Tests failed when using floating as connect_method
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tempest |
Fix Released
|
High
|
Zhenyu Zheng |
Bug Description
Tests
tempest.
tempest.
are under the same class, so when these tests setup, and using floating as connect_method only one floating IP was created.We first associated it to instance created at test1 and then we created another instance in test2 and tried to associate this IP again to the new server, since the IP was already associated to instance created in test1, error will happen when try to do this.
Actually this tests are about volumes, there is no need to create two instances for these tests, just create one instance and use it for all the tests.
Test Log:
tempest.
-------
Captured traceback:
~~~~~~~~~~~~~~~~~~~
Traceback (most recent call last):
File "/home/
File "/home/
File "/home/
**kwargs)
File "/home/
% server['id'])
File "/home/
File "/home/
File "/home/
File "/home/
resp, body = self.post(url, post_body)
File "/home/
return self.request(
File "/home/
method, url, extra_headers, headers, body, chunked)
File "/home/
resp, resp_body)
File "/home/
raise exceptions.
tempest.
Details: {u'message': u'Unable to associate floating ip 46.30.2.0 to fixed ip 192.168.0.126 for instance a6e6ffa1-
Changed in tempest: | |
assignee: | nobody → Zhenyu Zheng (zhengzhenyu) |
description: | updated |
Changed in tempest: | |
status: | New → In Progress |
Changed in tempest: | |
status: | Confirmed → In Progress |
Actually this is not about creating 1 server in test class, it is like we create only 1 floating ip in validation resources and try to associate the same FIP to all instance requesting validation.
Validation resources are cleaned up at teardwonClass level.
I think we did not tested connect_method as floating ip and i am sure there will more cases where single class try to use same FIP in multiple servers.
For this test fail case, we can clean the server during teardown itself instead if waiting till teardwonClass.
But as general solution we need to think more on this. We can either create multiple FIP as per test class requirement or we cleanup validation resource during tearDown.