Containers loose network until Reboot of the LXC

[sgtpepperaut]

Hello Proxmox Community,

I am experiencing a recurring issue with various LXC containers on my Proxmox setup losing network connectivity. Here are the specifics:

• Setup Details: All LXC containers are on the same network as the host and the router. 192.168.1.0/24
• Network Assignments: Some containers use DHCP, while others have static IPs assigned through the Proxmox GUI.
• Current Issue: The containers randomly lose network connectivity. Restarting the networking service on both the host and the containers (systemctl restart networking) does not resolve the issue. Only rebooting the containers seems to temporarily fix it. (i can not ping any other machine in the network and the other machines cant ping the container)
• Troubleshooting Done:
  • Initially suspected Cloudflare or Tailscale installations were causing interference, but after removing these services, the problem persists.

I am looking for advice on further troubleshooting steps I can take to diagnose and ideally resolve this issue. Any insights into log files to check, specific Proxmox configurations to revisit, or similar experiences would be greatly appreciated.

Thank you in advance for your help!

EDIT. i just found out that it affects ALL virtual interfaces including VMS.

 

[sgtpepperaut]

log from dmesg that migh be relevant:
[64909.665916] eth0: left promiscuous mode
[65198.281574] CIFS: Attempting to mount //192.168.1.5/plex-media
[65198.382190] eth0: entered promiscuous mode
[65204.420203] CIFS: VFS: Error connecting to socket. Aborting operation.
[65204.420212] CIFS: VFS: cifs_mount failed w/return code = -113
[65209.453397] eth0: left promiscuous mode
[65233.110536] ata7.00: Enabling discard_zeroes_data
[65472.618634] vmbr0: port 2(veth1018i0) entered blocking state
[65472.618641] vmbr0: port 2(veth1018i0) entered disabled state
[65472.618655] veth1018i0: entered allmulticast mode
[65472.618695] veth1018i0: entered promiscuous mode
[65472.687860] vmbr0: port 2(veth1018i0) entered blocking state
[65472.687867] vmbr0: port 2(veth1018i0) entered forwarding state
[65498.443235] CIFS: Attempting to mount //192.168.1.5/plex-media
[65498.967941] eth0: entered promiscuous mode
[65504.583730] CIFS: VFS: Error connecting to socket. Aborting operation.
[65504.583737] CIFS: VFS: cifs_mount failed w/return code = -113
[65510.037923] eth0: left promiscuous mode

 

[aaron]

Do you have the same IP address in use multiple times? That usually results in quite interesting network connectivity issues as well.

 

[sgtpepperaut]

Do you have the same IP address in use multiple times? That usually results in quite interesting network connectivity issues as well.

i do not think so. also would two machines on the same network (that have the same IP) cause two others who are correctly configured prevent communication between the correctly configured ones? i do not think so but im happy to learn that im wrong...

 

[sgtpepperaut]

i also found out that when i disconnect and reconnect the virtual network interface from the proxmox gui it starts working again.


I would really appreciate any help because im totally stuck

 

[B.Otto]

i do not think so. also would two machines on the same network (that have the same IP) cause two others who are correctly configured prevent communication between the correctly configured ones? i do not think so but im happy to learn that im wrong...

It would, since it screws up the routing tables on the virtual network between them. Then weird things happen, like machine A sends an IP packet but the reply packet gets sent to machine B, and such

 

[linkstat]

Hiello.
Could you tell us something about the hardware involved?
Is it a single node, a cluster?
Is it a server board or a workstation running Proxmox?
And network cards? (I've had some issues with consumer grade Realtek network cards in the past).
Another question, is: your network configuration the one Proxmox comes with by default? or are you using something else?

Regards!

 

[Fazio8]

Hello, I'm facing the same issues using vmbr and LXC.
If I restart the networking service on PVE my LXC loose connectivity (no pkts flowing according to tcpdump), until I restart the LXC.
Did you find any solution?
Thanks!