juju add-machine lxc:0 fails to start lxc vm

Looks like it's also declared as bug #1271144

fwiw this is fixed in trunk

I updated jujud to 1.17.3.1, but still get the same issue

So, I'd like to get a bit more detail from you about what you expect here. I believe that on MaaS we *are* supposed to be connecting to br0, because that is bridged onto the external network, which allows the LXC container to get a routable IP address from DHCP.
While "lxcbr0" is only routable on the local machine.

If you believe there is a reason to bind to lxcbr0, then this is a bug that we need to sort out.

If you don't, then this is a dupe of bug #1271144, because we just need to be starting/configuring/whatever the br0 network that we expect to be binding to.

Hi,
 I install either precise or trusty using Maas, so I have the bridge MaaS sets for me :

juju switch maas
juju bootstrap
juju ssh 0
juju add-machine lxc:0
juju ssh 0

ifconfig -a
....
lxcbr0 Link encap:Ethernet HWaddr f2:1f:29:cb:04:cb
          inet adr:10.0.3.1 Bcast:10.0.3.255 Masque:255.255.255.0
          adr inet6: fe80::f01f:29ff:fecb:4cb/64 Scope:Lien
          UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
          Packets reçus:0 erreurs:0 :0 overruns:0 frame:0
          TX packets:185855 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 lg file transmission:0
          Octets reçus:0 (0.0 B) Octets transmis:29065271 (29.0 MB)
...
but

/var/lib/juju/containers/juju-machine-0-lxc-6/lxc.conf
contains
lxc.network.link = br0

the result is
juju status :
...
    containers:
      0/lxc/6:
        agent-state-info: '(error: error executing "lxc-start": command get_cgroup
          failed to receive response)'
        instance-id: pending
        series: trusty
...

I dunno what changed, but 1 or 2 weeks ago it was just working
maybe some recent change in maas that make br0 is not set

After reading carefully, I agree this is a dupe of bug #1271144

juju ssh 0
ifup br0
is enough to have lxc vm running again

Hi, first time to report and help on a bug. Had the same problem with a brand new trusty 14.04 installation.

On the contrary of bug #1271144, the boostrap is fine and lxcbr0 is setup properly. But juju span the machine with the old config using br0, making the container fail to boot.

I found the bug here :
vim /var/lib/lxc/juju-machine-0-lxc-0/config

The config file has a bug in its generation and the following line is wrong :

lxc.network.link = br0

Replace it for :

lxc.network.link = lxcbr0

And the machine will start properly afterward.

Confirmed on a fresh install of 14.04.1

thanks Craym!

"

I found the bug here :
vim /var/lib/lxc/juju-machine-0-lxc-0/config

The config file has a bug in its generation and the following line is wrong :

lxc.network.link = br0

Replace it for :

lxc.network.link = lxcbr0

And the machine will start properly afterward."

Changed all lxc containers to lxcbr0 and confirmed boot.

I was using on a nonlocal node @ machine 2. but same exact results

This was marked as a duplicate, but just got a report of the same issue and resolution with 1.20.11 on maas. same s/br0/lxcbr0 fix.