LFS258 - Lab 9.2. Configure a NodePort

cafaray

Hello folks, I'm trying to complete this lab 9-2, but I realize that the pod doesn't have access to Internet (in the the control plane and workers, I already have access), also I see that I can not egress packages from inside the pods to the Internet, for example a simple nslookup google.com doesn't resolves.

Any idea about what is happening or am I missing something?

I'm working with VSI instances in IBMCloud with free in/out traffic.

Kind regards

chrispokorni

Hi @cafaray,

I would recommend checking the /etc/resolv.conf file of the IBM Cloud VS instance. If needed you could add (append) an extra nameserver as nameserver 8.8.8.8 in the resolv.conf file to fix the resolve issues.

Regards,
-Chris

cafaray

Thank you @chrispokorni , I did it, but still not working.

I guess the problem is with calico driver, I'm getting this error describing pod calico-node-*****

Readiness probe failed: calico/node is not ready: felix is not ready: Get "http://localhost:9099/readiness": dial tcp 127.0.0.1:9099: connect: connection refused
  Warning  Unhealthy  13m (x2 over 13m)  kubelet            Readiness probe failed: calico/node is not ready: BIRD is not ready: Error querying BIRD: unable to connect to BIRDv4 socket: dial unix /var/run/calico/bird.ctl: connect: connection refused
  Warning  Unhealthy  13m                kubelet            Readiness probe failed: 2022-05-01 09:03:47.003 [INFO][245] confd/health.go 180: Number of node(s) with BGP peering established = 0
calico/node is not ready: BIRD is not ready: BGP not established with 10.243.64.21

I've installed a couple of times in clean virtual machines, and I'm getting the same result. Maybe the labs 3.1, 3.2, and 3.3 are missing something, or obvious some steps that I'm not doing. My first sight was when the Lab 3.2 mentioned that calico-node must run two pods, but mine it's running just one.

Appreciate any help on this.
Kind regards

chrispokorni

Hi @cafaray,

The connection refused error seems to indicate that a firewall may be blocking traffic to a port required by a calico component. I would recommend to configure the VPC firewall to allow all ingress traffic, from all sources, all protocols, to all ports.

The Overview section of Lab Exercise 3.1 highlights networking considerations when provisioning the infrastructure. I would recommend revisiting this section for clarification.

Regards,
-Chris

cafaray

Hi @chrispokorni

I'm using this configuration for my VPC, also as I've mentioned at first time, from the VSI I have free IN/OUT traffic with no restrictitons (I'm attaching some screenshots of my VPC configuration).

(Sorry for the language in the Screenshots, I forgot change my local configuration. )

Here the VPC and networks: my VSIs are using second one.

Here the rules in the network

The problem is in the pod/container level, for any reason it's no communicating with the erternal world.

I'm doing many hours to solve this issue, I'm sure the problem is something stupid, but I'm not capable to find what I'm doing wrong.

Kind regards.

chrispokorni

Hi @cafaray,

As long as your VPC uses the firewall rule that allows all inbound traffic from all IP sources, all protocols (no restrictions of any kind to TCP or other protocols), to all port numbers (no restrictions to only allow ports 22, 80, 443, etc...), your containers should have access to the outside world.

Were any connectivity issues observed in earlier labs? External access to services is introduced as early as Lab 3.5...

Regards,
-Chris

cafaray

Hi,

Right, I can not complete Lab 3.5, but I have not attached importance to the situation, I considered that it was not necessary for the subsequent ones.

Regards.