Python not able to connect to grpc channel -> "failed to connect to all addresses" "grpc_status":14

Question

I'm getting the error below when trying to call a stub method. Any idea what is causing it?

[bolt.api.handlers] 2019-08-21 20:07:57,792 ERROR handlers:1066: 'ResourceHandler' object has no attribute 'ontology_service_handler'
Traceback (most recent call last):
  File "/bolt-webserver/bolt/api/onse/onse_handlers/ontology_service.py", line 17, in post
    ontology_id = await self.onse_stub.createOntology()
  File "/bolt-webserver/bolt/api/onse/onse_stub.py", line 41, in createOntology
    return self.stub.CreateOntology(ontology_messages_pb2.Ontology())
  File "/usr/local/lib/python3.6/site-packages/grpc/_channel.py", line 565, in __call__
    return _end_unary_response_blocking(state, call, False, None)
  File "/usr/local/lib/python3.6/site-packages/grpc/_channel.py", line 467, in _end_unary_response_blocking
    raise _Rendezvous(state, None, None, deadline)
grpc._channel._Rendezvous: <_Rendezvous of RPC that terminated with:
        status = StatusCode.UNAVAILABLE
        details = "failed to connect to all addresses"
        debug_error_string = "{"created":"@1566418077.791002345","description":"Failed to pick subchannel","file":"src/core/ext/filters/client_channel/client_channel.cc","file_line":3818,"referenced_errors":[{"created":"@1566418077.790965749","description":"failed to connect to all addresses","file":"src/core/ext/filters/client_channel/lb_policy/pick_first/pick_first.cc","file_line":395,"grpc_status":14}]}"

I've tried to provide ip address instead of hostname but still getting the same error . The OnseStub class is initialized right before calling createOntology method. The service is up and running. The failing call is done from a tornado web app (in case that might matter)

class OnseStub:

    def __init__(self, ontology_service_backend):
    self.channel = grpc.insecure_channel('localhost:51051')
    self.stub = ontology_service_pb2_grpc.OntologyServiceStub(self.channel)

    def __del__(self):
    if self.channel != None:
    self.channel.close() # close grpc channel

    async def createOntology(self):
    return self.stub.CreateOntology(ontology_messages_pb2.Ontology())

Powerhouse

Answer 1

What fixed it for me is adding the following option to client channel.

grpc.insecure_channel('localhost:50051', options=(('grpc.enable_http_proxy', 0),))

This is missing in the grpc quickstarts as well and is not really highlighted.

Arv

Instead of forcing insecure connection, you need to configure properly your CA to trust the grpc server certificate – Reject

Answer 2

This is common error, it can occured in different cases. But in most usual case decribed in https://github.com/grpc/grpc/issues/9987 can be fixed by unset http_proxy enviroment variable

if os.environ.get('https_proxy'):
 del os.environ['https_proxy']
if os.environ.get('http_proxy'):
 del os.environ['http_proxy']

Mundell

Answer 3

If you are using unix domain sockets, then it is possible you have stale socket files. Be sure to remove old files before making grpc calls.

Studied