-
Notifications
You must be signed in to change notification settings - Fork 828
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
seldon-core-operator CRD's incompatible with K8s v1.18 #1675
Comments
Thanks for these details. We will need to see what we need to add to the Kustomize definitions. |
Same problem! Downgrade Kubernetes to 1.17.0 and it fixes. |
FYI this affects local development as well, because kind images are not pinned to a specific version |
I can confirm. After downgrading my minikube using: minikube start --kubernetes-version v1.17.0 I can get past the issue in the seldon-core setup. |
The same error came up with the prometheus operator. I can confirm that when I adapted the schema to make the property required, it fixed the problem and the install works again on 1.18+ (though not sure if that is a proper "fix" for your operator) In the prometheus operator they fixed the problem by upgrading their CRD API to v1 for kubernetes >= 1.18 and keeping also the v1beta1 API for kubernetes < 1.16 |
Thanks for the info we will be looking to solve this post 1.2.0 release for 1.3.0 release |
I hit this problem when installing kubeflow (w/ |
Yes that is the core reason. We need to upgrade to v1 CRDs but need at same time to keep backwards compataibility for older k8s. So I think we will need options to create both. |
+1 need this backward compatible |
It should be. The helm chart checks for k8s version and deploys a v1beta1 CRD for k8s <1.18 and a v1 CRD for >- 1.18. |
kubectl version Helm install is resulting in the same error helm install seldon-core seldon-core-operator --repo https://storage.googleapis.com/seldon-charts --set usageMetrics.enabled=true --set ambassador.enabled=true |
The fix is in master and will be in next release @arunbenoyv |
Separate but related ques here: When we set the |
This is used for cases where we want the operator to create the CRD - its usually used just in Marketplaces where we can't install the CRD separately first |
So on a vanilla kubernetes install, it can be set to false and model
serving with mlflow and metrics in grafana will work?
…On Thu, Feb 17, 2022 at 10:00 AM cliveseldon ***@***.***> wrote:
This is used for cases where we want the operator to create the CRD - its
usually used just in Marketplaces where we can't install the CRD separately
first
—
Reply to this email directly, view it on GitHub
<#1675 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABHVQHEWDCNKBZGRSJBSAA3U3ULQ7ANCNFSM4MEF3WDA>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
You are receiving this because you commented.Message ID:
***@***.***>
|
Worth following our install methods to see if they fit your use. With Helm the chart will install the CRD. |
Installing the seldon-core-operator on a fresh v1.18 Kubernetes install throws an error:
I believe this is due to an API change in v1.18.
I was able to work around by installing a crd definition exported from a successful v1.17 install and installing seldon-core-operator with
crd.create=false
.Steps to reproduce:
The text was updated successfully, but these errors were encountered: