We need to have some automated tests that cover the combined use of durability and clustering.
This has been accomplished over some time. Currently there are both c++ and python client tests for persistent clusters. These tests are the identical tests run by qpid clustering, but by defining a store environment variable, these tests load the store and switch the queue and message declarations to persistent. Several of the python tests are store-only tests (which mean that when no store is loaded, they are ignored).
This issue tracks a long-term development task. Should it remain open?
I would close it.