[pkg-lxc-devel] Bug#1010469: Bug#1010469: Bug#1010469: Bug#1010437: autopkgtest: autopkgtest-build-lxc fails to build working lxc environment

[Julian Gilbey]

On Thu, May 11, 2023 at 11:59:41PM +0200, Pierre-Elliott Bécue wrote:
> 
> Julian Gilbey <jdg at debian.org> wrote on 11/05/2023 at 16:41:46+0200:
> [...]
> > Hi Pierre-Elliott,
> >
> > I was using debian testing (whatever state it was in at the time).
> >
> > I've just tried reinstalling lxc from scratch with the current debian
> > testing.  I haven't been able to get as far as reproducing this error,
> > as I've hit a different snag:
> > [...]

> > The resulting log file contains the cryptic error messages:
> >
> > lxc-start debian-sid 20230511122856.360 ERROR network -
> > ../src/lxc/network.c:netdev_configure_server_veth:711 - No such file
> > or directory - Failed to attach "vethQ4rt4x" to bridge "lxcbr0",
> > bridge interface doesn't exist
> >
> > That's super-weird; I have no idea what "vethQ4rt4x" is meant to mean.
> 
> It's the name the hosts give randomly to the interface it creates for
> the LXC container to get network.
> 
> Inside the container it'll be eth0, outside it's a veth intervace, named
> veth$RANDOM stuff.
> 
> The issue is in the message: you configured the container to bind this
> interface on a bridge named lxcbr0 that doesn't seem to exist on the
> host.

Hi Pierre-Elliott,

Thanks so much for the quick response, that's really helpful!

Unfortunately, this doesn't seem to be the issue, though:

# systemctl status lxc-net.service 
● lxc-net.service - LXC network bridge setup
     Loaded: loaded (/lib/systemd/system/lxc-net.service; enabled; preset: enab>
     Active: active (exited) since Thu 2023-05-11 20:35:48 BST; 13h ago
       Docs: man:lxc
    Process: 81843 ExecStart=/usr/libexec/lxc/lxc-net start (code=exited, statu>
   Main PID: 81843 (code=exited, status=0/SUCCESS)
      Tasks: 1 (limit: 76868)
     Memory: 1.3M
        CPU: 70ms
     CGroup: /system.slice/lxc-net.service
             └─81884 dnsmasq --conf-file=/dev/null -u dnsmasq --strict-order -->

May 11 20:35:48 euler systemd[1]: Starting lxc-net.service - LXC network bridge>
May 11 20:35:48 euler dnsmasq[81884]: started, version 2.89 cachesize 150
May 11 20:35:48 euler dnsmasq[81884]: compile time options: IPv6 GNU-getopt DBu>
May 11 20:35:48 euler dnsmasq-dhcp[81884]: DHCP, IP range 10.0.3.2 -- 10.0.3.25>
May 11 20:35:48 euler dnsmasq-dhcp[81884]: DHCP, sockets bound exclusively to i>
May 11 20:35:48 euler dnsmasq[81884]: reading /etc/resolv.conf
May 11 20:35:48 euler dnsmasq[81884]: using nameserver 10.0.0.243#53
May 11 20:35:48 euler dnsmasq[81884]: read /etc/hosts - 7 names
May 11 20:35:48 euler systemd[1]: Finished lxc-net.service - LXC network bridge>

And with some details snipped:

# ifconfig 
enp5s0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet [...]  netmask 255.255.255.0  broadcast 192.168.0.255
        inet6 [...]  prefixlen 64  scopeid 0x20<link>
        ether [...]  txqueuelen 1000  (Ethernet)
        [...]

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        [...]

lxcbr0: flags=4099<UP,BROADCAST,MULTICAST>  mtu 1500
        inet 10.0.3.1  netmask 255.255.255.0  broadcast 10.0.3.255
        ether 00:16:3e:00:00:00  txqueuelen 1000  (Ethernet)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

tun0: [...]

wlp3s0: [...]


# bridge vlan show
port              vlan-id  
lxcbr0            1 PVID Egress Untagged



So lxc-net was established, and it still didn't work :(  (And yes,
I've just checked that lxc-start still fails.)  But maybe the bridge
is meant to be in the lxc container itself?


So I'm still totally stumped.

Any further ideas/suggestions/things to check would be welcomely
received!

Best wishes,

   Julian