[BUG] Anonymous group, Thredds config page, not working after upgrade from 1.7.3 to 3.5.1

[tlvu]

Describe the bug
Going here https://lvupavics.ouranos.ca/magpie/ui/groups/anonymous/thredds

Result in this page:

With the following error in docker logs magpie.

[2021-03-25 15:02:18,563] INFO       [ThreadPoolExecutor-0_0][magpie.utils] Request: [GET lvupavics.ouranos.ca:443 /magpie/services/thredds/resources]
[2021-03-25 15:02:18,826] ERROR      [ThreadPoolExecutor-0_0][magpie.ui.utils] Unexpected API error under UI operation. [AttributeError: 'result' object has no attribute 'Resource']
Traceback (most recent call last):
  File "/opt/local/src/magpie/magpie/ui/utils.py", line 121, in wrap
    return func(*args, **kwargs)
  File "/usr/local/lib/python3.7/site-packages/pyramid/viewderivers.py", line 436, in rendered_view
    result = view(context, request)
  File "/usr/local/lib/python3.7/site-packages/pyramid/viewderivers.py", line 116, in _class_requestonly_view
    response = getattr(inst, attr)()
  File "/opt/local/src/magpie/magpie/ui/management/views.py", line 751, in edit_group
    info = self.get_remote_resources_info(res_perms, services, session)
  File "/opt/local/src/magpie/magpie/ui/management/views.py", line 808, in get_remote_resources_info
    res_perms = self.merge_remote_resources(res_perms, services, session)
  File "/opt/local/src/magpie/magpie/ui/management/views.py", line 823, in merge_remote_resources
    resources_for_service = merge(res_perms, service_values["service_sync_type"], service_id, session)
  File "/opt/local/src/magpie/magpie/cli/sync_resources.py", line 51, in merge_local_and_remote_resources
    remote_resources = _query_remote_resources_in_database(service_id, session=session)
  File "/opt/local/src/magpie/magpie/cli/sync_resources.py", line 244, in _query_remote_resources_in_database
    tree = get_resource_children(main_resource, session, models.REMOTE_RESOURCE_TREE_SERVICE)
  File "/opt/local/src/magpie/magpie/api/management/resource/resource_utils.py", line 196, in get_resource_children
    tree_struct_dict = tree_service_builder.build_subtree_strut(query)
  File "/usr/local/lib/python3.7/site-packages/ziggurat_foundations/models/services/resource_tree.py", line 76, in build_subtree_strut
    return self.service.build_subtree_strut(result=result, *args, **kwargs)
  File "/usr/local/lib/python3.7/site-packages/ziggurat_foundations/models/services/resource_tree_postgres.py", line 152, in build_subtree_strut
    new_elem = {"node": node.Resource, "children": OrderedDict()}
AttributeError: 'result' object has no attribute 'Resource'
[2021-03-25 15:02:18,839] INFO       [ThreadPoolExecutor-0_0][magpie.utils] Request: [POST lvupavics.ouranos.ca:443 /magpie/ui/error]

Note that the Jenkins test suite actually passed on that instance after the upgrade: https://daccs-jenkins.crim.ca/job/PAVICS-e2e-workflow-tests/job/master/239/console

Also when the magpie container was recreated at the moment of the upgrade, there was no error related to DB migration.

To Reproduce
Steps to reproduce the behavior:
My dev server autodeployed the PR bird-house/birdhouse-deploy#107

Expected behavior
Being able to see the permissions for Thredds for the anonymous group.

[fmigneault]

@tlvu
Can you create an admin user (or provide this instance's admin user credentials)?
I cannot view definitions from it without admin access.

[tlvu]

@tlvu
Can you create an admin user (or provide this instance's admin user credentials)?
I cannot view definitions from it without admin access.

Provided via private messaging.

[fmigneault]

@tlvu
Is it possible to obtain a data dump of the db by private message?
Using the same endpoint and config on a local instance doesn't cause the problem.
Must be tightly related to the 'remote' resources for which actual data must exist to do the sync.

[tlvu]

@tlvu
Is it possible to obtain a data dump of the db by private message?
Using the same endpoint and config on a local instance doesn't cause the problem.
Must be tightly related to the 'remote' resources for which actual data must exist to do the sync.

Provided tarball of /data/magpie_persist via private messaging.

[fmigneault]

@tlvu
My bad. I meant a DB snapshot before migration with magpie data similar to pavics/medus.
The error log is very specific to 'remote' resources, which depends on sync with thredds directories of actual data.

[tlvu]

@tlvu
My bad. I meant a DB snapshot before migration with magpie data similar to pavics/medus.
The error log is very specific to 'remote' resources, which depends on sync with thredds directories of actual data.

@fmigneault Sorry I do not have the state before the migration since the new broken Magpie was autodeployed directly on my dev VM the moment that PR was merged. Normally, PR is merged only after throughout testing so that has never been an issue before.