error handling in read_file seems incorrect in file.py

in LargeHTTPFile line 137 there is an iteration of all paths of the file.
but in _request an error is always raised if is_success returns False.

so the whole fallback to backup_dests doesnt seem to work, because the it will only ever try the first item in the list, and that either succeeds or fails with raising an error exiting the loop.

the call in line 146 to is_succes can never return false.
