firewaller dies if a machine has been removed directly via the provider API

[Junien F]

Hi,

Using juju 2.5.4, the firewaller worker dies if a deployed machine has been deleted using the provider API directly. This causes the worker to be continuously restarted. I believe it will log the following :

2019-05-23 10:00:21 ERROR juju.worker.dependency engine.go:636 "firewaller" manifold worker returned unexpected error: cannot respond to units changes for "machine-10": instances not found

and/or

2019-05-23 10:00:21 ERROR juju.worker.dependency engine.go:636 "firewaller" manifold worker returned unexpected error: machine 0 not provisioned

How to repro (at least on AWS) :

juju deploy ubuntu
juju run --unit ubuntu/0 "open-port 80"
aws ec2 terminate-instances --instance-ids <id_of_machine_0>
juju expose ubuntu

Thanks

[Joseph Phillips (manadart)]

Related to https://bugs.launchpad.net/juju/+bug/1814277.

[Canonical Juju QA Bot (juju-qa-bot)]

This bug has not been updated in 2 years, so we're marking it Low importance. If you believe this is incorrect, please update the importance.