[Bug]: Long running tasks: RabbitMQ-based framework tries to serialize exceptions that is shall not

[sanderegg]

Is there an existing issue for this?

• I have searched the existing issues

Which deploy/s?

production aws (e.g. osparc.io), staging aws (e.g. staging.osparc.io), production on-premise (dalco), staging on-premise (dalco), development (master), other (e.g. local)

Current Behavior

Found during debugging session with @GitHK from missing parts of MMUX service of @JavierGOrdonnez

Currently the new long running task framework introduced in https://github.com/ITISFoundation/private-issues/issues/74 tries to serialize exceptions that it was explicitely told to not serialize.

This creates confusion when debugging and misleads to consider issues from the framework when they are in the client code, or somewhere else.

Expected Behavior

When a client calls the long running tasks framework and tell it to transfer ExceptionA, the framework shall not try to serialize ExceptionB out of fun. This is not the intended behavior.
It shall only try to serialize ExceptionA.
If ExceptionB is raised in the server then that exception shall be transformed to a simple common exception that can be re-raise on the client side safely, not adding some weird serialization error around.

Steps To Reproduce

No response

Anything else?

No response

[sanderegg]

@GitHK can you check that one? is this still the case?

[GitHK]

Looks like so to me, still tracked by a parent issue