Task: review how we recover from failures while updating/saving root

[MVrachev]

What is the task about?

We need to make sure that we store our md files in a particular order, so that in case of a failure, we can easily recover.
Because of that, the order we need to follow regarding how we update and persist new metadata is the following:

1. bins/custom delegated target roles
2. targets
3. snapshot
4. timestamp
5. root

It's great we make sure that we solve the order from 1 to 4 with the implementation of _run_online_roles_bump() (see my comment here #501 (comment)), but the question is do we have the same guarantee for root?
That's something it's worth a review and if needed to make changes

References

Continuation of #494

Code of Conduct

• I agree to follow this project's Code of Conduct