Some providers release wrong resources when destroying hosted models
Bug #1536792 reported by
Cheryl Jennings
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Fix Released
|
Critical
|
Anastasia |
Bug Description
When destroying hosted models, resources associated with another model with the same name may also be released (i.e. security groups, disks, instances).
This currently affects Joyent, AWS and GCE providers.
In case of GCE the resources involved are Disks.
In case of AWS the involved resources are security groups, volumes and instances.
AWS and GCE will be fixed in Juju 2.0-alpha2. Joyent still needs to be fixed.
Changed in juju-core: | |
status: | New → Fix Committed |
importance: | Undecided → Critical |
milestone: | none → 2.0-alpha2 |
description: | updated |
description: | updated |
Changed in juju-core: | |
status: | Fix Committed → In Progress |
Changed in juju-core: | |
milestone: | 2.0-alpha2 → 2.0-beta1 |
Changed in juju-core: | |
milestone: | 2.0-beta1 → 2.0-beta2 |
Changed in juju-core: | |
milestone: | 2.0-beta2 → 2.0-beta3 |
Changed in juju-core: | |
assignee: | nobody → Anastasia (anastasia-macmood) |
Changed in juju-core: | |
status: | In Progress → Fix Committed |
Changed in juju-core: | |
status: | Fix Committed → Fix Released |
tags: | added: 2.0-count |
affects: | juju-core → juju |
Changed in juju: | |
milestone: | 2.0-beta3 → none |
milestone: | none → 2.0-beta3 |
To post a comment you must log in.
This is the scenario that I am running:
1. Bootstrap 2 controllers (by default, I'll get 2 host models with same name as well - "default").
2. Add a machine to each hosted non-admin model.
3. Remove one of the machines.
4. Make sure that the other controller has all its artifacts intact.
5. Destroy one controller
6. Make sure that the other controller is intact.
7. Destroy remaining controller.
JOYENT failures:
3. Remove on of the machine - I can't specify which model to remove from: admin@local: =default not found
>juju remove-machine 0 -m=default
ERROR model local.tags:
If I omit -m flag, the command returns fine but no machines are destroyed.
6. Once one controller is successfully destroyed, I cannot get status nor run list-models against the remaining controller. Command never returns.
7. Killing remaining controller fails with:
Unable to open API: open connection timed out
Unable to connect to the API server. Destroying through provider.