连接到在不同 pod 上运行的 gRPC 服务器

Question

I have a gRPC client and server running on my local and I'm able to communicate over them via HTTP2 through localhost. But when the apps are deployed in dev environment using Kubernetes and Docker, I'm not able to call the gRPC server from client app.

I connect to the client using this code:

ManagedChannel managedChannel = ManagedChannelBuilder.forAddress(host, port).usePlaintext().build();

In DEV: Here is my application.yml from server app.

grpc:
  port: 6565
  enabled: true

I have the gRPC server running on port 6565, so I have changed the following on client application.yml as well

taxgRPC:
  host: bpstax-ws.g-dev4.svc.rnq.k8s.copart.com
  port: 6565  

When my client calls the server, the server doesn't get any incoming requests and client fails giving the following exception:

Caused by: io.grpc.netty.shaded.io.netty.channel.AbstractChannel$AnnotatedConnectException: Connection refused: no further information: bpstax-ws.g-dev4.svc.rnq.k8s.copart.com/10.148.2.194:6565

When I log into the particular dev pod, I can see the port listening to the gRPC server.

[dedas@rndcks401 ~]$ kubectl exec -ti bpstax-g-dev4-v029-74mj2 -n g-dev4 -- bash
bash-4.1$ netstat -tlnp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address               Foreign Address             State       PID/Program name
tcp        0      0 0.0.0.0:6565                0.0.0.0:*                   LISTEN      1/java
tcp        0      0 0.0.0.0:38957               0.0.0.0:*                   LISTEN      1/java
tcp        0      0 0.0.0.0:8080                0.0.0.0:*                   LISTEN      1/java
tcp        0      0 0.0.0.0:8081                0.0.0.0:*                   LISTEN      1/java

I need some help on how to connect to this gRPC server. Please let me know if any clarifications/code is required.

Answer 1

I guess you have to provide a correct host to your gRPC client. Within your k8s env the best way to discover you server pod's ip address is though headless service .

The example definition:

apiVersion: v1
kind: Service
metadata:
  labels:
    app: server
  name: server-headless
  namespace: <YOUR_NAMESPACE>
spec:
  clusterIP: None
  ports:
  - name: tcp
    port: 6565
    protocol: TCP
    targetPort: 6565
  selector:
    app: <SERVER_POD_LABEL>
  type: ClusterIP

Having the headless service for your server app you can use service's FQDN as a host. This way you can be sure the service's FQDN is always resolved to the correct pod's ip/s.