openstack/heat
Code Issues Proposed changes

Wait for the policy to be done in tests

Browse Source 
As observed in the amqp job, we sometimes signal scale policy too fast,
without waiting for the medata to be set. It creates a timeout. We can
check the metadata of the policy before signaling again, to make sure
that we can move forward.

Change-Id: I9857803ef960efbd034f05985ec8fcc7272e2f70
This commit is contained in:
Thomas Herve
2017-04-10 23:44:26 +02:00
parent ffc1412ed5
commit 7109468b2f
4 changed files with 14 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
heat_integrationtests/common/test.py
View File

@@ -655,10 +655,14 @@ class HeatIntegrationTest(testscenarios.WithScenarios,
return matched
time.sleep(build_interval)
def check_autoscale_complete(self, stack_id, expected_num):
def check_autoscale_complete(self, stack_id, expected_num, parent_stack,
policy):
res_list = self.client.resources.list(stack_id)
all_res_complete = all(res.resource_status in ('UPDATE_COMPLETE',
'CREATE_COMPLETE')
for res in res_list)
all_res = len(res_list) == expected_num
return all_res and all_res_complete
if all_res and all_res_complete:
metadata = self.client.resources.metadata(parent_stack, policy)
return not metadata.get('scaling_in_progress')
return False