In the schema-disagg-abort test all schema operations (create and drop) issued their corresponding publish calls before the step_down_timestamp is set.

It looks like a leftover from the early days of schema epoch implementation. There is no reason to issue all outstanding publish events before step-down starts.

TODO:

1. Investigate that publish (for create and drop) can be called after step_down_timestamp is set. Establish correct timestamp for this publish.
2. Do not flush pending publish events before step-down starts. Let the publish events flow naturally across step-down.