Comment 19 for bug 1674148

Andres, I'm not sure I understand your reply in #16. It seems you think that we're not getting a disk back, but that's not the case. The issue is that they're not being mapped from labels as expected.

What we expect is that the first constraint in the list is mapped to the root filesystem's block device. When I looked at the MAAS code it appeared that there's code to do that. But from the output, you can see that the ID in constraints_by_type for "root" does not match the ID for the "/" block device in physicalblockdevice_set.