Python script won't load on startup

Question

I have tried like 5 methods to get this script loaded after raspi3 load.

#!/usr/bin/python3.7
#sudo python /home/pi/masterBaseInterface.py


import socket               # Import socket module
import thread
import time
import serial


def get_ip_address():
  ip_address = '';
  s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
  s.connect(("8.8.8.8",80))
  ip_address = s.getsockname()[0]
  s.close()
  return ip_address

def on_new_client(clientsocket,addr):
  while True:
    msg = clientsocket.recv(2048)
    if msg != "":
      if msg == "findingNemo":
        clientsocket.send('heyDad') 
      else:
        #msg = "5A00000d0a71"

        ser = serial.Serial(
          port='/dev/ttyUSB0',
          baudrate=115200,
          parity=serial.PARITY_NONE,
          stopbits=serial.STOPBITS_ONE,
          bytesize=serial.EIGHTBITS,
          timeout=2
        )

        print "Received message from client: "+msg
        msg = msg.replace('\n', ' ').replace('\r', '').replace(' ', '')      
        msg.strip() 

        print "Dispatching message to master base"
        ser.write(msg.decode('hex'))
        print "Message sent, waiting for response"
        #time.sleep(2)
        #output = ser.readline()
        output=""
        time.sleep(2)
        while ser.inWaiting() > 0:
          output += ser.read(10)

        if len(output)>0:
          output = output.replace('\n', ' ').replace('\r', '').replace(' ', '')
          print "Master base replies with "+output.encode('hex')
          clientsocket.send('OK,'+output.encode('hex'))  
          clientsocket.close()
          print "Message sent to client, connection closed"
          break
        else: 
          print "Master base not responding, sending error message to client"
          clientsocket.send('ER,noResponse') 
          clientsocket.close()
          print "Message sent to client, connection closed"
          break

        msg=""

      #clientsocket.send(output.encode('hex'))        
    #clientsocket.close()

s = socket.socket()         # Create a socket object
host = get_ip_address() # Get local machine name
port = 1000                # Reserve a port for your service.

print 'Server started with local IP '+ get_ip_address()
print 'Ready to serve clients'

s.bind((host, port))        # Bind to the port
s.listen(5)                 # Now wait for client connection.


while True:
   c, addr = s.accept()     # Establish connection with client.
   print 'Got connection from', addr
   thread.start_new_thread(on_new_client,(c,addr))
   #Note it's (addr,) not (addr) because second parameter is a tuple
   #Edit: (c,addr)
   #that's how you pass arguments to functions when creating new threads using thread module.
s.close()

Posting the script source because I don't know if it contains something illegal...

What I have tried? This crontab

@reboot sudo python /home/pi/masterBaseInterface.py

Also the rc.local thing

sudo nano /etc/rc.local
sudo python /home/pi/masterBaseInterface.py &

The bashrc too, but this ones only works when I login via ssh with putty

sudo nano /home/pi/.bashrc
echo Running at boot 
sudo python /home/pi/masterBaseInterface.py

I tried other things, but can't remmeber now. So, what's happening here?

Answer 1

If there is any error while running this on startup, you can’t see the error message. To check this, press ctrl+alt+F1. Here you can check the output of background processes. You can return to graphical display by ctrl+alt+F7. This is the way you can ensure whether there is any error while running the code or not.

Now come to your solution, if there is any error, I can check that #!/usr/bin/python3.7 is written in the first line which indicates python3 but you are writing python in rc.local which means python2.

You should try using python3 instead.

Answer 2

According to this site you may need to change your first line from this:

#!/usr/bin/python3.7

to this:

#!/usr/bin/env python3

The first line in my working script is this:

#!/usr/bin/env python

ETA: OK, you've tried this and it's still not working. My working script is launching from rc.local, and I've caught another difference. You've put this in your rc.local:

sudo python /home/pi/masterBaseInterface.py &

OTOH, my rc.local has this (equivalent):

/home/pi/masterBaseInterface.py &

You could try removing "sudo" and see if it works (I think rc.local is effectively running as root anyway). If it doesn't, you could try removing "python" as well, but you would need to make sure you have marked the script executable (chmod +x).

Answer 3

My suggestions are as follows:

Since you are running the script as a cron job, the PATH environment variable is not set, so the Python interpreter cannot be found and the script fails to run.

As your script already has #!/usr/bin/python3.7 on the first line, and assuming the path is indeed correct on your platform, then there is no need to prefix the commmand with python. You could just call your script directly like this:

/home/pi/masterBaseInterface.py

(Without python at the front) ... provided it has the executable flag set ie chmod +x /home/pi/masterBaseInterface.py

or add the full path to Python in your cron like this: @reboot sudo /usr/bin/python3.7 /home/pi/masterBaseInterface.py

(again make sure that the path is correct, you can run which python3 or possibly which python3.7 to verify.

Second suggestion would be to run the program in a screen session and log the console output to a file. If the execution fails you can see what actually happened in the console. One example here. An added bonus is that you can 'join' the console through SSH later and catch on the program while it's already running in its own screen session...

So I would always recommend using screen, especially for running unattended scripts. It is also convenient because you can attach and detach from sessions anytime, recover where you left after losing the SSH connection etc.

Third suggestion would be to add some logging and exception handling to your program, so that if it crashes while running in the background, you have details about the error in a log file. It helps a lot with debugging. Obviously, this will work only when Python is actually able to execute the script. But any program is always going to crash at some point for some reason. It's good to have a trace.