 systemd[1]: systemd-fsckd.service: Deactivated successfully.
Apr 10 12:51:29 BIO-TRI-HOST-1 pvedaemon[1552]: start VM 150: UPID:BIO-TRI-HOST-1:00000610:0000135C:67F7F711:qmstart:150:root@pam:
Apr 10 12:51:29 BIO-TRI-HOST-1 pvedaemon[1307]: <root@pam> starting task UPID:BIO-TRI-HOST-1:00000610:0000135C:67F7F711:qmstart:150:root@pam:
Apr 10 12:51:29 BIO-TRI-HOST-1 systemd[1]: Created slice qemu.slice - Slice /qemu.
Apr 10 12:51:29 BIO-TRI-HOST-1 systemd[1]: Started 150.scope.
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: tap150i0: entered promiscuous mode
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: vmbr0: port 2(tap150i0) entered blocking state
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: vmbr0: port 2(tap150i0) entered disabled state
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: tap150i0: entered allmulticast mode
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: vmbr0: port 2(tap150i0) entered blocking state
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: vmbr0: port 2(tap150i0) entered forwarding state
Apr 10 12:51:29 BIO-TRI-HOST-1 pvedaemon[1552]: VM 150 started with PID 1574.
Apr 10 12:51:29 BIO-TRI-HOST-1 pvedaemon[1307]: <root@pam> end task UPID:BIO-TRI-HOST-1:00000610:0000135C:67F7F711:qmstart:150:root@pam: OK
Apr 10 12:51:29 BIO-TRI-HOST-1 kernel: x86/split lock detection: #AC: CPU 1/KVM/1597 took a split_lock trap at address: 0x7efcb050
Apr 10 12:51:30 BIO-TRI-HOST-1 pvedaemon[1602]: starting vnc proxy UPID:BIO-TRI-HOST-1:00000642:000013D1:67F7F712:vncproxy:150:root@pam:
Apr 10 12:51:30 BIO-TRI-HOST-1 pvedaemon[1308]: <root@pam> starting task UPID:BIO-TRI-HOST-1:00000642:000013D1:67F7F712:vncproxy:150:root@pam:
Apr 10 12:51:30 BIO-TRI-HOST-1 pveproxy[1316]: proxy detected vanished client connection
Apr 10 12:51:30 BIO-TRI-HOST-1 pvedaemon[1606]: starting vnc proxy UPID:BIO-TRI-HOST-1:00000646:000013D9:67F7F712:vncproxy:150:root@pam:
Apr 10 12:51:30 BIO-TRI-HOST-1 pvedaemon[1306]: <root@pam> starting task UPID:BIO-TRI-HOST-1:00000646:000013D9:67F7F712:vncproxy:150:root@pam:
Apr 10 12:51:33 BIO-TRI-HOST-1 pvedaemon[1306]: <root@pam> end task UPID:BIO-TRI-HOST-1:00000646:000013D9:67F7F712:vncproxy:150:root@pam: OK
Apr 10 12:51:40 BIO-TRI-HOST-1 pvedaemon[1602]: connection timed out
Apr 10 12:51:40 BIO-TRI-HOST-1 pvedaemon[1308]: <root@pam> end task UPID:BIO-TRI-HOST-1:00000642:000013D1:67F7F712:vncproxy:150:root@pam: connection timed out
Apr 10 12:52:16 BIO-TRI-HOST-1 kernel: x86/split lock detection: #AC: CPU 0/KVM/1596 took a split_lock trap at address: 0x15f351853cc
Apr 10 13:06:23 BIO-TRI-HOST-1 systemd[1]: Starting systemd-tmpfiles-clean.service - Cleanup of Temporary Directories...
Apr 10 13:06:23 BIO-TRI-HOST-1 systemd[1]: systemd-tmpfiles-clean.service: Deactivated successfully.
Apr 10 13:06:23 BIO-TRI-HOST-1 systemd[1]: Finished systemd-tmpfiles-clean.service - Cleanup of Temporary Directories.
Apr 10 13:06:23 BIO-TRI-HOST-1 systemd[1]: run-credentials-systemd\x2dtmpfiles\x2dclean.service.mount: Deactivated successfully.
Apr 10 13:10:47 BIO-TRI-HOST-1 kernel: e1000e 0000:00:1f.6 enp0s31f6: Detected Hardware Unit Hang: TDH <db> TDT <26> next_to_use <26> next_to_clean <da>buffer_info[next_to_clean]: time_stamp <1000dd555> next_to_watch <db> jiffies <1000dd9c0> next_to_watch.status <0>MAC Status <40080083>PHY Status <796d>PHY 1000BASE-T Status <3800>PHY Extended Status <3000>PCI Status <10>
Apr 10 13:10:49 BIO-TRI-HOST-1 kernel: e1000e 0000:00:1f.6 enp0s31f6: Detected Hardware Unit Hang:

 

[2oldschool]

I think I am going to try switch the network card type and see if that makes a difference. I was planning on using this as an important home lab but the fact that a VM kills network traffic to the whole host has me scared.

 

[2oldschool]

I have changed all kinds of settings and found this, setting:

ethtool -K enp0s31f6 tso off gso off gro off

I thought it was my answer. I ran multiple vms stress tested pushing network traffic and everything stayed up for 24 hours, just now, dragging a 25mb file from a smb share via windows vm it just crashed everything. I have now ordered a new network card to put in this computer.

 

[jimmycav]

Looks similar to this: https://forum.proxmox.com/threads/e1000-driver-hang.58284/

Caused by this:

https://bugzilla.proxmox.com/show_bug.cgi?id=6273

If so you can pin the previously stable kernel (worked for me at least)

 

[2oldschool]

If I get a different ethernet card, this shouldn’t be an issue, correct?
I bought https://a.co/d/dXv8yIE will test Tuesday.

 

[jimmycav]

If I get a different ethernet card, this shouldn’t be an issue, correct?
I bought https://a.co/d/dXv8yIE will test Tuesday.

Maybe - I guess it depends if it's a different chipset and driver.

 

[2oldschool]

So far so good with the new network card.