Delete a customer's project
Deleting a project removes its resources from the cloud, then the project itself. Customers delete their own projects in the client portal. You delete one for them from Client Area > Projects, by opening the project and choosing Schedule for deletion.
Check before deleting
Before anything is deleted, OSIE checks the project against each cloud it lives on. It signs in to the project as the cloud's administrator user, and looks for resources OpenStack refuses to delete in their current state: a protected image, a volume that is still being attached or extended, a load balancer whose last change has not finished, or a database that is still being created. While any of these remain, the deletion is blocked. Resources whose deletion may still fail, such as a load balancer in an error state, are reported without blocking.
The check repeats when the deletion runs. A project that passes and is then changed can still be refused.
Force a deletion
When only resources block the deletion, you can force it. The project leaves OSIE at once and is no longer billed. Whatever OpenStack refuses to delete stays in the project on the cloud, and OSIE deletes that cloud project once it is empty.
You cannot force a deletion that OSIE has no access to: when the administrator user cannot sign in, or OpenStack refuses it access to the project. Fix that access first. Cloud diagnostics show what is missing, and OSIE's own roles on the mapped domain explains the roles OSIE needs in customer domains.
A deletion that stopped
A deletion stops when OSIE has lost access to the project since the check, before anything is deleted. It also stops when a cloud cannot take over what remains, or on an unexpected error. The project is then Disabled, and its page shows why the deletion stopped. Resolve the cause, then delete the project again.
Delete a project through the Admin API
Your own systems delete projects with Delete Project, for example to recreate a demonstration project every night. The request runs the same deletion and the same checks as the admin dashboard.
- When it starts. By default the deletion is scheduled, starts 5 to 6 minutes later, and can be cancelled
from the admin dashboard until then. With
now=trueit starts at once. - Resources in the project. A project that still holds cloud resources is refused with HTTP 400, unless the
request sets
cascade=trueto remove them. The defaults OpenStack creates in every project do not count. Withnow=truethe resources are always removed. - Blocked deletions. When the check before deleting finds something that blocks the deletion, the request
is refused with HTTP 409
PROJECT_DELETION_BLOCKED, anddetails.checklists what the check found, per cloud and per resource.
The response holds the project in its new status, SCHEDULED_FOR_DELETION or DELETE_IN_PROGRESS. Follow it
with Get Project: a 404 means the deletion has finished. A project
that turns DISABLED instead is a deletion that stopped. Send the DELETE again
once the cause is resolved. When the attempt that stopped began less than 20 minutes earlier, the new one waits
until those 20 minutes are up.