Race between test_enlist_node and test_commission_node
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| maas-test |
Fix Committed
|
High
|
Raphaël Badin | ||
Bug Description
I think I've stumbled on a possible race condition.
Originally, I observed this:
test_power_up_node starts
Node powers up
test_power_up_node ends
test_enlist_node starts
Node PXEs and enlists
test_enlist_node ends
test_commission
Node powers off
test_commission
Node NEVER powers back on.
I worked around this manually by pushing the power button on my SUT during test_commission
Then, I retried and observed THIS behaviour:
test_power_up_node starts
Node powers up
test_power_up_node ends
test_enlist_node starts
Node PXEs by this point and enlists
Node Powers Off
test_enlist_node ends
test_commission
Node powers up
Node PXEs and commissions
test_commission
So that makes it seem that there's a race between enlistment and commissioning that prevents test_commission
Related branches
- Jeroen T. Vermeulen (community): Approve
-
Diff: 23 lines (+13/-0)1 file modifiedmaastest/cases.py (+13/-0)
| Raphaël Badin (rvb) wrote | #1 |
| Changed in maas-test: | |
| status: | New → Triaged |
| importance: | Undecided → High |
| Changed in maas-test: | |
| assignee: | nobody → Raphaël Badin (rvb) |
| status: | Triaged → In Progress |
| summary: |
- possible race between test_enlist_node and test_commission_node + Race between test_enlist_node and test_commission_node |
| Changed in maas-test: | |
| status: | In Progress → Fix Committed |
Looks like a race indeed. It's likely that the BMC dropped the "power up" command from test_commission