'ERROR connecting with bootstrap config' for status on destroyed model

What version of juju were you running when you hit this?

Can you confirm that the client version matched the agent version?

Testing with beta11, I see the following message:

$ juju add-model foo
Added 'foo' model on aws/us-west-2 with credential 'cherylj' for user 'admin'

$ juju destroy-model foo -y

$ juju status
ERROR no model in focus

Please use "juju models" to see models available to you.
You can set current model by running "juju switch"
or specify any other model on the command line using the "-m" flag.

I was running 2.0b11 freshly bootstrapped.

Mark

I played with this and can get the b11 message if I destroy the model, switch to another model, and switch back to the dying model. In Azure it was taking > 60s for that model to disappear. In the meantime I could run juju status on the dying model and get an empty status output.

Once the model finally went away I'd get the

 ERROR connecting with bootstrap config: aborted

message.

I upgraded to beta12 to see what would happen and experienced a different message but the same behavior. In beta12 you get

juju status
ERROR unknown model: "55459f0f-56c8-4f0b-8cf7-b8a1042df2aa" (not found)

A sample output for steps to reproduce from b11, b12 is the same with only the error message in the end being different.

https://pastebin.canonical.com/161089/

Thanks for the repro Rick,

Mark

Pull request https://github.com/juju/juju/pull/5849

Note when printing the error message: an admin may have revoked access to a model the current user has cached in their models.yaml file, and there's not (yet) an API to query the controller for a list of accessible models, but nonetheless it will almost always be close enough simply to print the count of the models in the local models.yaml file. So we do that for now.