salaxander / kubernetes-the-hard-way Goto Github PK

Should be:
EXTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n controller-0 --output tsv | cut -f20)

After creating the SSH config file, I cannot SSH into the machine because of wrong permissions (working inside WSL in my case):

tomica@N1:~/.ssh$ ssh controller-0
Bad owner or permissions on /home/tomica/.ssh/config

Fix:
chmod 644 ~/.ssh/config

You can add a hint for less experienced users to not to verify the deployment or run az commands from anywhere but their "control" host (WSL in my case). Sometimes it's clear, sometimes it's not (chapter 8 - Verification and beyond).

Should be:
EXTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n ${INSTANCE_NAME} --output tsv | cut -f20)

It should be cut -f20 and cut -f17 (otherwise you'll get the wrong fields):
EXTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n ${instance} --output tsv | cut -f20)
INTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n ${instance} --output tsv | cut -f17)

Missing the EOF in The Controller Manager Client Certificate:
cat > kube-controller-manager-csr.json <<EOF
{
"CN": "system:kube-controller-manager",
"key": {
"algo": "rsa",
"size": 2048
},
"names": [
{
"C": "US",
"L": "Portland",
"O": "system:kube-controller-manager",
"OU": "Kubernetes The Hard Way",
"ST": "Oregon"
}
]
}
EOF

In the smoke test chapter, we expose an nginx service, and try to curl the node port on a worker node using

curl -I http://${EXTERNAL_IP}:${NODE_PORT}

This is failing for me still. I'm thinking it may be something with the security group rule as I'm getting a connection refused.
Any help troubleshooting this would be appreciated!

There is an error in the line where you're cutting out the instance's public IP address:
az vm show --show-details -g kubernetes-the-hard-way -n ${instance} --output tsv | cut -f19

Should be:
az vm show --show-details -g kubernetes-the-hard-way -n ${instance} --output tsv | cut -f20

Since the Azure CLI can't SSH to nodes without a public IP like the Google Cloud CLI, this tutorial contains a lot of lines like this:

EXTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n controller-0 --output tsv | cut -f19)
ssh azureuser@${EXTERNAL_IP}

Let's add a step to build an SSH config so that all the following steps become more like:

ssh azureuser@worker-0

Issue

The default limit of vCPUs in a region is 10. This chapter requires 12 to be created. This causes the creation of worker-2 to fail.

If the user submits the request to create this worker within the iterated series. The request will fail silently. If the user submits the request to create the worker independently, they will receive the following error:

Azure Error: InvalidTemplateDeployment
Message: The template deployment 'vm_deploy_5rAXEbqdNbfTEgfZ0S4u5i6R2SMKCvJO' is not valid according to the validation procedure. The tracking id is 'XXXXXXXXXXXXXXXXXXXX'. See inner errors for details. Please see https://aka.ms/arm-deploy for usage details.
Exception Details:
        Error Code: QuotaExceeded
        Message: Operation results in exceeding quota limits of Core. Maximum allowed: 10, Current in use: 10, Additional requested: 2. Please read more about quota increase at https://aka.ms/ProdportalCRP/?#create/Microsoft.Support/Parameters/{"subId":"XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX","pesId":"XXXXX","supportTopicId":"32447243"}.

Possible Solution

Increasing the limit is not exceptionally difficult, but does require submitting a support ticket and knowing how to interpret the error above correctly. It may be helpful to have a note in the section on creating workers that describes this potential issue and how to fix it.

Should be (EOF should be pulled out, in the same column with "done"; if not, shell waits for additional input):
for instance in controller-0 controller-1 controller-2 worker-0 worker-1 worker-2; do
EXTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n ${instance} --output tsv | cut -f20)
cat <<EOF | tee -a ~/.ssh/config
Host ${instance}
User azureuser
HostName ${EXTERNAL_IP}
IdentityFile ~/.ssh/id_rsa
ServerAliveInterval 120
EOF
done

If we want to be consistent, OU should be the same on every certificate:
"OU": "Kubernetes The Hard Way"

Should be:
EXTERNAL_IP=$(az vm show --show-details -g kubernetes-the-hard-way -n worker-0 --output tsv | cut -f20)

At the end of chapter 12 when we run:

kubectl exec -ti $POD_NAME -- nslookup kubernetes

The IP address resolves, but not name 'kubernetes'