Why does the code cloned from the official website not run normally after being applied to the cluster?

[wyhdil]

Hello, I have some questions regarding the source code from the official website that I would like to ask.

I cloned this version of the code from the official website. Then, I built the try-now-page, service, and operator modules into separate container images and deployed them to the cluster. And i tagged all these modules with the tag "0.8.0."

Then I discovered these two problems：

1.When I tried to deploy the operator using the theia-cloud-operator:0.8.0 image in the cluster, I encountered an error.


I would like to know if any modifications are required before directly packaging and deploying the operator's code as a container image in the cluster. Currently, it seems that we cannot directly apply the image to the cluster.

2.When I configured the operator to use the image version 0.8.1.OSWeek23-v1 (the default version used in Terraform) and the service to use the 0.8.0 image, I encountered a 473 error status code after the try-now-page module made a request to the service.


I would like to understand why the service is returning a 473 status code. Based on the code, it seems to be related to the App Definition. However, when I checked the cluster, I verified that the theia-cloud-demo App Definition does exist. So, at the moment, I'm unsure what is causing this error.

In short, I want the image built from the code to run normally when applied to the cluster. What kind of configuration is required?

Hope to get answers，thanks!

[sgraband]

As already described here mixing versions is never a good idea. As to the 0.8.0 version: Did you also use the helm chart that corresponds to that version? The value with which you apply the deployment also play a huge role.
Also could you paste the error log of the operator? This might contain significant hints on what is going on (Probably a wrongful parameter).
Lastly, i would always encourage to take the latest version of a framework, as it should be the most stable version and is the one that is currently supported.

[sgraband]

How are you installing Theia Cloud? Using terraform or via helm directly? If you are doing the first you also need to checkout the theia-cloud-helm repo at the correct tag.

I think the issue here is, that since you switched versions you need to cleanup old CRD versions. Please uninstall Theia Cloud and then run:
kubectl delete crd appdefinitions.theia.cloud
kubectl delete crd sessions.theia.cloud
kubectl delete crd workspaces.theia.cloud
on the cluster. Afterwards you can install it again and it should work.

[wyhdil]

I previously installed it using Terraform and have now deleted the CRDs you mentioned. But which version of Theia Cloud should I choose?

It appears that there is no exact matching version of 0.8.1.MS2 available here.

Regarding the manual installation of Theia Cloud, should I follow the instructions provided here?

[wyhdil]

Because helm install always fails, I manually downloaded the files. However, it seems that there are still some issues.

Here it seems that I am unable to use the name "theia-cloud-crds", so I changed it.

tptuser@linux-ide:~/Work/infra/lc/theiacloud-0.8.1.MS2/theia-cloud$ helm install theia-cloud-crds-0.8.1 theia-cloud-crds-0.8.1-v004-MS2.tgz --namespace theiacloud
NAME: theia-cloud-crds-0.8.1
LAST DEPLOYED: Tue Nov  7 13:46:15 2023
NAMESPACE: theiacloud
STATUS: deployed
REVISION: 1
TEST SUITE: None

After completing the installation, I retrieved the CRD resources.

tptuser@linux-ide:~/Work/infra/lc/theiacloud-0.8.1.MS2/theia-cloud$ kubectl get crd
NAME                                  CREATED AT
appdefinitions.theia.cloud            2023-11-07T05:46:16Z
certificaterequests.cert-manager.io   2023-09-01T02:04:38Z
certificates.cert-manager.io          2023-09-01T02:04:38Z
challenges.acme.cert-manager.io       2023-09-01T02:04:38Z
clusterissuers.cert-manager.io        2023-09-01T02:04:38Z
issuers.cert-manager.io               2023-09-01T02:04:38Z
orders.acme.cert-manager.io           2023-09-01T02:04:38Z
sessions.theia.cloud                  2023-11-07T05:46:16Z
workspaces.theia.cloud                2023-11-07T05:46:16Z

But when querying for detailed information, I encountered an issue.

Cr was not obtained here, there should be some errors in some links, but I have no idea how to solve it.

The current resource status in the cluster and the open status in the browser：

I hope you can give me some tips and help！Thanks!

[sgraband]

Yes the instructions you found are correct.
Regarding the version, there is a difference between the version of the helm chart and the version of Theia Cloud components. You need to use a helm chart that references the appVersion you want to use in the Chart.yaml.

Your helm install fails, because, like the error message indicates, you still have a helm chart deyployed with the name theia-cloud-crds. So this needs to be uninstalled first before you can reinstall using the same name again.

Note, that the theia-cloud-crds helm chart is created using the base terraform chart, therefore you might have missed uninstalling it.

Furthermore, when only installing the theia-cloud-crds helm chart, there are obviously no appdefinitions available, as this helm chart only provides the CRDs and not the CRs. Those are added with the theia.cloud chart.

[wyhdil]

I deleted the chart of the previous version of theia-cloud-crds and installed the 0.8.1. MS2 version, but it still doesn't seem to work

Open IDE in the browser and still report the status code of 473.

[sgraband]

In the last console output under App Version you can see, that your 3 helm charts that are installed are all for different app versions. This needs to be streamlined. Best case, to the newest version. Then you might need to adapt your values for the helm charts. Especially you need to use the correct images for that version.

If you still get a 473 you can also take a look at the logs of the service to see what is the reason for the failure there.

[wyhdil]

OMG, there are some problems now.
I replaced all three charts of theia-cloud with the latest version of 0.8.1. MS3, and then replaced the version of try-now-page, service, and operator with 0.8.1. MS3.

Now the pod looks normal, but when I opened it in the browser, I found a new problem and I don't understand why.

Here is the landing-page pod：

tptuser@linux-ide:~/Work/infra/lc/theiacloud-0.8.1.MS3/theia-cloud$ kubectl describe pod landing-page-deployment-6d5f5877b9-ttnnl  -n theiacloud 
Name:             landing-page-deployment-6d5f5877b9-ttnnl
Namespace:        theiacloud
Priority:         0
Service Account:  default
Node:             minikube/192.168.49.2
Start Time:       Tue, 07 Nov 2023 17:59:20 +0800
Labels:           app=landing-page
                  pod-template-hash=6d5f5877b9
Annotations:      checksum/config: 6097b0fc002d7b00562d57cb83713a973aa0d75c188b6d260fc212a4f8c37e24
Status:           Running
IP:               10.244.1.81
IPs:
  IP:           10.244.1.81
Controlled By:  ReplicaSet/landing-page-deployment-6d5f5877b9
Containers:
  landing-page-container:
    Container ID:   docker://a4218f2a382ca8c8ac759434155bb1b24391a9b3d05b75f4588e7067d137bd88
    Image:          theia-cloud-try-now-page:0.8.1.MS3
    Image ID:       docker://sha256:2acbdc593078bac941fbf54cd326fcff9d7127f60a7f5e11509e767a968b3ee0
    Port:           80/TCP
    Host Port:      0/TCP
    State:          Running
      Started:      Tue, 07 Nov 2023 17:59:20 +0800
    Ready:          True
    Restart Count:  0
    Readiness:      http-get http://:80/config.js delay=5s timeout=1s period=5s #success=1 #failure=3
    Environment:    <none>
    Mounts:
      /usr/share/nginx/html/config.js from landing-page-config (rw,path="config.js")
      /usr/share/nginx/html/logo.svg from landing-page-config (rw,path="logo.svg")
Conditions:
  Type              Status
  Initialized       True 
  Ready             True 
  ContainersReady   True 
  PodScheduled      True 
Volumes:
  landing-page-config:
    Type:        ConfigMap (a volume populated by a ConfigMap)
    Name:        landing-page-config
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                 node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type    Reason     Age    From               Message
  ----    ------     ----   ----               -------
  Normal  Scheduled  8m19s  default-scheduler  Successfully assigned theiacloud/landing-page-deployment-6d5f5877b9-ttnnl to minikube
  Normal  Pulled     8m19s  kubelet            Container image "theia-cloud-try-now-page:0.8.1.MS3" already present on machine
  Normal  Created    8m19s  kubelet            Created container landing-page-container
  Normal  Started    8m19s  kubelet            Started container landing-page-container

Why does this error occur?

[sgraband]

I would assume, that the landing page is not hosted at the address that you have entered. Can you please post the logs of the landing page, as well as the values you used?

kubectl logs --selector=app=landing-page --follow --tail=-1 -n theiacloud right after installing Theia Cloud.