Juju openstack lvd returns always "Container is already running a start operation"

[Ronny Lo Basso]

When I tried to set up Openstack development environment, the same and always error is returned :

container=juju-db2944-18 err="Container is already running a start operation" lvl=eror msg="stop hook failed" t=2017-05-23T18:11:05+0200
action=start created=2017-05-23T16:11:02+0000 ephemeral=false lvl=eror msg="Failed starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:11:10+0200 used=1970-01-01T00:00:00+0000
created=2017-05-23T16:11:02+0000 ephemeral=false lvl=info msg="Deleting container" name=juju-db2944-18 t=2017-05-23T18:11:11+0200 used=2017-05-23T16:11:05+0000
created=2017-05-23T16:11:02+0000 ephemeral=false lvl=info msg="Deleted container" name=juju-db2944-18 t=2017-05-23T18:11:11+0200 used=2017-05-23T16:11:05+0000
ephemeral=false lvl=info msg="Creating container" name=juju-db2944-18 t=2017-05-23T18:11:21+0200
ephemeral=false lvl=info msg="Created container" name=juju-db2944-18 t=2017-05-23T18:11:21+0200
action=start created=2017-05-23T16:11:21+0000 ephemeral=false lvl=info msg="Starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:11:24+0200 used=1970-01-01T00:00:00+0000
container=juju-db2944-18 err="Container is already running a start operation" lvl=eror msg="stop hook failed" t=2017-05-23T18:11:24+0200
action=start created=2017-05-23T16:11:21+0000 ephemeral=false lvl=eror msg="Failed starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:11:29+0200 used=1970-01-01T00:00:00+0000
created=2017-05-23T16:11:21+0000 ephemeral=false lvl=info msg="Deleting container" name=juju-db2944-18 t=2017-05-23T18:11:30+0200 used=2017-05-23T16:11:24+0000
created=2017-05-23T16:11:21+0000 ephemeral=false lvl=info msg="Deleted container" name=juju-db2944-18 t=2017-05-23T18:11:30+0200 used=2017-05-23T16:11:24+0000
ephemeral=false lvl=info msg="Creating container" name=juju-db2944-18 t=2017-05-23T18:11:40+0200
ephemeral=false lvl=info msg="Created container" name=juju-db2944-18 t=2017-05-23T18:11:40+0200
action=start created=2017-05-23T16:11:40+0000 ephemeral=false lvl=info msg="Starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:11:43+0200 used=1970-01-01T00:00:00+0000
container=juju-db2944-18 err="Container is already running a start operation" lvl=eror msg="stop hook failed" t=2017-05-23T18:11:43+0200
action=start created=2017-05-23T16:11:40+0000 ephemeral=false lvl=eror msg="Failed starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:11:48+0200 used=1970-01-01T00:00:00+0000
created=2017-05-23T16:11:40+0000 ephemeral=false lvl=info msg="Deleting container" name=juju-db2944-18 t=2017-05-23T18:11:48+0200 used=2017-05-23T16:11:43+0000
created=2017-05-23T16:11:40+0000 ephemeral=false lvl=info msg="Deleted container" name=juju-db2944-18 t=2017-05-23T18:11:49+0200 used=2017-05-23T16:11:43+0000
ephemeral=false lvl=info msg="Creating container" name=juju-db2944-18 t=2017-05-23T18:11:59+0200
ephemeral=false lvl=info msg="Created container" name=juju-db2944-18 t=2017-05-23T18:11:59+0200
action=start created=2017-05-23T16:11:59+0000 ephemeral=false lvl=info msg="Starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:12:02+0200 used=1970-01-01T00:00:00+0000
container=juju-db2944-18 err="Container is already running a start operation" lvl=eror msg="stop hook failed" t=2017-05-23T18:12:02+0200
action=start created=2017-05-23T16:11:59+0000 ephemeral=false lvl=eror msg="Failed starting container" name=juju-db2944-18 stateful=false t=2017-05-23T18:12:07+0200 used=1970-01-01T00:00:00+0000
created=2017-05-23T16:11:59+0000 ephemeral=false lvl=info msg="Deleting container" name=juju-db2944-18 t=2017-05-23T18:12:07+0200 used=2017-05-23T16:12:02+0000
created=2017-05-23T16:11:59+0000 ephemeral=false lvl=info msg="Deleted container" name=juju-db2944-18 t=2017-05-23T18:12:08+0200 used=2017-05-23T16:12:02+0000

vagrant@openstack:~$ juju version
2.1.2-xenial-amd64

[Stéphane Graber (stgraber)]

Looks like Juju is attempting to start the container twice. But that's not really the problem. The problem is that the container start itself failed.

We'd need all the files from /var/log/lxd/<container name>/ to investigate. That may be made somewhat difficult by the fact that Juju appears to clean everything up on failure.

The usual culprit for this is some invalid configuration in the Juju (or conjure-up) generated profile. Things like missing kernel modules or missing devices, but without the actual error message for the container in question, it's hard to tell exactly what's the issue.

[Ronny Lo Basso (ronny35)]

Hello,

Sorry for the missing logs.

[Ronny Lo Basso (ronny35)]

Hello,

No news for this topic ?

[Stéphane Graber (stgraber)]

lxc 20170523193141.848 ERROR lxc_conf - conf.c:instantiate_veth:2727 - failed to set mtu '9000' for veth pair (veth39WF6S and vethX3QWS9): Invalid argument

So the issue (rather well hidden by Juju) is that the container is configured to use an MTU of 9000 but the bridge it's connecting to won't allow it.

I seem to remember discussing this very issue with Adam Stokes (conjure-up) so this may have been fixed or should at least be known upstream. I'd recommend you get in touch with the conjure-up folks if the latest version doesn't fix this.

[Stéphane Graber (stgraber)]

Marking this bug as invalid for LXD.