Mu-authorization pushed data in the default graph http://mu.semte.ch/application on top of the specified graphs in the user_groups configuration. This leads to duplication of data.
There is a GraphCleanup specification that is designed to eliminate this. Also tried with different versions of the mu-authorization to no avail (version beta0.0.5, beta0.0.7, latest).
When removing a record in the Ember app, the resource sends the correct DELETE query to the authorization, but it is transformed into a SELECT query (presumably to check if any items actually need to be removed, before removing) on the wrong graph. The graph it refers to is http://mu.semte.ch/graphs/empty which can be found in the source code of the mu-authorization service. This problem should only appear when no active groups are found. How is this possible when data insertion works somewhat correctly (see previous paragraph for more problems)?
After every request, the following messages appear in the logs, sometimes immediately, but most after a few seconds:
identifier_1 |
identifier_1 | 09:06:56.715 [error] GenServer #PID<0.1360.0> terminating
identifier_1 | ** (stop) "Mint transport error"
identifier_1 | Last message: {:tcp_closed, #Port<0.435>}
identifier_1 | State: #PID<0.1360.0>
dispatcher_1 |
dispatcher_1 | 09:07:11.776 [error] GenServer #PID<0.1399.0> terminating
dispatcher_1 | ** (stop) "Mint transport error"
dispatcher_1 | Last message: {:tcp_closed, #Port<0.396>}
dispatcher_1 | State: #PID<0.1399.0>
Are these from the livereloading of the Ember app that times out? All requests have succeeded an properly closed, as far as I can tell from the Network monitoring in the browser tools (yes even the favicon.ico has been made to succeed properly).