Troubleshooting

Networking Issues

If you are having issues with networking, including but not limited to the following common cases, you need to verify all the steps in this section.

• The pods take a long time to start
• The pods don't have internet access
• The pods cannot ping each other

Node steps

Follow these steps on each node in your cluster.

• Verify the FORWARD policy is set to ACCEPT and that the br_netfilter module is loaded.

sudo iptables --policy FORWARD ACCEPT
sudo modprobe br_netfilter

• Verify that the IP forwarding is enabled.

sudo sysctl net.ipv4.ip_forward=1

• Verify that the kubelet service is running.

sudo systemctl status kubelet

• Flush the iptables rules.

sudo iptables --flush
sudo iptables --table nat --flush
sudo nft flush ruleset

Cluster-wide steps

Follow these steps once for the entire cluster from a kubectl client.

• Restart the Calico nodes.

kubectl rollout restart --namespace kube-system daemonset calico-node

• Restart the CoreDNS pods.

kubectl rollout restart --namespace kube-system deployment coredns

If still the issue persists, please contact BubbleRAN support.

Deployment Issues

Rarely the following issues might occur during the deployment.

• No pods appear -- Most likely the models are missing
• The pods stuck on ContainerCreating status -- Most likely the device are not available
• The pods stuck on CrashLoopBackOff status -- Most likely the configuration is wrong
• The pods stuck on ImagePullBackOff status -- Most likely you have not logged into the namespace

Solutions

1. Redeploy the network if possible.
2. Check you have logged into the namespace via the CLI.
3. Check for the networking issues.
4. Check via the cli extract infra, if the requested devices are detected and available.
5. Use the cli diag command to check whether the nodes are reachable and ready.

If the issue persists, please contact BubbleRAN support. You would receive a response within 24 hours.