Add sleep(2) when try to delete a container
Even though in devstack this test pass successfully, the same
don't happen in HA deployment. It takes a few seconds to
synchronize the deletion of the objects between the controllers
in order to delete the container. Without this time this test
fails consistently in tripleo as you can see in the logs at
http://bit.ly/2gMDl8d.
Adding the sleep makes the HA proxy have enough time to sync
the deletion of the objects, and the container is deleted
successfully.
Change-Id: I76f6493d1429166bef9804f2fbf2acdc02425489
diff --git a/tempest/api/object_storage/base.py b/tempest/api/object_storage/base.py
index 1b1ffd1..52b0a9c 100644
--- a/tempest/api/object_storage/base.py
+++ b/tempest/api/object_storage/base.py
@@ -13,6 +13,8 @@
# License for the specific language governing permissions and limitations
# under the License.
+import time
+
from tempest.common import custom_matchers
from tempest import config
from tempest.lib.common.utils import data_utils
@@ -102,6 +104,10 @@
The containers should be visible from the container_client given.
Will not throw any error if the containers don't exist.
Will not check that object and container deletions succeed.
+ After delete all the objects from a container, it will wait 2
+ seconds before delete the container itself, in order to deployments
+ using HA proxy sync the deletion properly, otherwise, the container
+ might fail to be deleted because it's not empty.
:param container_client: if None, use cls.container_client, this means
that the default testing user will be used (see 'username' in
@@ -121,6 +127,9 @@
for obj in objlist:
test_utils.call_and_ignore_notfound_exc(
object_client.delete_object, cont, obj['name'])
+ # sleep 2 seconds to sync the deletion of the objects
+ # in HA deployment
+ time.sleep(2)
container_client.delete_container(cont)
except lib_exc.NotFound:
pass