Error setting up HA for JobManager

[sfarsaci]

Flink supports HA for the job manager . I tried to configure it in my cluster but doesn't seem I can set more than one replica for the JobManager. When I try to do so I get the following error:
for: "k8s/metrics-aggregator.yaml": admission webhook "vflinkcluster.flinkoperator.k8s.io" denied the request: invalid JobManager replicas, it must be 1

This is an excerpt of my descriptor:

spec:
  flinkVersion: "1.13"
  image:
    name: docker.atl-paas.net/atlassian/innoweek-flink-k8s:0.0.0
  jobManager:
    podAnnotations:
      iam.amazonaws.com/role: arn:aws:iam::XXXXXXXXXX:role/sfarsaci-flink-k8s-dev
    accessScope: Cluster
    replicas: 3

Is HA for the job manager possible using flink-on-k8s-operator or am I doing something wrong?

[eddiewang]

@sfarsaci were you able to figure this out?

[aidan-melen]

We got the same error...

[regadas]

Thanks for the report y'all; I'll have a look.