Tacker should not add VNF to monitoring service if mgmt_url is None
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tacker |
Fix Released
|
Medium
|
Sripriya |
Bug Description
Tacker currently adds VNF instance to monitoring thread even though VNF is an ERROR state which can be due to multiple reasons:
Instance creation fails, mgmt IP addr assignment is not successful or any other internal ERROR for which Tacker service throws an ugly stack trace.
Traceback (most recent call last):
File "/usr/local/
func(*args, **kwargs)
File "/opt/stack/
self.
File "/opt/stack/
device_dict, action_cb)
File "/opt/stack/
device_
File "/opt/stack/
return json.loads(
File "/opt/stack/
raise TypeError("%s can't be decoded" % type(text))
TypeError: <type 'NoneType'> can't be decoded
This error can be misleading as the actual internal error is masked and it appears to the user that there is problem with monitoring itself.
Expected behavior:
Tacker should handle the actual ERROR gracefully and not mislead the user with stack trace dump
Changed in tacker: | |
assignee: | nobody → Sripriya (sseetha) |
Changed in tacker: | |
importance: | Undecided → Medium |
Fix proposed to branch: master /review. openstack. org/256104
Review: https:/