juju destroy-environment --force tries to delete unrelated container
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Won't Fix
|
Medium
|
Unassigned |
Bug Description
When there's no environment and I run juju destroy-environment charm-testing-hp --force --debug, it tries to delete an unknown container and fails:
2014-08-21 17:13:04 ERROR juju.cmd supercommand.go:323 failed to delete container:
caused by: Unauthorised URL https:/
caused by: request (https:/
If I manually specify a control bucket in environments.yaml, or if there's a .jenv, this does not happen.
It always tries to delete the same container. I wonder whether it's generating a control-bucket with an un-initialized psuedo-random number generator.
Note also that there's no debug message about deleting a container, just instances.
We see this frequently in CI, but it seems benign, so far.
$ juju destroy-environment charm-testing-hp --force --debug
2014-08-21 17:12:51 INFO juju.cmd supercommand.go:37 running juju [1.20.5-
2014-08-21 17:12:51 INFO juju.provider.
WARNING! this command will destroy the "charm-testing-hp" environment (type: openstack)
This includes all machines, services, data and other resources.
Continue [y/N]? y
2014-08-21 17:12:54 INFO juju.provider.
2014-08-21 17:12:55 DEBUG juju.provider.
2014-08-21 17:13:04 ERROR juju.cmd supercommand.go:323 failed to delete container:
caused by: Unauthorised URL https:/
caused by: request (https:/
This happens in HP Cloud US East. It does not happen in US West or in canonistack.
description: | updated |
tags: | added: openstack-provider |
description: | updated |
tags: | added: hp-cloud |
tags: | removed: hp-cloud |
Changed in juju-core: | |
status: | Triaged → Won't Fix |