Duplicate key error on juju.txns.stash creating model

[Mark Shuttleworth]

Just saw this:

  $ juju models
  MODEL OWNER STATUS LAST CONNECTION
  controller* admin@local available 12 seconds ago
  bigdata admin@local destroying never connected

  mark@maas:~$ juju add-model kubes

  ERROR failed to create new model: E11000 duplicate key error collection: juju.txns.stash index: _id_ dup key: { : { c: "statuseshistory", id: ObjectId('5765d034ffc8d5d51c65b348') } }
mark@maas:~$

It might be related to the destroy-model proceeding in the background.

Interestingly, though the model got created:

  $ juju models
  MODEL OWNER STATUS LAST CONNECTION
  controller* admin@local available 7 minutes ago
  kubes admin@local available never connected

Though I don't know if it's broken in some way or not.

[Cheryl Jennings (cherylj)]

As I have been triaging other bugs in 2.0, I have also seen this error in logs, most commonly in reference to the presence collection. In those cases, the pingers would restart and no error was surfaced up to the users.

@sabdfl - Was your controller machine on trusty or xenial? (if on trusty, it would be still be using mongo 2.4) I'll also try to dig up some other examples to see if there's some correlation between mongo version and this error.

[Mark Shuttleworth (sabdfl)]

In this case the controller is Xenial.

Mark