Details
-
Bug
-
Resolution: Won't Do
-
P3: Somewhat important
-
None
-
master
-
None
Description
Environment: coin-master connected to opennebula 5.8.5 on one-master-eighth.
See attached screenshot. At the end of the log, there are two lines after the line "provision succeeded". These were not seen in the coin log, I had to connect with VNC to see them.
Provision succeeded Exiting upon agent request:... Error running shutdown command:...
The coin log has the following lines after "provision succeeded":
agent:2020/01/07 09:57:55 agent.go:272: Provision succeeded
INFO:workitem.py:634(36235): Agent FINISHED PASS: provisioning/qtci-linux-SLES-15-x86_64-e8e160/Provision
INFO:workitem.py:704(36235): VM metrics
INFO:workitem.py:706(36235): Host metrics
ERROR:workitem.py:416(36235): Houston? Houston! Unhandled exception occurred while running a WorkItem. This is potentially serious! Investigate it immediately.
Traceback (most recent call last):
File "/home/diaposto/src/tqtc-coin-ci/src/workitem.py", line 380, in _run
await self.run()
File "/home/diaposto/src/tqtc-coin-ci/src/workitem.py", line 870, in run
raise AgentException("An instruction failed.")
File "/home/diaposto/src/tqtc-coin-ci/src/agentlauncher.py", line 188, in __aexit__
await self._exit_action()
File "/home/diaposto/src/tqtc-coin-ci/src/opennebulahardwarepool.py", line 361, in save_agent_as_template
await _block_until_vm_is_in_state(self.vm_id, one_vm_state.poweroff, max_iterations=6 * 5)
File "/home/diaposto/src/tqtc-coin-ci/src/opennebulahardwarepool.py", line 192, in _block_until_vm_is_in_state
await _block_until_object_is_in_state(getter, vm_id, state, "VM", ERROR_STATES, check_interval, max_iterations)
File "/home/diaposto/src/tqtc-coin-ci/src/opennebulahardwarepool.py", line 174, in _block_until_object_is_in_state
raise TimeoutError(message % (type, id, current_state, state))
concurrent.futures._base.TimeoutError: VM (id:4381) is still in a undesired state (3), while we wanted to see it in state: one_vm_state.poweroff
Because of the missing "error running shutdown command" it was quite hard to figure out what went wrong with this provisioning.
