Ubuntu 22.04 virtual machines on KVM often lose internet connectivity due to misconfigured default network bridges, incorrect MAC address assignments in virt-manager, or conflicts with systemd-networkd versus NetworkManager. The most common fix involves verifying that the VM is attached to the correct libvirt network interface, typically "default," and ensuring the guest OS recognizes the virtual ethernet adapter as a managed device rather than an unmanaged one.
Verifying the Virtual Network Configuration
The root cause of most internet disconnection issues in KVM environments is a mismatch between the virtual machine's network settings and the host's active libvirt networks. By default, KVM creates a NAT-based network named "default" which provides DHCP and DNS services to guests. If this service is stopped or disabled, all VMs relying on it will lose connectivity.
To check if the default network is active, run the following command on the host system:
virsh net-list --all
If the "default" network shows as inactive, start it immediately using:
virsh net-start default
For persistent connectivity across reboots, enable autostart for the network:
virsh net-autostart default
This step ensures that the underlying infrastructure supporting your VM's internet access remains operational regardless of host restarts. If you are using a custom bridge instead of the default NAT, ensure that the physical interface connected to the internet is correctly enslaved to the bridge device via
Ubuntu server network interfaces configuration.
Checking virt-manager Adapter Settings
Once the host network is confirmed active, the next step is to verify how the virtual machine itself is configured to connect. Open virt-manager, right-click your Ubuntu 22.04 VM, and select "Open" to view its details. Navigate to the hardware section labeled "NIC" or "Network Source."
Ensure that the model type is set to "VirtIO" for optimal performance, although "E1000" or "RTL8139" are also valid fallbacks. The critical setting here is the "Source mode" and "Device name." For standard internet access, the source should be set to "Virtual network 'default': nat". If it is set to "Bridge device" but no bridge exists, or if it points to a non-existent interface, the VM will fail to obtain an IP address.
Additionally, check the MAC address field. Sometimes, copying a VM from another host can result in duplicate MAC addresses on the same local network, causing ARP conflicts that prevent proper routing. You can regenerate a unique MAC address by clicking the refresh icon next to the MAC field in the NIC properties. This resolves many subtle connectivity issues where packets are dropped due to address collisions.
Guest OS Network Manager Conflicts
Even with perfect host and VM settings, the Ubuntu 22.04 guest may still fail to connect if its internal network management tools are conflicting. Ubuntu 22.04 uses Netplan as the primary configuration tool, which sits above either NetworkManager or systemd-networkd. A common issue arises when NetworkManager marks the virtual ethernet interface (usually eth0 or ens3) as "unmanaged."
Check the status of NetworkManager inside the VM:
nmcli general status
Look for your network interface in the list. If it says "unmanaged," NetworkManager is ignoring it, preventing automatic connection. To fix this, edit the NetworkManager configuration file:
sudo nano /etc/NetworkManager/NetworkManager.conf
Ensure there is no "unmanaged-devices" line blocking your interface. If present, comment it out or remove the specific pattern matching your virtual NIC. After saving, restart the service:
sudo systemctl restart NetworkManager
Alternatively, if you prefer using Netplan directly, ensure your YAML configuration in
/etc/netplan/ correctly identifies the interface and requests DHCP. A typical Netplan config for internet access looks like this:
network:
version: 2
ethernets:
eth0:
dhcp4: true
Apply these changes with
sudo netplan apply. If the interface name differs, use
ip a to identify the correct name before editing.
DNS Resolution and Connectivity Testing
Sometimes the VM connects to the network but cannot resolve domain names, making it appear as though there is no internet. This is often due to the guest not receiving DNS servers via DHCP from the KVM default network. Verify connectivity by pinging an external IP address:
ping 8.8.8.8
If this succeeds but
ping google.com fails, the issue is strictly DNS-related. Check
/etc/resolv.conf inside the VM. It should contain nameserver entries provided by the host. If it is empty or contains only localhost, NetworkManager may have failed to update it. Manually adding a public DNS server like Google's (
8.8.8.8) or Cloudflare's (
1.1.1.1) to the Netplan configuration can force resolution:
network:
version: 2
ethernets:
eth0:
dhcp4: true
nameservers:
addresses: [8.8.8.8, 1.1.1.1]
After applying, test browsing again. If you suspect deeper layer issues, such as ARP problems, refer to guides on
troubleshooting ARP requests outside subnet on Linux router scenarios, though this is less common in standard NAT setups.
Frequently Asked Questions
Why does my VM have an APIPA address (169.254.x.x)?
An APIPA address indicates that the VM failed to obtain an IP from the DHCP server. This usually means the libvirt default network service is down or the VM is disconnected from any network source in virt-manager. Restarting the virsh net service and reconnecting the NIC typically resolves this.
Can I use a bridged network instead of NAT?
Yes, bridged networking allows the VM to appear as a distinct device on your physical LAN, giving it a direct IP from your home router. This requires creating a bridge interface on the host and selecting "Bridge device" in the VM's NIC settings, which is useful for hosting servers but more complex to configure than NAT.
What if ping works but websites won't load?
This is almost always a DNS issue. The network path is open, but the VM doesn't know how to translate domain names to IPs. Checking
/etc/resolv.conf and ensuring it has valid nameserver entries, or manually configuring them in Netplan, will fix the problem.
How do I reset the KVM network completely?
You can destroy and recreate the default network by running
virsh net-destroy default followed by
virsh net-undefine default, then recreating it with
virsh net-define /usr/share/libvirt/networks/default.xml and starting it again. This clears any corrupted state or stale leases.
Does disabling IPv6 help with connectivity issues?
In some older KVM versions or specific host configurations, IPv6 misconfigurations can block IPv4 traffic or cause timeouts. Disabling IPv6 in the guest's Netplan configuration or kernel parameters can sometimes serve as a workaround if other troubleshooting steps fail.