I have a few remote virtual machines, on which I want to deploy some Mongodb instances and then make them accessible remotely, but for some reason I can't seem to make this work.

These are the steps I took:

• I started a Kubernetes pod running Mongodb on a remote virtual machine.
• Then I exposed it through a Kubernetes NodePort service.
• Then I tried to connect to the Mongodb instance from my laptop, but it didn't work.

Here is the command I used to try to connect:

$ mongo host:NodePort   

(by "host" I mean the Kubernetes master).

And here is its output:

MongoDB shell version v4.0.3
connecting to: mongodb://host:NodePort/test
2018-10-24T21:43:41.462+0200 E QUERY    [js] Error: couldn't connect to server host:NodePort, connection attempt failed: SocketException:
Error connecting to host:NodePort :: caused by :: Connection refused :
connect@src/mongo/shell/mongo.js:257:13
@(connect):1:6
exception: connect failed

From the Kubernetes master, I made sure that the Mongodb pod was running. Then I ran a shell in the container and checked that the Mongodb server was working properly. Moreover, I had previously granted remote access to the Mongodb server, by specifying the "--bind-ip=0.0.0.0" option in its yaml description. To make sure that this option had been applied, I ran this command inside the Mongodb instance, from the same shell:

db._adminCommand( {getCmdLineOpts: 1}

And here is the output:

{
"argv" : [
    "mongod",
    "--bind_ip",
    "0.0.0.0"
],
"parsed" : {
    "net" : {
        "bindIp" : "0.0.0.0"
    }
},
"ok" : 1
}

So the Mongodb server should actually be accessible remotely.

I can't figure out whether the problem is caused by Kubernetes or by Mongodb.

As a test, I followed exactly the same steps by using MySQL instead, and that worked (that is, I ran a MySQL pod and exposed it with a Kubernetes service, to make it accessible remotely, and then I successfully connected to it from my laptop). This would lead me to think that the culprit is Mongodb here, but I'm not sure. Maybe I'm just making a silly mistake somewhere.

Could someone help me shed some light on this? Or tell me how to debug this problem?

EDIT:

Here is the output of the kubectl describe deployment <mongo-deployment> command, as per your request:

Name:                   mongo-remote
Namespace:              default
CreationTimestamp:      Thu, 25 Oct 2018 06:31:24 +0000
Labels:                 name=mongo-remote
Annotations:            deployment.kubernetes.io/revision=1
Selector:               name=mongo-remote
Replicas:               1 desired | 1 updated | 1 total | 1 available | 0 unavailable
StrategyType:           RollingUpdate
MinReadySeconds:        0
RollingUpdateStrategy:  1 max unavailable, 1 max surge
Pod Template:
  Labels:  name=mongo-remote
  Containers:
   mongocontainer:
    Image:      mongo:4.0.3
    Port:       5000/TCP
    Host Port:  0/TCP
    Command:
      mongod
      --bind_ip
      0.0.0.0
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Conditions:
  Type           Status  Reason
  ----           ------  ------
  Available      True    MinimumReplicasAvailable
OldReplicaSets:  <none>
NewReplicaSet:   mongo-remote-655478448b (1/1 replicas created)
Events:
  Type    Reason             Age   From                   Message
  ----    ------             ----  ----                   -------
  Normal  ScalingReplicaSet  15m   deployment-controller  Scaled up replica set mongo-remote-655478448b to 1

For the sake of completeness, here is the yaml description of the deployment:

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: mongo-remote
spec:
  replicas: 1
  template:
    metadata:
      labels:
        name: mongo-remote
    spec:
      containers:
        - name: mongocontainer
          image: mongo:4.0.3
          imagePullPolicy: Always
          command:
          - "mongod"
          - "--bind_ip"
          - "0.0.0.0"
          ports:
          - containerPort: 5000
            name: mongocontainer
      nodeSelector:
        kubernetes.io/hostname: xxx