Waiting for Deployment ingress-nginx-controller or my-gitea to become ready never ends #430
Comments
|
Hmm interesting. Can you give me the output of this command? I am guessing there was an issue with pod creation. kubectl describe pod -n ingress-nginxDo you see any errors in the ingress nginx pod? |
|
Thanks @nabuskey , in fact, 1st, when I tried to run anu kubectl command, I was getting
I dont see any ingress-nginx pod created so the command fails with the following error |
|
Another strange thing is that when I issue the command So it is trying port 8080 and not usual 6443 port, why it is looking for 8080 in spite of the fact in .kube/config it is indicated |
|
Were you able to resolve this? You shouldn't need to install kubeadm or kubelet. All you need is docker. The kubectl output seems to indicate your kubeconfig is misconfiugred somehow. Do you have a valid kubeconfig? |
|
No Sir. I think yes it is correct, here is the content of .kube/config
and when I run
It points to the right cluster An here is the hosts file
When I run
|
|
You seem to have another cluster created by docker? When idpbuilder creates a cluster, by default, it creates a cluster inside of a docker container, and it is named
|
|
Hi, no thers is no such container localdev-control-plane. Do you mean I should stop kubernetes to run in Docker desktop settings? For the command I get
|
If you are running a k8s cluster with docker desktop, that's fine. I don't foresee it being an issue. If you run this command, what do you get? You shouldn't need to run idpbuilder with root privilege as long as docker is setup correctly. |
|
Nothing
I ran
I waited almost for 5 minutes then errored as follows, and localdev-control-plane disappeared from the containers list |
|
What do you get if you run |
|
This is what I get which I think should be correct as container does not exist |
|
I see kind deletes them if it fails creating nodes. I opened a PR to make this process much easier to debug. This should help once it gets merged. In the meantime, are you able to create kind cluster using kind cli? Installation should be something like: # For AMD64 / x86_64
[ $(uname -m) = x86_64 ] && curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.24.0/kind-linux-amd64
# For ARM64
[ $(uname -m) = aarch64 ] && curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.24.0/kind-linux-arm64
chmod +x ./kind
sudo mv ./kind /usr/local/bin/kindThen run |
|
Good morning, I did another try, I stopped kubernetes in docker desktop and reissued
|
|
and it continues, I will leave it in this loop until if possible I receive a response from your side
|
|
Sorry, I did not see your response that you posted at midnight. So what should I do now? delete manually localdev-control-plane in docker desktop interface 1st then apply the commands |
This seems awfully similar to what I was helping @cmoulliard with. He was using Fedora 41 as well. If you still have it running, could you run this command? Do you see error messages like these? If so, the fix was:
apiVersion: v1
data:
allow-snippet-annotations: "true"
proxy-buffer-size: 32k
use-forwarded-headers: "true"
worker-processes: "4"
kind: ConfigMap
metadata:
labels:
app.kubernetes.io/component: controller
app.kubernetes.io/instance: ingress-nginx
app.kubernetes.io/name: ingress-nginx
app.kubernetes.io/part-of: ingress-nginx
app.kubernetes.io/version: 1.11.2
name: ingress-nginx-controller
namespace: ingress-nginx
|
|
@nabuskey Yes, the command and I was able to issue
So then what does this prove and help for running idpbuilder correctly? |
This proves that your system can create kind cluster correctly and we should be able to do the same. Where you get stuck is very likely in the components we install (argocd, ingress-nginx, and gitea). I think my previous comment here is the next step. #430 (comment) |
|
OK, so I understand you need to update idpbuilder, how much time can this take? Should I now create the file /tmp/cm.yaml and run idpbuilder create again? |
|
You don't need a new version of idpbuilder. You should be able to use your existing one. |
|
@nabuskey I did the steps you mentioned ,
when I run Enclosed 2 files, for describe command ingress-nginx-controllerDescribe.txt and again, idpbuilder is stuck in a loop |
|
Looks like your node doesn't have enough CPU resources. What machine are you running this on? Ingress-nginx only requests 100m CPU so I am surprised you are getting this. Run something like:
If your machine is underpowered, you could run it in Codespaces. |
|
It is a Hyper-VM has 4 CPUs and 8Gig of RAM |
|
Hmm does the node actually reflect available resources? By default, it should make all resources available on the machine. |
|
As you can see I am in the right context but it is erroring
|
|
OK, here is the content of kubectl describe nodes, it seems everythinhg is OK I issued I issued I got but still cant access the argo site nor gitae |
|
Oh the node has 1 CPU available. And 95% of it is already utilized. Therefore K8s cannot schedule the pod. And So. Why isn't it picking up the number of CPUs you have available. hmmm. |
|
Really, I am not a k8s expert, have you came across this? how can we tell the node to use more CPUs? |
|
In docker settings, I increased the CPU to 2 instead of 1, now it shows Capacity: Allocated resources: cpu 1050m (52%) 100m (5%) I was able to access the argo site
gitea also is accessible Can you please share some tutorials to go further. I agree with you we shoul;d be able to be more able to debug setup process |
|
Can you share what this command shows please ? |
|
Also, can you please tell me if it is possible to access argo and gitea remotely, such as from my lapto or another desktop inside my network and outside it as I dont want to access the VM and work |
Yes you can on your laptop using the urls documented here: https://cnoe.io/docs/reference-implementation/installations/idpbuilder/usage#basic-usage You can also create a cluster on a remote machine and access it from another one using as parameter Remark: To discuss such questions, can you then contact the idpbuilder members using slack - https://github.com/cnoe-io/idpbuilder?tab=readme-ov-file#community please ! |
|
Thanks @cmoulliard , here is the outpu of |
I don't see any disk or cpu/memory issue according to Can you access the following url locally: https://argocd.cnoe.localtest.me:8443/applications |
|
Yes as I said earlier, I am able now to access both, gitea took some more time to get ready that argocd container. However, as agreed with @nabuskey idpbuilder needs to give the right details for not being able to go through |
|
@cmoulliard & @nabuskey please let me ask this last question and I will switch to slack. At https://cnoe.io/docs/reference-implementation/integrations/reference-impl As you know that I was struggling to make local-dev up and running, is it possible to add the different componenets indivisually (Argo Workflows , backstage, Crossplane.........)? If yes, what is the shape of the command and params? Thanks again |
Yes. You can install a package individually or several packages as a package is locally a folder containing manifest files + Argocd application(set) file. Example to install top of the kubernetes cluster (= kind + ingress + gitea + argocd as core packages) the custom package1 |
If you are talking about the ability to select components to install from the ref impl, it's not possible right now. A stack is a tightly coupled package bundle that work together, at least for now. To make it more modular, we have to think about how to go about it cleanly. We discussed about this in previous meetings but we haven't reached any conclusions.
It's possible with SSH tunneling or any other tunneling mechanisms that map the port to your local loop back interface. For example, for my own development setup, I use SSH and SSM agent. For example: ssh -L 8443:localhost:8443 user@hostThen open your browser, go to |
|
@cmoulliard , I did clone, cd to stacks and tried to run
This happens with any other package I tried |
|
@nabuskey I tried |
|
I assume docker is running, correct? e.g. commands like Can you try it with the newest version (v0.8.1)? https://github.com/cnoe-io/idpbuilder/releases/tag/v0.8.1 |
|
Yes @nabuskey |
|
It should override it, but if it doesn't you can download it from the release page I linked above. The version should be |
|
OK Il will try |
|
I updated to verion 0.8.1, now whateever I run I get
when I run idpbuilder create --use-path-routing --package ref-implementation I get |
|
Ah you need to re-create the cluster because --use-path-routing flag is incompatible with your previous run. So run: idpbuilder create --use-path-routing --package ref-implementation --recreate |
|
Thanks, but it errors wheather I use --use-path-routing or not! so what is the difference between a sstandard idpbuilder create and idpbuilder create --use-path-routing? |
|
Did you pass the |
|
OK I will do but please tell me what is the bdifference between using -use-path-routing and not using? |
|
Path based routing vs domain based routing is documented here. If you recreate the cluster, you will lose your config changes. Depending on what config changes you made, there are ways to make it reproducible. What did you change? If you want to just test something out in a new cluster, you can always create a new one with the name and port flag. E.g --name new-cluster --port 7443. |
|
Thanks, I did recreate with the flag --use-path-routing, in spite of the fact during create I got also and endless loop, I rebooted, it seems that all namespaces, pods are there up and running but I cant access https://argocd.cnoe.localtest.me:8443/ nor https://gitea.cnoe.localtest.me:8443/
then |
|
@nabuskey please ignore my last email as I recreated the cluster without the flag --use-path-routing , everything went though no loop of waiting and I was able to access argocd and gitea. |
|
@nabuskey really very interesting this tool idpbuilder, it seems that there has been a lot of efforts behind and I appreciate it.
|
|
Hey @eliassal, I've been busy with KubeCon. Sorry for the late reply. Are you still having issues? |
What is your environment, configuration, and command?
Fedora 41 64 bit hyper-V VM
Followed instructions at https://cnoe.io/docs/reference-implementation/installations/idpbuilder/quick-start using RPM package
Docker destop works fine
What did you do and What did you see instead?
I ran
idpbuilder create
VM named local-dev-control created .On the command line, since 3 hours, I have the following output that never ends, it gets printed every 1 minute as if idpbuilder is not able to continue
When I try to access https://argocd.cnoe.localtest.me:8443/ or https://gitea.cnoe.localtest.me:8443/ I get
Here is a snpashot of the terminal
Ths log of the VM in docker desktop does not seem has any issue
Additional Information. .
I killed the process with Ctrl + C, surprisingly, I got a message indicating ArgoCD was created successfyly, I can get its password but still cant access it in firefox
The text was updated successfully, but these errors were encountered: