Just after bootstrap, juju model-defaults is called to set apt-https-proxy and apt-http-proxy but returns a non-zero exit code:
juju model-defaults -c foundations-maas apt-https-proxy=http://squid.internal:3128 apt-http-proxy=http://squid.internal:3128
A juju-crashdump from the controller is attached and has the following errors in machine-0.log around the time of the failure:
2018-08-25 20:55:36 ERROR juju.worker.dependency engine.go:587 "lease-clock-updater" manifold worker returned unexpected error: updating global clock: lease operation timed out
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "lease-manager" manifold worker returned unexpected error: lease operation timed out
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "is-responsible-flag" manifold worker returned unexpected error: lease manager stopped
2018-08-25 20:55:37 ERROR juju.api.watcher watcher.go:92 error trying to stop watcher: write tcp 127.0.0.1:48512->127.0.0.1:17070: write: broken pipe
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "not-alive-flag" manifold worker returned unexpected error: watcher has been stopped (stopped)
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "not-dead-flag" manifold worker returned unexpected error: watcher has been stopped (stopped)
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "valid-credential-flag" manifold worker returned unexpected error: watcher has been stopped (stopped)
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "api-caller" manifold worker returned unexpected error: api connection broken unexpectedly
2018-08-25 20:55:37 ERROR juju.worker.dependency engine.go:587 "api-caller" manifold worker returned unexpected error: api connection broken unexpectedly
I cannot reproduce this on current 2.6.4. I must have been fixed as part of other work in the area.
If you are still seeing this behavior, please open a new report with latest information - logs, etc.