Ah. I know! The magic smoke is leaking from places. But, yes. That would explain it, and apparently we have a ticket for this, but it doesn't have a high priority so it'll be fixed after new year's.
The way the drive deletion works is that we put drive in a special state, called deleting, when you delete them. We then filter them out from responses, until the deletion actually happens (which generally doesn't take very long, but we throttle it so it can take several minutes. Problem solved by not filtering them out of responses... Except that when you try to attach it to a guest or put it in a tag, the status of the drive isn't checked, so the operation succeeds (for guests, it'll fail when you try to start it), but you just cannot see the resource. Sorry for the confusion (took me a while to realise what's happening since I was doing the test manually, with a non-existent resource, instead of a deleting one.) --- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds-labs/pull/41#issuecomment-30212652
