flink-on-k8s-operator

flink-on-k8s-operator copied to clipboard

Hello. I deployed the operator to my Kubernetes cluster (v1.20.4) by following the User Guide. The only change was to fix the controller-gen version to 0.2.4 mentioned on #266. Then...

heojay

More reviewers & slack channel

3

I see we have a lot of open PRs & issues and looks like we will benefit from more reviewers. @functicons As the repo owner, do you have any objection...

shashken

make deploy 有错误

2

make deploy IMG=127.0.0.1:5000/flink-operator:v1.7 /home/zhangyun/gopath/bin/controller-gen "crd:maxDescLen=0,trivialVersions=true" rbac:roleName=manager-role webhook paths="./api/v1beta1/..." output:crd:artifacts:config=config/crd/bases go mod tidy kubectl apply -f config/crd/bases Warning: apiextensions.k8s.io/v1beta1 CustomResourceDefinition is deprecated in v1.16+, unavailable in v1.22+; use apiextensions.k8s.io/v1 CustomResourceDefinition The...

VincentZhangy

Entrypoint for submitting a job

6

Question: The operator is using `flink run` as the entrypoint, while the flink's official docker is using `java org.apache.flink.container.entrypoint.StandaloneJobClusterEntryPoint`, what's the difference between them? I'm asking because I have an...

Yufan-l

Does flink operator support flink1.12?

1

flink operator supports per-job mode and session mode deployment. It is found that the per-job mode is not supported in flink1.12 anymore. Does flink-operator still support job cluster flink1.12?

william-wang

a working example of a flink session cluster is here: https://github.com/mgunter-platform/flink-on-k8s-operator/blob/master/config/samples/flinkoperator_v1beta1_flinksessioncluster.yaml

mgunter-pivotal

adds connection timeout to job submitter

3

Hi all, While experimenting with the Flink operator, I experienced some occasional glitches when the very first `curl` done by the job submitter hangs for a very long time before...

sv3ndk

auto start from savepoint when upgrade the image?

2

I custom build flink image with jars. at now for savepoints ``` autoSavepointSeconds: 300 restartPolicy: FromSavepointOnFailure ``` This is only for `OnFailture`. Can we auto restore from latest savepoint when...

jiamo

Support for TaskManager Autoscaling

5

TaskManager Autoscaling is an important feature for production Flink workloads. Currently, I can't create a [Horizontal Pod Autoscaler](https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/) for the FlinkCluster custom resource. K8s provides support for autoscaling custom resources...

TsvetanKonstantinov

Not Able to trigger a new deployment when image pull failed or job failed

1

There is an issue for operator that it is not able to trigger a new deployment when image pull failed or job failed. More specifically, when I make a deployment...

yanghui16355