In helper_disagg.py, there is a function called reopen_disagg_conn, which is designed to use reconfig as a follower to disable shutdown checkpoints. In certain tests where this function is invoked, shutdown checkpoints are disabled, and the tests are specifically designed to rely on the last explicitly triggered checkpoint, assuming no shutdown checkpoint occurs.

This ticket aims to review whether this workflow is reasonable. The discussion is also related to context from WT-16571, where we plan to introduce a step-down checkpoint when detecting a dirty btree. However, this change has caused some test failures due to the aforementioned workflow.