juju should surface blocked volume deletion better

[Junien F]

Hi,

In bug 1722818, we investigated model removal failures that were actually simply caused by a process holding up a volume on an instance. This is information is easily seen _if_ you know about the "juju storage --format yaml" command (see example output below).

However, "juju status" only reported "attempt 1 to destroy model failed (will retry): model not empty, found X machines, X volumes, X filesystems (model not empty)".

Could you please either make "juju status" display the "message" below, or have "juju status" mention the "juju storage" or even "juju storage --format yaml" commands as a way to diagnose the problem ?

Thanks

 0/0:
    provider-id: filesystem-0-0
    volume: "0"
    attachments:
      machines:
        "0":
          mount-point: /srv/mnt/jenkins
          read-only: false
          life: dying
    pool: cinder
    size: 102400
    life: dying
    status:
      current: detaching
      message: |-
        umount failed: umount: /srv/mnt/jenkins: target is busy
                (In some cases useful info about processes that
                 use the device is found by lsof(8) or fuser(1).): exit status 32
      since: 02 Nov 2017 14:51:24Z

[Andrew Wilkins (axwalk)]

There have been many requests for storage in status, it just keeps getting deferred. Hopefully this is enough of an impetus.