-2

I have a local Kubernetes Running on Windows using Docker. In My Windows Host I am running a Process on Port 4000.

Inside my POD I am doing curl windows_host_ip:4000 But not able to get the response. Connection from POD is being established but the response is giving timeout. Logs

  • Hostname was NOT found in DNS cache
  • Trying 192.168.18.10...
  • connect to 192.168.18.10 port 4000 failed: Connection refused
  • Failed to connect to 192.168.18.10 port 4000: Connection refused
  • Closing connection 0 curl: (7) Failed to connect to 192.168.18.10 port 4000: Connection refused root@ordermanagement-64694dd8b8-2ktm8:/apps/ordermanagement# curl -v http://192.168.18.10:4000/ordermanagement/order/orders
  • Hostname was NOT found in DNS cache
  • Trying 192.168.18.10...
  • Connected to 192.168.18.10 (192.168.18.10) port 4000 (#0)

GET /ordermanagement/order/orders HTTP/1.1 User-Agent: curl/7.38.0 Host: 192.168.18.10:4000 Accept: /

  • Recv failure: Connection reset by peer
  • Closing connection 0 curl: (56) Recv failure: Connection reset by peer

Please let me know how can I allow incoming traffic to my POD from my Host System

Bhupendra Nath
  • 321
  • 1
  • 2
  • 10
  • 1
    Does this answer your question? [How to access host's localhost from inside kubernetes cluster](https://stackoverflow.com/questions/65123401/how-to-access-hosts-localhost-from-inside-kubernetes-cluster) – Phil Oct 19 '21 at 06:01
  • Looks like you trying to access external IP from internal container?.. Kubernetes has own network and own routing inside it. Simply → external network is not accessible from internal kubernetes's network. But you can try to access localhost of your windows machine using `host.docker.internal` host. Don't know about kubernetes, but it works in [Docker](https://stackoverflow.com/a/67272342/13946204) – rzlvmp Oct 19 '21 at 06:21
  • 1
    @rzlvmp no, it does not work in k8s – Phil Oct 19 '21 at 09:04
  • Below is the trace log of curl --trace trace.log http://kafka-service.default:4000/ordermanagement/order/orders == Info: Hostname was NOT found in DNS cache == Info: Trying 192.168.18.10... == Info: Connected to kafka-service.default (192.168.18.10) port 4000 (#0) => Send header, 118 bytes (0x76) == Info: Recv failure: Connection reset by peer == Info: Closing connection 0 Connection from POD to Host is being established correctly issue is coming as POD is not able to get the response back. – Bhupendra Nath Oct 20 '21 at 08:18

1 Answers1

0

This is the expected behavior. A pod is not supposed to have access to the host's network, process, filesystem, etc. directly otherwise anyone breaking into a pod can get the entire information of not only the host system but also about other pods running in the host.

You could look into creating a Kubernetes endpoint for the host process inside the cluster, please see this documentation here for manually managing services endpoints.

You can also potentially achieve this by running your pod as Privileged pod, however, it is not a good security practice to run privileged pods.

Privileged - determines if any container in a pod can enable privileged mode. By default, a container is not allowed to access any devices on the host, but a "privileged" container is given access to all devices on the host. This allows the container nearly all the same access as processes running on the host. This is useful for containers that want to use Linux capabilities like manipulating the network stack and accessing devices.

You can turn a pod into a privileged one by setting the privileged flag to true (by default a container is not allowed to access any devices on the host).

Manish Joshi
  • 3,550
  • 2
  • 21
  • 29