Charts: [stable/prometheus-operator] Error: customresourcedefinitions.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" already exists (with no existing CRDs)

[RuBiCK]

BUG REPORT

Version of Helm and Kubernetes: Kubernetes v1.11.5 | Helm v2.12.0

Which chart: stable/prometheus-operator

What happened: When installing stable/prometheus-operator fails with customresourcedefinitions.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" already exists
even checked that the CRD doesn't exist previously

What you expected to happen: Install without errors

How to reproduce it (as minimally and precisely as possible):
First I ensure there is no previous version installed:

$ helm ls --all
NAME            REVISION    UPDATED                     STATUS      CHART                       APP VERSION NAMESPACE
kube-dashboard  2           Mon Dec 31 11:13:48 2018    DEPLOYED    kubernetes-dashboard-0.10.2 1.10.1      kube-system

I remove CRD as described in doc:

$ kubectl get crd --all-namespaces
NAME                                    CREATED AT
alertmanagers.monitoring.coreos.com     2018-12-31T10:42:46Z
eniconfigs.crd.k8s.amazonaws.com        2018-12-31T09:00:10Z
prometheuses.monitoring.coreos.com      2018-12-31T10:42:46Z
prometheusrules.monitoring.coreos.com   2018-12-31T10:42:46Z
servicemonitors.monitoring.coreos.com   2018-12-31T10:42:46Z

$ kubectl delete crd/alertmanagers.monitoring.coreos.com crd/prometheuses.monitoring.coreos.com crd/prometheuses.monitoring.coreos.com crd/prometheusrules.monitoring.coreos.com crd/servicemonitors.monitoring.coreos.com
customresourcedefinition.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" deleted
customresourcedefinition.apiextensions.k8s.io "prometheuses.monitoring.coreos.com" deleted
customresourcedefinition.apiextensions.k8s.io "prometheusrules.monitoring.coreos.com" deleted
customresourcedefinition.apiextensions.k8s.io "servicemonitors.monitoring.coreos.com" deleted

There is not CRD:

$ kubectl get crd --all-namespaces
NAME                               CREATED AT
eniconfigs.crd.k8s.amazonaws.com   2018-12-31T09:00:10Z

With the environment clean (no CRD or previous helm installation), the installation fails:

$ helm install stable/prometheus-operator --name "monitoring" --namespace monitoring
Error: release monitoring failed: customresourcedefinitions.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" already exists

If I delete helm installation and CRD, it happens the same error

$ kubectl get crd --all-namespaces
NAME                               CREATED AT
eniconfigs.crd.k8s.amazonaws.com   2018-12-31T09:00:10Z

$ helm install stable/prometheus-operator --name "monitoring" --namespace monitoring --debug
[debug] Created tunnel using local port: '64985'

[debug] SERVER: "127.0.0.1:64985"

[debug] Original chart version: ""
[debug] Fetched stable/prometheus-operator to <local path>/.helm/cache/archive/prometheus-operator-1.1.0.tgz

[debug] CHART PATH: <local path>/.helm/cache/archive/prometheus-operator-1.1.0.tgz

Error: release monitoring failed: customresourcedefinitions.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" already exists

[philoserf]

This issue has already been reported here and identified as an issue with helm/helm.

[wirehead]

To be more specific: This is tracked by helm/helm#5054 and upgrading to Helm 2.12.1 should rectify the problem.

[albertorm95]

I'm having that issue, what should I do to resolve it?

[thavlik]

I'm having this issue as well after running helm delete --purge on prometheus-operator. Can't reinstall, even with --replace.

[solohin]

The same for
```bash
~$ helm install --name prometheus-operator stable/prometheus-operator

--namespace git2prod-system
--set grafana.enabled=false

Error: customresourcedefinitions.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" already exists

~$ helm version
Client: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}
Server: &version.Version{SemVer:"v2.14.1", GitCommit:"5270352a09c7e8b6e8c9593002a73535276507c0", GitTreeState:"clean"}```

[thavlik]

I was able to get past the issue by deleting all of the CRDs created by prometheus-operator (list them with kubectl get crd -A). I'm not able to retrieve the snippet right now, but I'll post back later today with it.

[solohin]

kubectl delete crd prometheusrules.monitoring.coreos.com
kubectl delete crd servicemonitors.monitoring.coreos.com
kubectl delete crd alertmanagers.monitoring.coreos.com
kubectl delete crd prometheuses.monitoring.coreos.com
kubectl delete crd alertmanagers.monitoring.coreos.com
kubectl delete crd podmonitors.monitoring.coreos.com

helm install --name prometheus-operator stable/prometheus-operator\
 --set grafana.enabled=false

[thavlik]

Yep, that's it!

[Aaron-ML]

Just to note I'm still seeing this with helm 2.14.3

[abdennour]

Delete all monitoring CRDs.
workaround of @solohin works with me !!

it seems that I deployed prometheus from https://operatorhub.io and I forgot to delete it before using helm.

[lchilka]

I am trying to deploy prometheus-operator for the first time using helm and yet i face the error "Error: customresourcedefinitions.apiextensions.k8s.io "alertmanagers.monitoring.coreos.com" already exists".
I checked the crds with "kubectl get crd -A" and found "alertmanagers.monitoring.coreos.com" was one of them.
Does "alertmanagers.monitoring.coreos.com" come out of the box with Kubernetes?

[aqt01]

kubectl delete crd thanosrulers.monitoring.coreos.com

kubectl delete crd prometheusrules.monitoring.coreos.com
kubectl delete crd servicemonitors.monitoring.coreos.com
kubectl delete crd alertmanagers.monitoring.coreos.com
kubectl delete crd prometheuses.monitoring.coreos.com
kubectl delete crd alertmanagers.monitoring.coreos.com
kubectl delete crd podmonitors.monitoring.coreos.com

helm install --name prometheus-operator stable/prometheus-operator\
 --set grafana.enabled=false

Great!
Just adding here another CRDs

kubectl delete crd thanosrulers.monitoring.coreos.com