Image v1 json client timeout while checking that image was deleted
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Glance |
Invalid
|
Undecided
|
Unassigned | ||
tempest |
Invalid
|
Undecided
|
Unassigned |
Bug Description
Description:
- Icehouse Openstack cloud
- Tempest commit 4b1b8cfd4526203
I'm getting tearDown class error after executing image tests
<testcase classname="" name="tearDownClass (tempest.
<failure type="testtools
File "/home/
cls.
File "/home/
raise exceptions.
TimeoutException: Request timed out
After some investiogation, i've found that client tried to check deleted VM by
def is_resource_
try:
except exceptions.
return True
return False
But get_image_meta able to get metadata of deleted image, it just has status = deleted.
My quick fix was:
def is_resource_
try:
_, meta = self.get_
if meta['status'] == 'deleted':
except exceptions.
return True
return False
information type: | Public → Public Security |
information type: | Public Security → Public |
Glance shows deleted images when doing a show or a list with a filter such as the "changes-since" filter .
With Admin credential, by default deleted image is visible with show but with non-admin credential show will not show the deleted image. If show is run as non-admin credential then, It will return NotFound error.
Tempest images tests are run as non-admin credential and expect NotFound from self.get_ image_meta( id) (when image is deleted) which is correct.
IMO, there is no issue in Tempest's is_resource_ deleted( self, id) function. If this could have been the issue as mentioned by you in quick fix then, it should fail every time. On gate those tests pass.
Can you please provide glance log so that timeout error can be debugged more?