Storage engine migration is recruiting two storage servers on same processes

The current migration logic intends to drop / add each storage engine one at a time without both old and new engines overlapping in lifetime, but testing shows this not to be the case.

This is particularly problematic since Redwood and SQLite do not share the same cache budget. While they could, in general it is hard to enforce that all storage engines share the same cache, for example RocksDB would also not share a cache budget with SQLite.

No file, test, or entry point is named. Start by tracing the storage engine migration logic and reproducing the observed overlap between old and new engines; done means migration no longer keeps both engines alive concurrently and avoids conflicting cache budgets.

Written by the indexing model from the issue text.