Fixed sleep time when starting hidra via control server

[schooft]

After starting hidra, the control server waits a hard-coded 10s until it acquires the configuration from the new hidra instance. However, in rare cases the startup can take longer than 10s and then an incomplete configuration is acquired. For example, the request_port might not be updated yet, resulting in the following error when trying to initialize a new Transfer object:

Failed to signal_socket start (connect, REQ):
'tcp://asap3-pXX.desy.de:random' Traceback (most recent call last):
   File "/usr/lib/python3/dist-packages/hidra/utils/utils_network.py",
line 651, in start_socket
     socket.connect(endpoint)
   File "zmq/backend/cython/socket.pyx", line 583, in
zmq.backend.cython.socket.Socket.connect
   File "zmq/backend/cython/checkrc.pxd", line 25, in
zmq.backend.cython.checkrc._check_rc
zmq.error.ZMQError: Invalid argument

The 10s sleep in hidra_control/server.py already has the comment

# TODO exchange that with proper communication to statserver