K3s install fails to deploy web container

[engineertdog]

I have the issue described in https://github.com/ctrliq/ascender-install/blob/main/docs/issues/ascender_api_waiting.md but the cause is not DNS. I used the config file https://github.com/ctrliq/ascender-install/blob/main/docs/installation/k3s/k3s/k3s.basic.config.yml and disabled the usage of /etc/hosts, and replaced the node names of course.

However, unlike the guide where it only points to DNS being the issue, I only have two pods running. The API service /api/v2/ping is reachable in a sense, but accessing the endpoint returns a 404.

sudo kubectl get pod -n ascender
NAME                                               READY   STATUS    RESTARTS      AGE
ascender-app-postgres-15-0                         1/1     Running   2 (81m ago)   142m
awx-operator-controller-manager-7d455fb4b5-2thq4   2/2     Running   4 (81m ago)   143m

One other thing to note, the installation docs say to run config_vars.sh, but that's not possible without first running setup.sh after setting the install type to k3s. https://github.com/ctrliq/ascender-install/blob/main/README.md#using-the-installer

[cigamit]

They only time I have really seen this is when I don't have enough resources for the other containers to start. How much CPU/Ram does that server have?

You can usually follow the logs of the operator and it will give you a hint of where it errors (the operator is just running a playbook on a loop)

[engineertdog]

I gave the Ascender server 8 cores and 10 GB of RAM, and I gave the Ledger server 8 cores and 8 GB of RAM.

SS from Ascender:

The controller logs is a wall of text, I can look through it, but may take some time.

One thing to note, after it completes this step, it says the Web Deployment is ready, but there's no pod.

[cigamit]

Do a
kubectl get all -n ascender
So we can see the deployments and everything else.

[engineertdog]

Sure thing, here's everything that's deployed to the NS.

[cigamit]

I finally saw this happen on another install. Check your custom.config.yml and ensure that you don't have a blank

  ee_images:
    - name: ''
      image: ''

settings in there. As the operator pukes on it for some reason. I need to figure out why the config vars is adding it as blank also.