Comment 31 for bug 1410876

We are experiencing this in our automated OpenStack bare metal testing (UOSCI). It appears to be a race, as it is intermittently an issue. It is frequent enough to cause noticeable test failures though.

The observable symptom is that juju-deployer times out, as some container units are forever in an "allocating" & "Waiting for agent initialization to finish" state.

The impact is that we cannot reliably test OpenStack deployments on baremetal.

Juju version is 1.24.2-0ubuntu1~14.04.1~juju1 fro ppa:juju/stable.

# from the machine log:
2015-07-21 19:38:35 DEBUG juju.worker.logger logger.go:45 reconfiguring logging from "<root>=DEBUG" to "<root>=WARNING;unit=DEBUG"
2015-07-21 19:41:05 ERROR juju.provisioner.lxc lxc-broker.go:163 failed to start container: lxc container cloning failed: error executing "lxc-clone": lxc_container: utils.c: mkdir_p: 267 Not a directory - failed to create directory '/run/lock/lxc//var/lib/lxc/juju-trusty-lxc-template/snaps'; lxc_container: lxccontainer.c: do_snapshot_destroy: 3210 Could not destroy snapshot %s - failed to allocate a pty; Insufficent privileges to control juju-trusty-lxc-template
2015-07-21 19:41:05 ERROR juju.provisioner provisioner_task.go:597 cannot start instance for machine "4/lxc/0": lxc container cloning failed: error executing "lxc-clone": lxc_container: utils.c: mkdir_p: 267 Not a directory - failed to create directory '/run/lock/lxc//var/lib/lxc/juju-trusty-lxc-template/snaps'; lxc_container: lxccontainer.c: do_snapshot_destroy: 3210 Could not destroy snapshot %s - failed to allocate a pty; Insufficent privileges to control juju-trusty-lxc-template
2015-07-21 19:41:06 ERROR juju.provisioner.lxc lxc-broker.go:163 failed to start container: lxc container cloning failed: cannot clone a running container
2015-07-21 19:41:06 ERROR juju.provisioner provisioner_task.go:597 cannot start instance for machine "4/lxc/1": lxc container cloning failed: cannot clone a running container

# More detail:
http://paste.ubuntu.com/11920907/