Improve session testing by using more realistic workflows

A follow up to #617. That issue was caused by using synthesized test session data instead of real test session data. We could perhaps fix this by flowing the data from ASFQuart or LDAP. We could also create an actual test service account, though that potentially carries risks even though the password would not be needed.

We should also try to solve the `read` and `write` asymmetry in ASFQuart.

Start by reading issue #617 and the existing session-testing workflow, then compare the possible data sources in ASFQuart and LDAP. Evaluate the risks of an actual test service account and investigate the reported read/write asymmetry in ASFQuart; done means the tests use realistic session data and the asymmetry is addressed.

Written by the indexing model from the issue text.