Juju on maas, when issuing a juju upgrade-charm connection hangs forever without any obvious errors.
jujumanage@maas:~ $ juju upgrade-charm landscape-postgresql --debug --show-log
2014-12-22 17:18:52 INFO juju.cmd supercommand.go:37 running juju [1.20.9-trusty-amd64 gc]
2014-12-22 17:18:52 DEBUG juju.conn api.go:187 trying cached API connection settings
2014-12-22 17:18:52 INFO juju.conn api.go:270 connecting to API addresses: [x.x.x.19:17070 os-1.maas:17070 os-1.maas:17070 x.x.x.12:17070 x.x.x.1:17070 os-2.maas:17070 os-2.maas:17070 x.x.x.27:17070 x.x.x.13:17070 x.x.x.1:17070 os-3.maas:17070 os-3.maas:17070 x.x.x.26:17070 x.x.x.14:17070 x.x.x.1:17070]
2014-12-22 17:18:52 INFO juju.state.api apiclient.go:242 dialing "wss://x.x.x.19:17070/environment/e810b542-2eed-492a-8df0-62cb08e2dc01/api"
2014-12-22 17:18:52 INFO juju.state.api apiclient.go:176 connection established to "wss://x.x.x.19:17070/environment/e810b542-2eed-492a-8df0-62cb08e2dc01/api"
2014-12-22 17:18:53 INFO juju.utils http.go:66 hostname SSL verification disabled
...Hangs...
Simultaneous packet captures on maas node and juju master show a session is established on 17070, several PSH packets are exchanged, then the connection remains open with no traffic exchanged until the 'juju upgrade-charm' command is control-c'd (11 minutes later in one test). If helpful I can scrub these captures and attach. Strace of jujud during an upgrade attempt is attached.
Have tried a jujud restart on juju master, and a mongo cluster restart by way of successful juju-backup runs. Other juju commands, (juju run, juju ssh, juju stat, juju get) are successful.
juju-core: 1.20.9-0ubuntu1~14.04.1~juju1
This was raised against 1.20.9, does it still occur with a recent Juju?