maas admin machine test --help
A node in the 'ready', 'allocated', 'deployed', 'broken', or any failed
state may run tests. If testing is started and successfully passes from
'broken' or any failed state besides 'failed commissioning' the node
will be returned to a ready state. Otherwise the node will return to
the state it was when testing started.
The MAAS CLI help page explicitly states that after tests are completed, if the tests pass, the node’s status will revert to its state before testing. I have also confirmed this through the web UI; for a deployed machine, upon completion of testing, the machine’s status remains in the deployed state. This behavior is by design.
However, there is a scenario where, for a machine already in a deployed state, if during testing you select “allow SSH access and prevent machine powering off,” then upon successful completion of the test, although the machine’s status reverts to deployed, in reality, the machine is running squashfs instead of the actual OS installed on the hard drive.
In this situation, if someone checks the machine’s status via the MAAS CLI (machine read) or the MAAS web , she/he would not be able to know that the machine is actually still running the squashfs introduced by the testing process.
I would like the MAAS web interface to flag this condition to reflect the machine’s true state accurately.