Reputation: 141
When in a multithreaded application i call rpyc.Connection.serve_all()
from a thread other threads are unable to immediately use the connection.
I think that serve_all
is blocking the connection and the other threads are able to access it only when there is some timeout.
This code should reproduce the issue
Server:
#!/usr/bin/env python3
import rpyc
import rpyc.utils.server
import threading
class Service(rpyc.Service):
def on_connect(self):
print("New connection")
def on_disconnect(self):
print("Connection closed")
def exposed_get_status(self):
return "Test string"
server = rpyc.utils.server.ThreadedServer(Service, port = 12345)
t = threading.Thread(target = server.start)
t.daemon = True
t.start()
t.join()
Client:
#!/usr/bin/env python3
import rpyc
import threading
import datetime
con = rpyc.connect('localhost',12345)
def delayed():
print("Time is up")
now = datetime.datetime.now()
print(con.root.get_status())
print(str(datetime.datetime.now() - now))
timer = threading.Timer(10,delayed)
print("Starting timer")
timer.start()
print("serving all")
con.serve_all()
Sample output (from the client):
$ python3 testrpyc.py
Starting timer
serving all
Time is up
Test string
0:01:30.064087
I'm using RPyC 3.4.3 on Python 3.5.4rc1 (debian sid) installed with pip.
I think I'm misusing serve_all, but I can't find anything in the docs.
Upvotes: 1
Views: 1079
Reputation: 141
(Answering myself)
I opened a issue on github and it seems that this is normal. The solution is to use perform IO only on a single thread for any resource.
Upvotes: 0