Red Hat Bugzilla – Bug 1480577
Removing OpenStack Cloud provider doesn't delete network data from database
Last modified: 2017-12-05 10:12:49 EST
Created attachment 1312117 [details]
Screenshot of duplicated entities
Description of problem:
After removing OpenStack cloud provider, all network information except network providers are still in the database.
Version-Release number of selected component (if applicable):
downstream 58-z 18.104.22.168
Steps to Reproduce:
1. Add OpenStack cloud provider
2. Navigate to Networks->Providers, Networks->Subnets etc. and
check added data
3. Remove OpenStack cloud provider
4. Check network data once again
1. All network data except network providers are still in the database
2. From other entities like subnets and routers disappeared Network Manager
3. Networks->Topology should display network data, if they are in database
1. All network data should disappear with removing cloud provider
2. Other entities should disappear with network provider entirely, not just
3. No network data are shown in Networks->Topology
1. Topology can't display the data correctly, because root object "network
provider" for them no longer exists in the database.
2. After adding OpenStack provider once again, other entities like subnets and
routers were duplicated (one with and one without network manager)
Gilles, could you take a look at this one?
Unfortunately, I haven't been able to reproduce the same symptoms on manageiq/fine (5.8.x) branch.
The closest I've obtained is, after multiple tests, I've seen a couple of times a delay to occur where the networks are not yet removed. This is caused by the delete task being delayed as the task is running in the background in async mode.
Could you please verify and test latest version of 5.8.x?
sorry for the late response, I was on vacation and I'm a bit busy right know, but tomorrow I'll take a look at it and retest it.
I was able to reproduce this bug with automated tests (still in dev, repo integration_tests PR #5133). If I added providers this way
pytest_generate_tests = testgen.generate(
classes=[AzureProvider, EC2Provider, OpenStackProvider], scope='module')
pytestmark = pytest.mark.usefixtures('setup_provider')
which automatically and successfully added those providers,
manually checking the cloud provider and selecting
configuration->remove Cloud Providers
after test run caused this bug again.
In my testing appliance are other network entities still there after half an hour, which would be a quite long delay.
I understand you're facing the issue on aforementioned PR#5133 meanwhile I believe that repo gets updates from downstream.
Would you be able to verify against latest Fine branch?
Few more times I wasn't able to reproduce the bugs either.