Don't use a timeout in shutdown sequence

[palango]

Problem Definition

Currently a timeout is used in the shutdown sequence if the ethereum node in not reachable.

Solution

Instead of using a client that tries to reconnect and giving it a timeout we should use a client that doesn't try to reconnect.

[palango]

@hackaugusto The original cause of this issue is gone, but there is still a timeout:

raiden/raiden/raiden_service.py

Lines 337 to 341 in d68f002

	try:
	with gevent.Timeout(self.shutdown_timeout):
	self.blockchain_events.uninstall_all_event_listeners()
	except (gevent.timeout.Timeout, RaidenShuttingDown):
	pass

[andrevmatos]

@palango I should be removing that one in #2208, then the only stop/shutdown timeout left should be the one in cli's _run_app

[palango]

Nice, then just add this as a fixed issue there.

[andrevmatos]

Fixed by #2208