Attaching juju status tabular output, not sure what else is needed, from what I can tell, the agents on the machines are alive.
Showing that the processes are running on one of the systems in question:
ubuntu@juju-machine-0-lxc-2:/etc/init$ ls juju*
jujud-machine-0-lxc-2.conf jujud-unit-trusty-client-0.conf
ubuntu@juju-machine-0-lxc-2:/etc/init$ sudo service jujud-machine-0-lxc-2 status
jujud-machine-0-lxc-2 start/running, process 846
ubuntu@juju-machine-0-lxc-2:/etc/init$ sudo service jujud-unit-trusty-client-0 status
jujud-unit-trusty-client-0 start/running, process 887
ubuntu@juju-machine-0-lxc-2:/etc/init$ ps -ef |grep 846
root 846 1 0 Jun11 ? 00:00:29 /var/lib/juju/tools/machine-0-lxc-2/jujud machine --data-dir /var/lib/juju --machine-id 0/lxc/2 --debug
ubuntu 1909 1868 0 14:12 pts/1 00:00:00 grep --color=auto 846
ubuntu@juju-machine-0-lxc-2:/etc/init$ ps -ef |grep 887
root 887 1 0 Jun11 ? 00:00:43 /var/lib/juju/tools/unit-trusty-client-0/jujud unit --data-dir /var/lib/juju --unit-name trusty-client/0 --debug
ubuntu 1911 1868 0 14:12 pts/1 00:00:00 grep --color=auto 887
Status History Command, not sure if this is useful or not:
dpb@helo:slaves[0]$ juju status-history trusty-client/0
TIME TYPE STATUS MESSAGE
11 Jun 2015 12:19:25-06:00 workload unknown Waiting for agent initialization to finish
11 Jun 2015 12:19:25-06:00 agent allocating
11 Jun 2015 13:06:57-06:00 workload maintenance installing charm software
11 Jun 2015 13:07:01-06:00 agent executing running install hook
11 Jun 2015 13:17:50-06:00 agent error hook failed: "install"
11 Jun 2015 13:19:38-06:00 workload maintenance installing charm software
11 Jun 2015 13:19:39-06:00 agent executing running install hook
11 Jun 2015 14:52:40-06:00 agent executing running leader-elected hook
11 Jun 2015 14:52:48-06:00 agent executing running config-changed hook
11 Jun 2015 14:53:17-06:00 agent executing running start hook
11 Jun 2015 14:53:21-06:00 workload unknown
11 Jun 2015 14:53:25-06:00 agent idle
12 Jun 2015 00:54:11-06:00 agent executing running leader-elected hook
12 Jun 2015 00:54:33-06:00 agent executing running config-changed hook
12 Jun 2015 00:55:00-06:00 agent idle
14 Jun 2015 00:50:10-06:00 agent executing running leader-elected hook
14 Jun 2015 00:50:35-06:00 agent executing running config-changed hook
14 Jun 2015 00:50:47-06:00 agent idle
Pretty sure this is the cause (from one of the "failed" units):
2015-06-15 15:51:51 ERROR juju.worker. uniter. filter filter.go:137 tomb: dying uniter. filter filter.go:137 tomb: dying uniter. operation leader.go:111 we should run a leader-deposed hook here, but we can't yet uniter. filter filter.go:137 tomb: dying uniter. filter filter.go:137 tomb: dying uniter. filter filter.go:137 tomb: dying uniter. filter filter.go:137 tomb: dying uniter. filter filter.go:137 tomb: dying
2015-06-15 15:51:51 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: worker stopped
2015-06-15 15:54:00 ERROR juju.worker.
2015-06-15 15:54:00 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: worker stopped
2015-06-15 15:54:24 WARNING juju.worker.
2015-06-15 15:54:24 ERROR juju.worker.
2015-06-15 15:54:24 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: worker stopped
2015-06-15 15:54:48 ERROR juju.worker.
2015-06-15 15:54:48 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: worker stopped
2015-06-15 16:00:04 ERROR juju.worker.
2015-06-15 16:00:04 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: worker stopped
2015-06-15 17:08:00 ERROR juju.worker.
2015-06-15 17:08:00 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: worker stopped
2015-06-15 17:08:07 ERROR juju.worker.
2015-06-15 17:08:07 ERROR juju.worker runner.go:219 exited "uniter": leadership failure: unable to make a leadership claim: no active lease manager