Issues with Trove

# Issues with Trove

(Openstack = Rocky / Trove = 2.16.0)

## Some background
We have an existing Openstack which was deployed using Kolla/Kayobe. I have managed to add Trove and it is appearing in the Horizon dashboard.
Created a test MySQL guest image using the trovestack integration script.
Added this image to glance then created the datastore using that image.
At first, when trying to build my instance it would just hang at "building" in the dashboard. Once checking the logs I could see it was sitting at the login prompt and had its IP address but was unable to gain access, via SSH or at the console.
I decided to mounted the original image created from the integration script, using guestfish, and was able to create a new user and added in my SSH keys, then created a new instance with the edited image.
I can now gain access to this instance via SSH, however, it is now dropping into an "Error" state on the Horizon Database dashboard and it looks like Trove is unable to communicate with the instance.

## Troubleshooting
After digging around our setup and it looks like we are missing the `trove-guestagent.conf` file, both on the instance and in the host config files.

I cloned the Trove files from GitHub and found a trove-guestagent.conf.sample file.
I have found a similar-looking file on the actual taskmanager docker container called `/etc/trove/trove.conf`[1] but I would have expected more in it, I think ...., it has some ports defined but all URLs are 0.0.0.0.

The syslogs and the cloud-init logs[2] in the trove instance are saying the guest agent is failing to start, there is a critical error just before it tries to load the `/usr/local/bin/guest-agent` file (this only has 4 lines in it, would you expect more? then there are errors at almost every related python file it tries to run after.

When trying to run `/opt/guest-agent-venv/bin/trove-guestagent` manually on the instance here is what I get back:
```
2020-05-06 09:17:50.261 11276 CRITICAL root [-] Unhandled error: RuntimeError: Manager class not registered for datastore manager None
2020-05-06 09:17:50.261 11276 ERROR root Traceback (most recent call last):
2020-05-06 09:17:50.261 11276 ERROR root   File "/opt/guest-agent-venv/bin/trove-guestagent", line 8, in <module>
2020-05-06 09:17:50.261 11276 ERROR root     sys.exit(main())
2020-05-06 09:17:50.261 11276 ERROR root   File "/opt/guest-agent-venv/lib/python3.5/site-packages/trove/cmd/guest.py", line 46, in main
2020-05-06 09:17:50.261 11276 ERROR root     raise RuntimeError(msg)
2020-05-06 09:17:50.261 11276 ERROR root RuntimeError: Manager class not registered for datastore manager None
2020-05-06 09:17:50.261 11276 ERROR root
```

Then running it again with the (/usr/local/bin/guest-agent) config file defined you get:
```
root@manzi-test-0405:~# /opt/guest-agent-venv/bin/trove-guestagent --config-file /etc/trove/conf.d/guest_info.conf
2020-05-06 09:21:23.530 11296 INFO trove.guestagent.module.driver_manager [-] Initializing module driver manager.
2020-05-06 09:21:23.538 11296 INFO trove.guestagent.module.driver_manager [-] Loading Module driver: ping
2020-05-06 09:21:23.541 11296 INFO trove.guestagent.module.driver_manager [-] Loading Module driver: new_relic_license
2020-05-06 09:21:23.541 11296 INFO trove.guestagent.module.driver_manager [-] Loaded module driver: ping
2020-05-06 09:21:23.541 11296 INFO trove.guestagent.module.driver_manager [-] Loaded module driver: new_relic_license
2020-05-06 09:21:23.542 11296 CRITICAL root [-] Unhandled error: TypeError: sequence item 0: expected str instance, NoneType found
2020-05-06 09:21:23.542 11296 ERROR root Traceback (most recent call last):
2020-05-06 09:21:23.542 11296 ERROR root   File "/opt/guest-agent-venv/bin/trove-guestagent", line 8, in <module>
2020-05-06 09:21:23.542 11296 ERROR root     sys.exit(main())
2020-05-06 09:21:23.542 11296 ERROR root   File "/opt/guest-agent-venv/lib/python3.5/site-packages/trove/cmd/guest.py", line 64, in main
2020-05-06 09:21:23.542 11296 ERROR root     rpc_api_version=guest_api.API.API_LATEST_VERSION)
2020-05-06 09:21:23.542 11296 ERROR root   File "/opt/guest-agent-venv/lib/python3.5/site-packages/trove/common/rpc/service.py", line 49, in __init__
2020-05-06 09:21:23.542 11296 ERROR root     self.manager_impl = profiler.trace_cls("rpc")(_manager)
2020-05-06 09:21:23.542 11296 ERROR root   File "/opt/guest-agent-venv/lib/python3.5/site-packages/osprofiler/profiler.py", line 243, in decorator
2020-05-06 09:21:23.542 11296 ERROR root     for attr_name, attr in inspect.getmembers(cls):
2020-05-06 09:21:23.542 11296 ERROR root   File "/usr/lib/python3.5/inspect.py", line 304, in getmembers
2020-05-06 09:21:23.542 11296 ERROR root     value = getattr(object, key)
2020-05-06 09:21:23.542 11296 ERROR root   File "/opt/guest-agent-venv/lib/python3.5/site-packages/trove/guestagent/datastore/manager.py", line 175, in guestagent_log_defs
2020-05-06 09:21:23.542 11296 ERROR root     guestagent_log = guestagent_utils.build_file_path(log_dir, log_file)
2020-05-06 09:21:23.542 11296 ERROR root   File "/opt/guest-agent-venv/lib/python3.5/site-packages/trove/guestagent/common/guestagent_utils.py", line 103, in build_file_path
2020-05-06 09:21:23.542 11296 ERROR root     file_name = os.extsep.join([base_name] + list(extensions))
2020-05-06 09:21:23.542 11296 ERROR root TypeError: sequence item 0: expected str instance, NoneType found
2020-05-06 09:21:23.542 11296 ERROR root
```

Which is exactly what is happening in the syslogs.

Have you ever seen anything like this?
Would you recommend copying over the trove.conf file from the taskmanager docker to the instance and see what happens?


[1] (trove.conf) https://pastebin.com/DTywJ6xS
[2] (Trove Logs) https://pastebin.com/RFPNCtee