search

GoogleCloudPlatform / flink-on-k8s-operator

[DEPRECATED] Kubernetes operator for managing the lifecycle of Apache Flink and Beam applications.
Apache License 2.0
658 stars 266 forks source link

Helm fails with latest operator image while make deploy works #384

Open shravangit20 opened 3 years ago

shravangit20 commented 3 years ago

Hi,

I am tryin to use the latest Flink Operator image and tried to deploy with Helm and Make deploy. Helm fails to deploy while make works with the same latest image used. Is this something to do with the kubectl/k8 version I am using (1.18) ? If yes, only Helm has this issue? Any help is appreciated.

After Helm install the pod stuck in containercreating state forever: helm install dna-flink-operator . --set operatorImage.name=gcr.io/flink-operator/flink-operator:v1beta1-7,flinkOperatorNamespace.name=flink-operator-system,flinkOperatorNamespace.create=false

NAME READY STATUS RESTARTS AGE pod/flink-operator-controller-manager-9bff44956-5pl6x 0/2 ContainerCreating 0 19m

After make deploy: NAME READY STATUS RESTARTS AGE pod/flink-operator-controller-manager-5b5f5fcd95-hvngf 2/2 Running 0 2m47s

Thanks, Shravan

shravangit20 commented 3 years ago

Can anyone please help with this? I also noticed that the make deploy works with the latest operator image but the Flink cluster with 1.12 does not get deployed. I kept my yaml same as 1.11 except for the flink image version.

Are there any known compatibility issues across these latest image versions of operator and Flink? Please share any trouble shooting ideas if these were already addressed.

Thanks, Shravan

shravangit20 commented 3 years ago

Could someone confirm if the operator image version 1.8/1.7/1.6 are stable and ready to be used ? I see the last tag verified in the git repo is 1.5 and want to confirm the same.

functicons commented 3 years ago

The latest operator image version v1beta1-8 is good, but seems we missed the tag in Git repo, will add it.