(1) Servers I own that are physically located in my house.
(2) Servers I own that are co-located at a hosting facility.
(3) (physical) Servers that I am renting from a hosting facility.
(4) Virtual servers that I am renting from a hosting facility.
I assume that most law enforcement officials feel that 2-4 shouldn't need a warrant because they are not located in my house, so therefore I don't get protection from unreasonable search/seizure.
Also, at what point do #3 or #4 start to encroach onto Dropbox-esque territory where I own the data, but they are holding it for me under contract?
In the first two cases, you are talking about your own property. Unless you have a contract with the colocation facility that says otherwise, the presence of your machine at such a facility shouldn't somehow make it searchable, just like if I put a closed briefcase full of papers down in a public library and walk away from it for 5 minutes, the police cannot suddenly search that.
As I understand it, things are shakier in case (3) and (4), because your provider absent a contract with you that says otherwise could choose to comply with subpoenas (providers commonly claim that they'll respond to LEOs only if they have a "subpoena or a warrant", as if that was a feature; you should assume that subpoenas are mere formalities and that any LEO can obtain one practically on demand). It's possible that your remedy in such a case would be civil, which would be cold comfort if your goal was to have evidence excluded from a trial.
There are a bunch of lawyers on HN now, and I'm sure one of them will jump in here to correct me.
However, the commentary, and the bill (perhaps) assumes that email resides on a providers server - that the end user does not control it.
How does the bill address warrantless access (or access of any kind) to email hosted on the users own mail server ?