search

kubernetes / ingress-nginx

Ingress-NGINX Controller for Kubernetes
https://kubernetes.github.io/ingress-nginx/
Apache License 2.0
17.28k stars 8.21k forks source link

apply ingress rule error after install ingress-nginx: x509 certificate is not valid ingress-nginx-controller-admission.ingress-nginx.svc #5968

Closed liminghua999 closed 3 years ago

liminghua999 commented 4 years ago 
k8s cluster install by binary  （i  also try v1.18.0）
[root@m-etc-1 ssl-nginx-webhook]# kubectl get no
NAME      STATUS   ROLES    AGE   VERSION
m-etc-1   Ready    <none>   20h   v1.18.6
m-etc-2   Ready    <none>   20h   v1.18.6
m-etc-3   Ready    <none>   20h   v1.18.6
n-1       Ready    <none>   20h   v1.18.6
n-2       Ready    <none>   20h   v1.18.6
n-3       Ready    <none>   20h   v1.18.6
slb-1     Ready    <none>   20h   v1.18.6
slb-2     Ready    <none>   20h   v1.18.6

and calico install is ok
and coredns install is OK
and ingress-nginx install is OK  
ingress-nginx deploy file:
https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/provider//baremetal/deploy.yaml

[root@m-etc-1 cfg]# kubectl api-versions | grep admissionregistration.k8s.io
admissionregistration.k8s.io/v1
admissionregistration.k8s.io/v1beta1

AND:  --enable-admission-plugins=NamespaceLifecycle,LimitRanger,ServiceAccount,ResourceQuota,NodeRestriction,DefaultStorageClass,DefaultTolerationSeconds,MutatingAdmissionWebhook,ValidatingAdmissionWebhook,PersistentVolumeClaimResize,PodPreset 

[root@m-etc-1 ssl-nginx-webhook]# kubectl get all -n ingress-nginx
NAME                                       READY   STATUS      RESTARTS   AGE
pod/ingress-nginx-admission-create-v27qd   0/1     Completed   0          20h
pod/ingress-nginx-admission-patch-599bf    0/1     Completed   0          20h
pod/ingress-nginx-controller-dsg2j         1/1     Running     1          14h
pod/ingress-nginx-controller-jwjxk         1/1     Running     1          14h

NAME                                         TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
service/ingress-nginx-controller             ClusterIP   10.244.81.145    <none>        80/TCP,443/TCP   20h
service/ingress-nginx-controller-admission   ClusterIP   10.244.170.231   <none>        443/TCP          20h

NAME                                      DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
daemonset.apps/ingress-nginx-controller   2         2         2       2            2           in=ingress      20h

NAME                                       COMPLETIONS   DURATION   AGE
job.batch/ingress-nginx-admission-create   1/1           5s         20h
job.batch/ingress-nginx-admission-patch    1/1           5s         20h

I try apply ingress rule,but I get error:
[root@m-etc-1 ~]# kubectl apply -f ingress-nginx.yml
Error from server (InternalError): error when creating "ingress-nginx.yml": Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": Post https://ingress-nginx-controller-admission.ingress-nginx.svc:443/extensions/v1beta1/ingresses?timeout=30s: x509: certificate is valid for kubernetes, kubernetes.default, kubernetes.default.svc, kubernetes.default.svc.cluster, kubernetes.default.svc.cluster.local, not ingress-nginx-controller-admission.ingress-nginx.svc

WHY???
liminghua999 commented 4 years ago

No one meet this error?? I try the k8s cluster (install by kubeadm)

k8s version: v1.18.6
docker version: 19.03.12 
os: centos 7.6
helm3

apply ingress rule still report same error ;

If I disable webhooks ,then I apply ingress rule,it is OK;

21ki commented 4 years ago

kubectl delete -A ValidatingWebhookConfiguration ingress-nginx-admission

costela commented 3 years ago

@liminghua999 can you please check if the output of these two commands match?

$ kubectl -n ingress-nginx get validatingwebhookconfigurations ingress-nginx-admission -ojsonpath='{.webhooks[0].clientConfig.caBundle}'

$ kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}'
21ki commented 3 years ago

[root@master01 ~]# kubectl -n ingress-nginx get validatingwebhookconfigurations ingress-nginx-admission -ojsonpath='{.webhooks[0].clientConfig.caBundle}' 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 [root@master01 ~]# kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}' 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

md-waldron commented 3 years ago

I am seeing this problem also:

Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": Post "https://ingress-nginx-controller-admission.kube-system.svc:443/extensions/v1beta1/ingresses?timeout=30s": x509: certificate signed by unknown authority

This is in minikube with Ingress-nginx installed using minikube addons enable ingress

I can 'resolve' it using kubectl delete -A ValidatingWebhookConfiguration ingress-nginx-admission but this does not seem like a solution, its just a hack to get it working.

ghfalcon7 commented 3 years ago

I'm running into the same issue, did anyone solve this?

marvinnitz18 commented 3 years ago

I'm running into the same issue, did anyone solve this?

Me too, is it new ?

KaivalyaDabhadkar commented 3 years ago

I am seeing this problem also:

Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": Post "https://ingress-nginx-controller-admission.kube-system.svc:443/extensions/v1beta1/ingresses?timeout=30s": x509: certificate signed by unknown authority

This is in minikube with Ingress-nginx installed using minikube addons enable ingress

I can 'resolve' it using kubectl delete -A ValidatingWebhookConfiguration ingress-nginx-admission but this does not seem like a solution, its just a hack to get it working.

Hi , I am also facing this same issue, did you find any resolution for this other than deleting the Webhook configuration?

md-waldron commented 3 years ago

No, not yet. It's annoyingly, but the work around is not too troublesome, so we have not spent more time trying to fix it.

On Tue, 24 Nov 2020, 11:45 KaivalyaDabhadkar, notifications@github.com wrote:

I am seeing this problem also:

Internal error occurred: failed calling webhook " validate.nginx.ingress.kubernetes.io": Post " https://ingress-nginx-controller-admission.kube-system.svc:443/extensions/v1beta1/ingresses?timeout=30s": x509: certificate signed by unknown authority

This is in minikube with Ingress-nginx installed using minikube addons enable ingress

I can 'resolve' it using kubectl delete -A ValidatingWebhookConfiguration ingress-nginx-admission but this does not seem like a solution, its just a hack to get it working.

Hi , I am also facing this same issue, did you find any resolution for this other than deleting the Webhook configuration?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kubernetes/ingress-nginx/issues/5968#issuecomment-732891723, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABBOUR5V5CFA2BFTIMLELFLSROMFZANCNFSM4PTALKBA .

sschne commented 3 years ago

@md-waldron What version of minikube, kubernetes and which minikube driver are you using? Can you describe what you are doing to reproduce this issue?

md-waldron commented 3 years ago

@sschne I updated my docker and k8s recently and I no longer seem to have the issue. Here is the startup I get when starting minikube, seems all the version info you wanted is reported in that:

😄 minikube v1.14.0 on Darwin 11.0.1 ✨ Using the virtualbox driver based on existing profile 👍 Starting control plane node minikube in cluster minikube 🔄 Restarting existing virtualbox VM for "minikube" ... 🐳 Preparing Kubernetes v1.19.2 on Docker 19.03.12 ... 🔎 Verifying Kubernetes components... 🔎 Verifying ingress addon... 🌟 Enabled addons: storage-provisioner, default-storageclass, dashboard, ingress 🏄 Done! kubectl is now configured to use "minikube" by default

cloud-66 commented 3 years ago

in kubernetes 1.17.13 i have the same issue and don't know how to solve it.

raider444 commented 3 years ago

Server Version: version.Info{Major:"1", Minor:"19", GitVersion:"v1.19.4", GitCommit:"d360454c9bcd1634cf4cc52d1867af5491dc9c5f", GitTreeState:"clean", BuildDate:"2020-11-11T13:09:17Z", GoVersion:"go1.15.2", Compiler:"gc", Platform:"linux/amd64"}

docker://19.3.13

The same problem.

codeclown commented 3 years ago

Borrowing from answers above, here's what resolved this for me. Issue started when I deleted an entire namespace that had contained nginx-ingress, and then tried to reinstall everything via helm. Kept getting upon helm upgrade --install:

Error: UPGRADE FAILED: failed to create resource: Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": Post "https://foobar-ingress-nginx-controller-admission.foobar.svc:443/networking/v1beta1/ingresses?timeout=10s": x509: certificate signed by unknown authority (possibly because of "x509: ECDSA verification failure" while trying to verify candidate authority certificate "nil1")

Solution:

# Find name of the ingress-nginx-admission resource
kubectl get -A ValidatingWebhookConfiguration
# Delete it
kubectl delete -A ValidatingWebhookConfiguration <name>
# Example:
kubectl delete -A ValidatingWebhookConfiguration foobar-ingress-nginx-admission
timd73 commented 3 years ago

I'm encountering this issue, and I don't have a valid workaround, because my ingress-nginx is deployed with fluxcd, so every time I delete the ValidatingWebhookConfiguration, it is recreated.

I am not sure if it's related, but I followed @costela's instructions, and this command has no output:

kubectl -n ingress-nginx get validatingwebhookconfigurations ingress-nginx-admission -ojsonpath='{.webhooks[0].clientConfig.caBundle}'

This command

kubectl -n ingress-nginx get validatingwebhookconfigurations ingress-nginx-admission -ojsonpath='{.webhooks[0].clientConfig}'

shows

{"service":{"name":"ingress-nginx-controller-admission","namespace":"ingress-nginx","path":"/networking/v1beta1/ingresses","port":443}}

Is this the problem, no caBundle? Can someone help solve this?

Or is there a workaround to delete the ValidatingWebhookConfiguration from the source yaml, which I took from here? Can someone explain how exactly to do that?

damienleger commented 3 years ago

My fix for this issue without deleting the validatingwebhookconfigurations. Inspired by https://github.com/kubernetes/ingress-nginx/issues/5968#issuecomment-700287814

CA=$(kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}')
kubectl patch validatingwebhookconfigurations ingress-nginx-admission --type='json' -p='[{"op": "add", "path": "/webhooks/0/clientConfig/caBundle", "value":"'$CA'"}]'
imranrazakhan commented 3 years ago

@liminghua999 Why you close this issue? I am still getting it with K8S 1.21.1 and 0.46.0 ingress.

kristopher-bredemeier-gcx commented 3 years ago

I ran into the the same problem. I fixed it by deleting and reapplying the ingress-nginx-admission-patch job.

juozasget commented 3 years ago

Running the ingress-nginx-admission-patch job fixed it for me as suggested by @kristopher-bredemeier-gcx

If you are looking for an easy way to trigger the ingress-nginx-admission-patch job. Helm chart version upgrade will trigger it. apiVersion: batch/v1 kind: Job metadata: name: ingress-nginx-admission-patch annotations: helm.sh/hook: post-install,post-upgrade

dannystaple commented 3 years ago

I've seen this where I'd accidentally introduced quotes into the class-name for the controller in the args for Deployment-ingress-nginx-controller.yml - doing some substitution so I could consider multiple controllers (on different network segments). Don't put quotes there, and avoid underscores.

The log for the ingress controller deployment will show: Invalid value: "ingress-controller-leader-\"nginx\"": a lowercase RFC 1123 subdomain must consist of lower case alphanumeric characters,

marv254 commented 3 years ago

My fix for this issue without deleting the validatingwebhookconfigurations. Inspired by #5968 (comment)

CA=$(kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}')
kubectl patch validatingwebhookconfigurations ingress-nginx-admission --type='json' -p='[{"op": "add", "path": "/webhooks/0/clientConfig/caBundle", "value":"'$CA'"}]'
xiazemin commented 3 years ago

Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": an error on the server ("") has prevented the request from succeeding

felipess19 commented 3 years ago

the solution of @marv254 works for me. thks.

beastob commented 2 years ago

Inspired by @marv254 's comment.

I had a similar issue with missing caBundle in the validatingwebhookconfigurations/ingress-nginx-admission, and it was installed via ArgoCD and a customized ingress-nginx helm chart.

Since ingress-nginx-admission-patch has an annotation "helm.sh/hook": post-install, caBundle will only appear in the validatingwebhookconfigurations after the Chart is fully installed.

In my case, my custom chart contains some resources that required the admission webhook with the caBundle, and that caused Helm chart to never reaches the 'post-install' stage.

My solution is to add the annotation "helm.sh/hook": post-install to my custom resource as well.

danivendetta commented 2 years ago

Hi

I've having the same problem. And based in the comment

My fix for this issue without deleting the validatingwebhookconfigurations. Inspired by #5968 (comment)

CA=$(kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}')
kubectl patch validatingwebhookconfigurations ingress-nginx-admission --type='json' -p='[{"op": "add", "path": "/webhooks/0/clientConfig/caBundle", "value":"'$CA'"}]'

I've been investigating. I don't use Helm, and I've had to configure each component individually.

There're two Jobs for nginx-admission-webhooks, create and patch. And Based in the Image's documentation we have a command line with their especifications:

https://github.com/jet/kube-webhook-certgen#patch

  --patch-validating              If true, patch validatingwebhookconfiguration (default true)

Doing that I've got a correct validatingwebhookconfiguration config.

I hope that help you.

damienleger commented 2 years ago

@danivendetta you put the same link twice, typo I think.

danivendetta commented 2 years ago

@danivendetta you put the same link twice, typo I think.

@damienleger Thanks, I've edit the post.

dpankros commented 2 years ago

I'm just adding this in case someone else hits the same issue as me. In my case, we use Hasicorp Consul to secure our k8s mesh. As part of that process, I had used to the consul.hashicorp.com/transparent-proxy-exclude-inbound-port annotation on Nginx to exclude inbound traffic but encrypt the mesh traffic, but I did not include 8443, which the admission webhook uses. When the webhook tried to access the controller via the admission service it failed because the port was not excluded and was therefore hitting the envoy sidecar (serving up tls for the service mesh). When I added 8443 to the exclude inbound port annotation, the error went away.

tontondematt commented 2 years ago

Hi

I've having the same problem. And based in the comment

My fix for this issue without deleting the validatingwebhookconfigurations. Inspired by #5968 (comment)

CA=$(kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}')
kubectl patch validatingwebhookconfigurations ingress-nginx-admission --type='json' -p='[{"op": "add", "path": "/webhooks/0/clientConfig/caBundle", "value":"'$CA'"}]'

I've been investigating. I don't use Helm, and I've had to configure each component individually.

There're two Jobs for nginx-admission-webhooks, create and patch. And Based in the Image's documentation we have a command line with their especifications:

https://github.com/jet/kube-webhook-certgen#patch

  --patch-validating              If true, patch validatingwebhookconfiguration (default true)

Doing that I've got a correct validatingwebhookconfiguration config.

I hope that help you.

can you please detail what you have done

mrnonz commented 2 years ago

Should we reach a conclusion about the root cause? and how can this be avoided in the future?

AnthonyWC commented 2 years ago

Ran into this issue as well; in my case I deployed multiple ingress controllers in different namespaces and created 2 different ValidatingWebhookConfiguration (as they are non name-spaced cluster wide).

What fixed it (for me) was deleting both ValidatingWebhookConfiguration and then re-apply CA patches for each one.

hnikt-jonasfh commented 1 year ago

We also ran into this issue when reinstalling the nginx ingress via a customized Helm chart. The ingress deployment went fine, but any ingress object would get the x509 signed by unknown certificate authority error, and the validationwebhookconfiguration had no caBundle prior to fixing. The patching solution as explained by @damienleger fixed this for us, although we do have to run a replace sync for the ingresses that failed due to our configuration of Argo CD, but that's not related to the issue discussed here. I also wonder why this issue is closed when it is still a live issue.

Our customized Helm chart uses nginx ingress helm chart version 4.4.x, according to the dependencies. Controller image tagged at version 1.5.1.

dbeltman commented 1 year ago

For anyone banging their heads against this:

TLDR: Metallb/Loadbalancer was not running/working correctly causing the service not to be available. TLS/x509errors threw me on a wild goose chase.

My only indication was that the service created by nginx-ingress was not put in a ready state ( nothing apparent except ArgoCD saying it wasnt ). When investigating i saw that i had a config error in my metallb chart values. After fixing that the service finally became available, all errors disappeared and i could create ingresses again. (be sure to terminate sync-action in flux/argo if applicable and retry sync)

Jeansen commented 1 year ago

My fix for this issue without deleting the validatingwebhookconfigurations. Inspired by #5968 (comment)

CA=$(kubectl -n ingress-nginx get secret ingress-nginx-admission -ojsonpath='{.data.ca}')
kubectl patch validatingwebhookconfigurations ingress-nginx-admission --type='json' -p='[{"op": "add", "path": "/webhooks/0/clientConfig/caBundle", "value":"'$CA'"}]'

I faced the same issue after having kille the Ingress controller pod. When it got recreated a deployed Jaeger operator complaind not being able to call the ingress-nginx-controller-admission webhook on port 443. After I applied the patch, all was fine again.

leewoobin789 commented 11 months ago

this seems to occur even in the latest version. is there any planned prevention mechanism for the race condition that ValidatingWebhookConfiguration is missing the CA? patching ValidatingWebhookConfiguration manually sounds like a quick fix, not a promising solution.

micheljung commented 9 months ago

Also facing this issue with 4.8.4. Can this be reopened?

blackbass64 commented 3 months ago

The patching solution as explained by @damienleger fixed this for us, although we do have to run a replace sync for the ingresses that failed due to our configuration of Argo CD, but that's not related to the issue discussed here. I also wonder why this issue is closed when it is still a live issue.

When managing ingress-nginx Helm chart with Argo CD, the job for patching the admission webhook certificate won't work because Argo CD doesn't fully support Helm hooks. Instead, leverage Argo CD's built-in resource hook for a smoother integration. These hooks automate injecting the CA certificate from Secret into the ValidatingWebhookConfiguration after Argo CD finishes syncing, ensuring the certificate is available when needed.

Here's the Fix (values.yaml):

controller:
  admissionWebhooks:
    annotations:
      argocd.argoproj.io/hook: PostSync
MattLamont commented 3 months ago

The patching solution as explained by @damienleger fixed this for us, although we do have to run a replace sync for the ingresses that failed due to our configuration of Argo CD, but that's not related to the issue discussed here. I also wonder why this issue is closed when it is still a live issue.

When managing ingress-nginx Helm chart with Argo CD, the job for patching the admission webhook certificate won't work because Argo CD doesn't fully support Helm hooks. Instead, leverage Argo CD's built-in resource hook for a smoother integration. These hooks automate injecting the CA certificate from Secret into the ValidatingWebhookConfiguration after Argo CD finishes syncing, ensuring the certificate is available when needed.

Here's the Fix (values.yaml):

controller:
  service:
    admissionWebhooks:
      annotations:
        argocd.argoproj.io/hook: PostSync

@blackbass64 This fix also worked great for me, but I believe the values.yaml file needs a correction. It should be:

controller:
  admissionWebhooks:
    annotations:
      argocd.argoproj.io/hook: PostSync