Tolerate range failure when discovering redis servers

[worr]

Currently, when a seco.range.RangeException is raised, we immediately call sys.exit. Putting aside the fact that this is bad behavior in a library, we don't try and test other range servers.

Now, we continue the loop and only abort when we have exhausted the range servers available to us. It's also necessary to remove setting self._redis_corelist to None, as this causes the library to abort later.