You learn remarkably little about a storage system while it is running normally. Normal operation is what it was designed for, tested against and demonstrated in. Everything interesting about it surfaces during a migration, because a migration is the one situation where two systems must simultaneously hold a coherent opinion about the same data.
That is also why cutover plans are where I have seen the most experienced people in a room get quiet.
The questions a cutover forces you to answer out loud
Who is authoritative right now, and how does every participant find out? Note that both, briefly is an answer, and it is the answer that causes incidents. If any component can be told late, you need to know exactly how late and what it is permitted to do in the meantime.
What happens to a write that arrives during the transition? Not the happy path. The one that arrives after the source stopped accepting and before the target started serving. Every system has that window. The only question is whether you have described it or discovered it.
How do you roll back once the first byte has landed on the target? If the answer is going to be worked out during the rollback, you do not have a plan, you have an intention.
How do you prove afterwards that nothing was lost? Not the tool reported success. An independent check — ideally one that would fail loudly if the tool had lied to you. Tools are written by people who were also under deadline.
Rehearse the rollback, not the migration
The migration gets rehearsed because the migration is the plan, and rehearsing the plan feels like preparation. The rollback is the thing you will actually be executing at three in the morning, in front of an audience, having already been awake for nineteen hours.
I have never regretted the hour spent walking through a rollback. I have regretted the opposite.
Why it is worth seeking this work out
Every one of those questions is a distributed systems question with a storage accent, and there is no way to answer them abstractly. Teams that have taken a real system through a real cutover reason about consistency differently afterwards — not because they read something, but because they were forced to say precisely what they meant by it, in writing, with a date attached and their name on it.
If you are early in your career and a migration comes up, volunteer. It is the fastest concentrated education in this field I know of, and it is unpopular, which means it is available.