On 5/22/24 19:22, Arun Mani J wrote: > I have attached the screenshot of `nmcli` inside guest (clipboard doesn't work, but that's for another day may be): https://imgur.com/NlDtDtc > > The guest is stuck in two states basically - connecting and after a few seconds it reaches disconnected state. > > $ cat /proc/sys/net/ipv4/ip_forward > 1 > > (I'm using bridge instead of brctl because it is not available and seems to be deprecated in favor of bridge, please let me know if that's not the case) > > $ sudo bridge link show virbr0 > > The command gave no output, so I tried ip link (apologies if that doesn't help) So, when I'm running a VM what I get is: # bridge link show virbr0 18: vnet0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 master virbr0 state forwarding priority 32 cost 2 Something looks broken and vnet0 is not getting plugged into the bridge. > > $ ip link show virbr0 > 4: virbr0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN mode DEFAULT group default qlen 1000 > link/ether 52:54:00:78:76:0f brd ff:ff:ff:ff:ff:ff > > (It says state is DOWN ??) > Yeah, this is suspicious too. > $ for i in nat filter mangle; do sudo iptables -t $i -L -v ; done Firewall rules look more or less the same. Now, I'm assuming your guest is under qemu:///system URI, right? Can you enable debug logs, start the guest and then share the logs somewhere (e.g. pastebin-like service) please? In the logs there should be a section where libvirt creates vnetX interface and plugs it into the bridge. https://libvirt.org/kbase/debuglogs.html If you're running a monolithic daemon (libvirtd) then all the work happens inside it and logs from it are sufficient. But if you're running split daemons then I'd need to see logs from virtqemud and virtnetworkd. Michal
