Node stuck in Commissioning state after virsh power on failed
Bug #1277658 reported by
Jason Hobbs
This bug report is a duplicate of:
Bug #1273222: MAAS doesn't check return values of power actions.
Edit
Remove
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| MAAS |
New
|
Undecided
|
Unassigned | ||
Bug Description
After enlisting a node and configuring it to use virsh power, I started commissioning it. virsh was configured incorrectly - no ssh-key - setup - and a PowerActionFail() error was raised in the cluster worker. The error didn't bubble up enough to change the state of the node, and I had to recover by manually changing the node's state in the db.
I'm using the dev environment, revno 1891, on trusty.
Revision history for this message
| Jason Hobbs (jason-hobbs) wrote | #1 |
Revision history for this message
| Julian Edwards (julian-edwards) wrote | #2 |
To post a comment you must log in.
I think this is actually fine in terms of what the node is doing, it's waiting to get powered up so it can complete the commissioning cycle. There is no other natural state for the node at this stage, going backwards to "Declared" is no use.
(You needn't have hacked the DB state, you can just power the node up manually.)
The bug here is that external power script failures are not displayed to the user. Also, we probably need a "retry" for power.
I think therefore this is a duplicate of bug 1273222